ERSA Dortmund 2002 - Semantic Scholar

ERSA 2002 Dortmund: 42nd Congress of the European Regional Science Association (August 27th–31st 2002)

The Irish Indigenous Software Industry: Explaining the Development of a Knowledge-intensive Industry Cluster in a Less Favoured Region MIKE CRONE Northern Ireland Economic Research Centre, 22-24 Mount Charles, Belfast BT7 1NZ, UK E-mail: [email protected]

Web-site: www.qub.ac.uk/nierc

ABSTRACT Over the last decade, the Republic of Ireland has emerged as a hotbed of software development activity. The software industry was one of the leading sources of employment growth in the Irish economy during 1990s. Although foreign direct investment has been part of the story, it is the emergence of a vibrant indigenous software industry that is particularly notable. The indigenous industry presently employs 14,000 people and generates annual revenues of €1.4 billion. It is highly export-oriented, has growing productivity and is increasingly innovative. Several of its leading firms have floated on the Nasdaq stock exchange. The emergence of a cluster of knowledge-intensive businesses, in what has been traditionally regarded as a less favoured region, is conceptually interesting and certainly of relevance to policy-makers in comparable regions. The paper sets the ambitious goal of explaining the development of the Irish indigenous software industry, drawing on an examination of several existing studies and various additional secondary data sources. It begins by describing the key characteristics of the industry. It then reviews some of the key factors and processes that have contributed to its development, within the context of the recent literature on clusters of knowledge-intensive industry. The initial establishment of the cluster can be attributed to the presence of some favourable factor conditions coupled with the firm and market-building efforts of pioneering entrepreneurs, who used a range of astute company development strategies. The Irish State played an important role in creating the key factor conditions and in supporting the development of indigenous software firms with various direct and indirect measures. More recently, the development of the indigenous software industry has been sustained by various agglomeration economies and localised collective learning processes. Some important international ties have complemented these local dynamics.

CONTENTS

ABSTRACT ................................................................................................................................................... 1 INTRODUCTION .......................................................................................................................................... 1 PROFILE OF THE IRISH SOFTWARE INDUSTRY................................................................................... 2 Overview.................................................................................................................................................... 2 The ‘Overseas’ Software Industry.............................................................................................................. 5 The Indigenous Software Industry ............................................................................................................. 6 RELEVANT LITERATURE AND EXPLANATORY FRAMEWORK ..................................................... 12 Agglomeration economies ....................................................................................................................... 12 Collective learning/knowledge dissemination processes ......................................................................... 13 Two phases of cluster development: towards an explanatory framework................................................ 14 EXPLAINING THE ORIGINS AND INITIAL ESTABLISHMENT OF THE INDIGENOUS SOFTWARE INDUSTRY .................................................................................................................................................. 17 A window of opportunity ......................................................................................................................... 17 Favourable factor conditions.................................................................................................................... 19 Spark of entrepreneurship ........................................................................................................................ 22 Firm-building and market-building strategies .......................................................................................... 24 Supporting role of State and semi-State institutions ................................................................................ 28 EXPLAINING HOW THE DEVELOPMENT OF THE INDIGENOUS SOFTWARE INDUSTRY HAS BEEN SUSTAINED..................................................................................................................................... 31 An improved environment for entrepreneurship and firm-building ......................................................... 31 Agglomeration economies ....................................................................................................................... 32 Localised collective learning/knowledge dissemination processes.......................................................... 35 CONCLUSION............................................................................................................................................. 44 REFERENCES ............................................................................................................................................. 50 FIGURES AND TABLES ............................................................................................................................ 54

INTRODUCTION

Over the last decade, the Republic of Ireland has emerged as a hotbed of software development activity and a key node in the global software industry. The software industry has been one of the leading sources of employment growth in Ireland’s ‘Celtic Tiger’ economy, with employment in the sector growing at a rate of 16% per annum during the 1990s (compared to 6% for the economy as a whole). By 2000, the Irish software industry comprised approximately 900 companies, with an estimated 30,000 employees and a combined annual turnover of €10 billion (National Software Directorate statistics cited in HotOrigin Ltd, 2002).

One ingredient in the emergence of the Irish software industry has been a major influx of foreign direct investment in the sector, starting in the mid-1980s and continuing through to the present day (Coe, 1997; NSD, 1997; Ó Riain, 1997). As a result the ‘overseas’ segment of the Irish software industry currently employs around 16,000 people and includes most of the world’s leading software multinationals (e.g. Microsoft, Oracle, IBM/Lotus, Symantec, Sun Microsystems)1. The second component of the Irish software story is the parallel emergence of a vibrant ‘indigenous’ software industry2. The indigenous industry comprises over 700 firms, employs around 14,000 people and generates annual revenues of €1.4 billion. It is highly export-oriented, has growing productivity and is increasingly innovative. Several of its leading firms have floated on the Nasdaq stock exchange and some firms are global players within their particular market niches.

This paper attempts to explain the development of the indigenous segment of the Irish software industry, drawing on a review of previous studies and an analysis of various secondary data sources. The paper contributes to the existing literature by presenting a case study of the development of an indigenous cluster of knowledge-intensive businesses in what has been traditionally regarded as a less favoured region. Most existing empirical studies have focussed on the development of such clusters in ‘core’ regions3. The story of the Irish indigenous software also has wider relevance to academic

1

debates about the processes involved in cluster formation and to policy-makers in regions who might seek to emulate Ireland’s success in this key knowledge economy sector.

The structure of the paper is as follows. It begins with a descriptive account of the key characteristics of the Irish indigenous software industry (IISI). Relevant sections of the existing literature are then reviewed with an emphasis on theoretical and empirical studies of clustering in knowledge-intensive industries. An explanatory framework is derived from this literature and is then used as the basis for a two-stage account of the development of the IISI. The paper concludes by commenting on the wider implications of this case study and drawing out the possible lessons for policy-makers in similar regions.

PROFILE OF THE IRISH SOFTWARE INDUSTRY

Overview

The National Software Directorate provides the best series of statistics on the Irish software industry. Headline statistics are available for the period 1991-1999 covering employment, the number of companies, revenue and exports - disaggregated into indigenous and overseas segments (Table 1)4. The data show that the Irish software industry experienced a dramatic and significant expansion on all four available indicators during the 1990s. Total employment expanded from around 7,800 in 1991 to almost 25,000 in 1999 - an increase equating to 15.6% per annum. Average employment growth in the second half of the 1990s (20.6% p.a.) was almost twice that in the first half (10.9% p.a.). The number of companies in the Irish software industry grew from 365 in 1991 to more than 800 in 1999. The number of companies also expanded faster in the late 1990s than the early 1990s. Both total revenue and total exports of the Irish software industry grew at an average annual rate of over 15% during the 1990s.

2

It is important to highlight some significant contrasts between the overseas (foreignowned) and indigenous (Irish-owned) segments of the Irish software industry (Tables 1 and 2). First, the balance of employment between the indigenous and overseas segments remained fairly constant throughout the rapid growth of the 1990s, although employment growth in the indigenous segment – a very rapid 24% per annum – was faster than in the overseas segment in the late 1990s (Table 1). Second, the make up of the overseas and indigenous segments is very different. There are many more indigenous firms than overseas firms but the former are much smaller on average (Tables 1 and 2). Third, the number of overseas software companies in Ireland increased during the 1990s due to ongoing inward investment but average employment size among overseas firms also increased during the period (Tables 1 and 2). Fourth, the indigenous software industry was characterised by a particularly high start-up rate and a low closure rate during the 1990s (O’Gorman et al, 1997). As a result of the birth of so many new firms, mean employment size in the indigenous industry remained small (Table 2).

A fifth observation is that both total revenue and total exports of the Irish software industry are dominated by the overseas sector, which accounted for over 80% of revenue and 85% of exports in 1999. However, the indigenous segment increased its share of revenues and exports through the 1990s; e.g. exports by indigenous firms increased by 37% per annum (Table 1). Both average revenue per company (company size) and average revenue per employee (productivity) in the indigenous sector also increased steadily and significantly during the 1990s (Table 2). Finally, although it appears that overseas software firms are far more productive (Table 2) it is actually the case that overseas software companies in Ireland having unusually high sales per employee (O’Gorman et al, 1997). This is because the revenue figures for the foreign-owned segment are inflated by the presence of high visibility US packaged software firms (e.g. Microsoft, Lotus, Oracle, Symantec, Novell and Claris) who have chosen Ireland as their base for selling into the EU market (see below).

The existing literature on the Irish software industry is helpful in making sense of the above contrasts between the indigenous and overseas segments of the Irish software

3

industry. Ó Riain (1998) identifies “two quite different and relatively autonomous global production/innovation chains”. The first of these software production/innovation chains which has also been documented by Coe (1997) - is described as ‘software logistics and localisation’. This production/innovation chain dominates the overseas sector and consists mainly of the subsidiaries of major US packaged software MNCs, which carry out software localisation, testing and distribution functions in Ireland. These ‘low-end’ activities are part of a wider international packaged software production/innovation chain, which sees ‘higher’ functions - including software development and marketing conducted in the United States (Coe, 1997; Ó Riain, 1997). The firms concerned are among the leading software companies in the world and they produce large volumes of (typically) standardised products for a mass market. There is also an “extensive vendor base of printers, translation bureaux and other suppliers” (Ó Riain, 1997, p.12-13) in Ireland which supports the US packaged software MNCs5.

The second of Ó Riain’s software production/innovation chains dominates within the indigenous segment and consists primarily of small and medium sized Irish firms producing niche software products (often in small volumes) for export (Ó Riain, 1998). The leading companies in this sub-sector “are gaining growing recognition in international technical markets, are building partnerships with US firms and in some cases even going public in the US” (Ó Riain, 1997, p.12-13). The typical presence of higher functions such as development and marketing at these firms means this sub-sector has “a more rounded software development business model” (Ó Riain, 1998, p.5) than the ‘logistics and localisation’ sub-sector.

The final theme addressed in this overview is the geographical distribution of the Irish software industry. The overriding feature of the geography of the Irish software industry is the dominance of Dublin; 83% of all software employment and 76% of all software companies were located in the greater Dublin area in 1998 (Sheridan and Sterne, 1999). This concentration on Dublin is much stronger than in other industry sectors. For example, the Dublin region accounts for just 27% of Ireland’s industrial establishments and 24% of its industrial employment6. Dublin also has only 40% of Ireland’s services

4

employment (O’Gorman et al, 1997, p.15). The secondary centres of software industry activity in Ireland are the Mid-West region (notably Limerick-Shannon) with 6% of software employment, Galway with 4% of national employment, followed by the Cork & South-West region (Stern and Sheridan, 1999). The key message here is that when we talk about the Irish software industry we are, to a large extent, effectively talking about a greater Dublin software industry.

The ‘Overseas’ Software Industry7

An analysis of the overseas segment of the Irish software industry reveals it to be composed of at least five major sub-sectors: 1. Localisation, Manufacturing and Distribution of Software Packages. These are the (often US-owned) packaged software MNCs discussed above. The Irish operations of these firms are typically specialised in relatively low value added, sometimes low skilled, activities including porting of legacy products on new platforms, localisation (text translation, changing formats, etc), disk duplication, assembly, packaging, and distribution/fulfilment (Coe, 1997; Ó Riain, 1997). 2. Software Supporting Subcontractors (SSS). The very existence of this sub-sector can be attributed to the fact that some of the subsidiaries of packaged software MNCs outsource or sub-contract certain functions within Ireland. This phenomenon has stimulated the development of a specialised vendor base of localisation/translation bureaux, printers, disk manufacturers, and logistics specialists. 3. Software Development Centres (SDC). The other major sector of the overseas segment of the Irish software industry is composed of software development centres. These dedicated software development operations undertake work of varying degrees of sophistication but typically employ a much higher percentage of computer science graduates and software engineers than the LMD firms. Some SDCs are part of major computer services or IT consulting MNCs (e.g. EDS, IBM, ICL and Accenture) and other are operated by non-software firms, including electronics MNCs (e.g. Motorola, Ericsson).

5

4. International Software and Computer Services Firms Serving Local Markets. There are a number of firms in the overseas segment that serve purely local (domestic) markets (Coe, 1997; Ó Riain, 1997). The Irish IT market is relatively small, accounting for just 0.6% of the EU total in 1994.8. 5. Acquired Former Indigenous Companies. NSD statistics indicate that foreign firms had acquired 12 indigenous firms, with over 2,000 employees and annual revenues of IR£174 million, by 1999 (Table 1). Some high profile examples include Kindle Banking Systems (acquired by Misys of UK), Aldiscon (by Logica of UK), Euristix (by FORE Systems of USA) and Saville Systems (by ADC of Canada).

The relative significance of these different sectors within the overseas software industry is suggested by an analysis of the Top 20 overseas software firms compiled by the author (Table 4)9. The analysis reveals that 5 of the Top 20 firms specialised in LMD of software packages, accounting for 36% of employment in the Top 20. The related SSS category accounted for 5 more firms and another quarter of employment in the Top 20. Software development centres account for a quarter of employment in the Top 20, with the remainder of the Top 20 being accounted for by two acquired indigenous firms (who also undertake software development work) and two IT consultancy firms (Table 4). This finding is not radically different from Coe’s (1999) assertion that the software product manufacturing and software development sectors each account for approximately half of the jobs among the leading overseas software firms.

The Indigenous Software Industry

Size analysis of employment and revenue

The aggregate figures in Table 1 conceal considerable variations within the indigenous software industry. An NSD size analysis for 1998 (the most recent available) shows the indigenous industry is composed of a very large number of small companies and a relatively small number of ‘larger’ companies (Table 5). However, it is these larger firms that dominate employment. The 34 largest indigenous firms (those with more than 50

6

employees) accounted for 43% of all indigenous employment in 1998 and the 10 largest firms (more than 100 employees) for almost a quarter. In contrast the 428 smallest firms (1-24 employees) accounted for only 35% of indigenous employment.

The number of larger indigenous firms (50+ employees) has increased substantially over the last decade or so, from only 4 in 1989 to 14 in 1992, 24 in 1995, and 34 in 1998. My investigations suggest there were more than 60 indigenous firms in this group by 2001, including at least 20 with more than 100 employees. Note also that larger firms accounted for a disproportionate share of the new jobs created in the indigenous industry in the mid1990s; firms with more than 50 employees accounted for 43% of indigenous employment by 1998 compared to 33% in 1992.

Larger indigenous firms are even more dominant in terms of revenues. The 34 largest firms (50+ employees) in 1998 earned 56% of all indigenous software industry revenues, and the largest 10 (100+ employees) earned 36% of all revenues (Table 6). Also note the positive relationship between firm size and productivity, with mean revenue per employee being much higher in the largest firms (Table 6).

Export orientation and internationalisation

The Irish indigenous software industry is increasingly export oriented. In 1998, 62% of the revenues earned by the indigenous software industry came from exports (Table 6). Significantly, exporting is not limited to just a few leading indigenous software firms. According to NSD statistics, 81% of indigenous software companies, some 450 firms, were exporting by 1997, and 44% of indigenous firms obtained at least half of their revenues from export sales. Note also that the share of companies exporting increased steadily during the 1990s. However, although a high percentage of indigenous firms are exporters, export revenues for the indigenous sector are dominated by a small number of leading firms (Table 6). In 1998, the largest 34 firms (50+ employees) earned more than two-thirds of export revenues and the largest 10 indigenous firms (100+ employees) earned nearly half. There is therefore a positive correlation between firm size and export

7

orientation; the largest firms obtain a greater share of their revenues from exports (Table 6).

It is also interesting to examine the market orientation of exports from the indigenous software industry (Table 8). As we might expect, the UK and Western European markets are important but, perhaps surprisingly, the United States is the largest export market with 43% of revenues. Coe (1999) argues that the penetration of the “massive and highly competitive US market” is a sign of the growing maturity and international competitiveness of Irish indigenous companies.

The growth in exports in the indigenous software industry has been mirrored by an increase in the number of overseas offices operated by Irish firms. Overseas offices are established to assist with product sales and marketing, and possibly also to give local customer support services. The total number of overseas offices operated by indigenous software firms nearly doubled from just over 50 in 1991 to nearly 100 in 1997 (Source: NSD web-site). The leading location for overseas offices is the UK but the number of UK offices increased only marginally between 1991 and 1997. More notable, however, is the increase in North American offices from 14 in 1995 to 31 in 1997 (see Crone, 2002, for more details). As we have seen earlier, this focus on North America as the primary target market for leading Irish indigenous software product firms is reflected in the high share of export revenues earned from the US. Unfortunately no more current data is available on the number of overseas offices operated by Irish software firms or on the level of employment in them.

In recent years, another form of internationalisation has become important, namely acquisition activity by Irish firms. Based on an analysis of the TIU TechWatch news archive (www.techwatch.ie), Irish software firms made nearly 60 acquisitions in 19992001, including 37 cases where the target was overseas (i.e. not Irish). The overseas targets included 15 firms in the United States and 13 in the UK. Leading indigenous companies SmartForce, Baltimore Technologies, IONA Technologies, Riverdeep, Trintech, Datalex and Parthus were all engaged in multiple overseas acquisitions (see

8

Crone, 2002, for details). The Irish companies seem to be using these acquisitions to broaden their technical competencies and extend their geographical coverage (although the motives require further detailed investigation). It is notable that the most acquisitive Irish firms are publicly listed. Companies that have reached this stage must have international or global aspirations and might be expected to engage in expansionary acquisition activity.

Specialisations: products versus services

The Irish indigenous software industry can be further sub-divided according to a number of specialisations. The primary distinction is between firms developing software products and those providing various software-related services. Software product companies typically “develop a software programme which is then copied many times and sold to many different customers” whereas software services “are provided uniquely to each customer as required, rather than involving repeated selling of copies of a standardised product” (O’Gorman et al, 1997, p.8). Arora et al (2001) argue that many of the products sold by leading Irish indigenous software companies are ‘service-intensive products’ that require a variety of related services (such as installation, training, system integration, support and maintenance).

In terms of the relative importance of products and services within the indigenous software industry there is the following evidence. First, Ó Riain (2000, p.16) finds that “firms are split relatively evenly between a focus on product development (41% obtaining over 50% of their revenues from products) and providing services (37% of all firms obtaining over 50% of their revenues from services)” (based on his 1997 survey). Second, Arora et al (2001) classified over 500 indigenous companies according to their specialisms using data from the company directory on the NSD web-site, supplemented by information from company web-sites (Table 7). This analysis suggests 44% of indigenous software firms are involved in the development of software products. Third, HotOrigin (2001, p.5) claim “there are at least 250 local companies engaged directly in the development of software products”. Note that the product-oriented business model

9

has been adopted by many of the most successful indigenous firms. Also note that product-oriented firms tend to be more focused on exporting whereas service-oriented firms are more likely to serve the domestic market (Ó Riain, 2000).

Other Irish indigenous software firms focus on providing software services, such as bespoke (once-off or customised) development of programmes for individual customers, consultancy and technical training. Arora et al’s (2001) analysis of specialisations in the indigenous software sector indicates the relative importance of different service activities (Table 7). Customised development services are the most common type of service activity, being offered by more than half of all indigenous software firms (Table 3.5). Other common service activities are Internet services, multimedia services and localisation.

Specialisations: product and industry focus

In a sub-sectoral analysis of the industry conducted by the National Software Directorate in the late 1990s, the main sub-sectors identified were niche product development, general business applications, services/bespoke development, banking and finance applications, multimedia/CBT, software development tools/software systems and telecommunications (see Crone, 2002, for details). The highest revenue earning and most productive of these sectors were multimedia/CBT and software development tools/software systems. The most export-oriented sectors were Multimedia/CBT, telecommunications, software tools/ systems software, and banking and finance applications. The services/bespoke development sector was the least productive and the least export-oriented.

In a more recent assessment, HotOrigin (2001) distinguished between specialisations on the basis of (a) product/technology focus and (b) industry focus. In terms of product/technology focus, the indigenous product development industry is heavily concentrated on enterprise application integration (including middleware) and wireless technologies (applications and infrastructure). Other strong niches in Ireland are said to

10

be CRM, e-learning, Java components and XML-based tools (HotOrigin Ltd, 2001, p.8). In terms of industry focus, the main target sectors are the financial services industry, the telecommunications industry and high technology industries. In summarising the current standing of the Irish indigenous industry, HotOrigin claim it is well placed to compete in some of the key high growth areas in software globally (i.e. CRM, wireless infrastructure and applications, e-learning and enterprise application integration).

Leading Firms

In identifying the leading indigenous Irish software firms it is obvious to start with those that have successfully floated on public stock markets (Table 3). HotOrigin (2001) identify the seven publicly listed indigenous software product companies as SmartForce, IONA Technologies, Baltimore Technologies, Trintech, Riverdeep, Parthus and Datalex, all of which are quoted on the high-profile US Nasdaq exchange10. These seven firms have 4,900 employees worldwide and have a combined annual turnover of €875 million (HotOrigin, 2002). They “have shown that they can compete and win at an international level” (HotOrigin Ltd, 2001, p.11). Whilst the firms mentioned above are the undoubted leaders of the Irish software industry, they have now been joined by an array of other firms, who are said to be on the brink of breaking through to international level (e.g. Fineos, CR2, Eontec, Norkom, WBT Systems). Most recently, several fast-growth startups have burst onto the scene, demonstrating their intent to internationalise from the outset (e.g. Cape Clear, Macalla, Network365, Orbiscom, Xiam). The specialisations of the leading indigenous firms include:


financial services applications/solutions [e.g. Fineos, Eontec, CR2, Norkom];




e-security/secure payment solutions [e.g. Baltimore, Eurologic, Trintech, cardBASE];




e-learning/computer based training [e.g. SmartForce, Riverdeep, WBT Systems];




open systems-based middleware [e.g. IONA Technologies, Cape Clear, Macalla];




telecommunications software [e.g. Euristix (now Marconi), Network365, Xiam].

Geographical distribution

11

Finally in this section, it is interesting to consider the extent to which the indigenous industry follows the overall geographical pattern described earlier. In 1991, the fledgling indigenous industry consisted of 291 companies, of which 71% were located in Dublin (NSD, 1992). By 1995, there had been little change in this pattern. O’Gorman et al (1997) using data from the ‘Connect Ireland’ database found that “two-thirds of indigenous software firms (including nearly all of the largest ones) are in the Dublin area” (p.15). My analysis of a list of 126 leading indigenous firms, compiled for this study, shows 76% of these firms are located in Dublin, including nearly all of the Top 30 (Table 9). In terms of employment, Dublin is even more dominant with 87% of employment in the Top 126 firms (Table 9). Hence Dublin firms are larger on average than firms in the rest of the country. Overall it might have been expected that the initial concentration of indigenous firms in Dublin would lessen with time but this has clearly not happened. This suggests there are strong agglomerative forces operating in the Irish indigenous software industry.

RELEVANT LITERATURE AND EXPLANATORY FRAMEWORK

This section reviews the key points from the recent theoretical and empirical literature on clustering of knowledge-intensive industries and introduces an explanatory framework, based on this literature, which is later used to help explain the development of the IISI. Two main approaches to explaining clustering can be identified. The first, more traditional, approach focuses on the role of ‘agglomeration economies’ or ‘external effects’ and is based on the ideas of Marshall (1890) and recent work along similar lines by Krugman (1986, 1991). The second, more recent, approach relies on the notion of ‘collective learning’ processes associated with the evolution of ‘innovative milieu’ (Camagni, 1991, 1995)11. Such approaches have recently been in the ascendency within the economic geography literature.

Agglomeration economies

12

The pure agglomeration economies approach focuses on cost savings that accrue to a firm from its location within a concentration of similar or related firms. The idea of agglomeration economies - sometimes known as ‘external economies’ of industrial clustering – can be traced back to the work of Marshall (1890). Marshall proposed three reasons why firms would continue to be localised within the same area. These relate to the costs savings that result from (1) the development of a local pool of specialised labour, (2) the increased local provision of inputs specific to an industry, and (3) the maximum flow of information and ideas between proximate firms (Gordon and McCann, 2000, p.516). In the case of high-technology industries, a fourth factor that might be added to Marshall’s list of agglomeration economies is that clustering can bring common infrastructure benefits, such as access to major telecommunications hubs/networks (Swann and Prevezer, 1996). As Gordon and McCann (2000) observe, “the key to Marshall’s approach is that what links each source of economic benefits is the fact that this accrues to firms within the local area primarily because of their geographical proximity. None of the sources or results of these benefits is internal to a particular firm, but each is external to all the firms” (p.516).

Marshall’s ideas have recently been revisited and modified by the economist Paul Krugman (1986, 1991). Krugman’s approach to regional agglomeration (i.e. clustering) involves two basic concepts. First, there is the idea that regional specialization evolves for accidental reasons, and second, the notion that once these regions become established, they are sustained by the agglomeration economies or ‘external scale economies’ (Pinch and Henry, 1999). In terms of external scale economies Krugman identifies the same three types of factor as Marshall; i.e. labour market pooling, the creation of specialist suppliers, and the development of knowledge spillovers. However, Krugman concentrates upon the first two types of factor arguing that knowledge spillovers are limited to a few high-technology agglomerations, are national or international - rather than regional - in scope, and are difficult to model quantitatively (Pinch and Henry, 1999).

Collective learning/knowledge dissemination processes

13

It has been recognised by some authors that, when it comes to explaining the clustering of knowledge-intensive industries, spillovers of knowledge (Marshall’s third factor) are – contrary to Krugman’s assertions - likely to be particularly important (e.g. Swann and Prevezer, 1996). In an empirical context, both Pinch and Henry (1999) and Keeble and Nachum (2001) have shown that the pure cost-based agglomeration economies approach, as endorsed by Krugman, has limited utility in accounting for the clustering of certain knowledge-intensive industries12. In searching for an alternative explanation for clustering in knowledge-intensive industries, Pinch and Henry (1999) and Keeble and Nachum (2001) focus on the role of knowledge spillovers and suggest that a ‘collective learning approach’ has greater utility than the pure agglomeration economies approach.

The collective learning approach is “distinguished by its identification of particular processes as crucial for the development of a localised collective learning capacity and resultant enhanced innovativeness in a knowledge-based cluster” (Keeble and Nachum, p.10). It can be traced to the work of Camagni (1991, 1995). The crux of this approach is the identification of the crucial role of knowledge spillovers in the development of clusters of knowledge-intensive industry. Methodologically, it follows that explanations of clustering in such industries should focus on identifying specific ‘collective learning processes’ or ‘knowledge dissemination processes’ and explaining their role in enhancing the dominance of the cluster. In the case of the British motor sport cluster, for example, this approach leads Pinch and Henry (1999) to assert that “the critical role of knowledge helps to account for the geographical clustering of firms in Motor Sport Valley. The continual ‘churn’ of information and people in the region provides a locus of innovation and a wealth of expertise that no other region can replicate at present” (p.825). The key point is that motor sport firms can enhance their innovativeness and competitive edge by tapping into this knowledge pool (i.e. locating in the cluster). Examples of the types of collective learning/knowledge dissemination processes that are likely to underpin successful clusters of knowledge intensive industry are identified in Table 10.

Two phases of cluster development: towards an explanatory framework

14

When it comes to explaining the development of a particular cluster over time, Bresnahan et al (2001) deduce from their study of nascent ICT clusters around the world that “the processes of starting and sustaining a cluster have different economics” (p.10 – emphasis added). They also stress that “founding a new cluster, or the early firms in a new cluster, is a very different entrepreneurial and economic activity than founding a firm in an established cluster” (Bresnahan et al, 2001, p.10). It follows that, in seeking to explain the development of a particular cluster, it is useful to distinguish between (1) explaining the initial establishment or origins of a cluster and (2) the question of what sustains a cluster once it is established. Pinch and Henry (1999) also employ this type of two-phase approach in their explanation of the development of ‘Motor Sport Valley’.

In terms of explaining the processes of cluster development, we can identify a number of approaches in the existing literature to the second question, namely explaining what sustains a cluster once it is established. For example, Krugman focuses on the role of external scale economies (as discussed above). Bresnahan et al (2001) also argue that the second, sustained development, phase can be explained with reference to external effects theory and agglomeration economies (although they place more emphasis on knowledge spillovers and circulation than Krugman). In explaining what sustains the ‘Motor Sport Valley’ cluster, Pinch and Henry (1999) focus on the role of collaborative learning processes and localised knowledge spillovers (as discussed above).

When it comes to the first question, however, the existing literature is somewhat less clear on how to explain the origination and initial establishment of a cluster (Bresnahan et al, 2001)13. Certainly, in the very early stages of cluster development, when only a handful of pioneering firms are present there are unlikely to be significant knowledge externalities or other agglomeration economies, and the scope for collective learning is also likely to be limited. In the case of Krugman’s theories on clustering, the origins of clusters are attributed simply to ‘accidents of history’. However, Pinch and Henry (1999) are critical of Krugman on this point and argue for a more informed examination of the long-term processes operating in particular clusters. Thus, Pinch and Henry (1999) attribute the origins of the ‘Motor Sport Valley’ cluster to a paradigmatic technological

15

shift within the global motor sport industry (from heavy front-engined cars to lightweight mid-engined cars with an increased emphasis on aerodynamics). Britain was able to replace Northern Italy as the centre of the global motor sport industry because (partly by co-incidence) it possessed a series of newly-relevant factor conditions for competing in motor sport that were absent in Italy. Chief among the British advantages were expertise in aluminium engines and aeronautical engineering (aerodynamic design and composite materials became increasingly important for competitiveness). The example of Pinch and Henry (1999) suggests the importance of identifying the structural factors (a window of opportunity) and factor conditions that may lead to a cluster becoming established in a particular place.

Even with the structural opportunity and favourable pre-conditions for success, the establishment of a cluster still requires a further ingredient. Here Bresnahan et al (2001) point to the need for “a spark of entrepreneurship” to get a cluster going, followed by “years of firm- and market-building efforts” (p.11). More specifically, what is required is “significant and systematic efforts by the ‘pioneers’ of a cluster to promote organisational and technological capabilities of various sorts, create new firms and institutions etc.”(Bresnahan et al, 2001, p.13). With regards to firm-building, Bresnahan et al (2001, p.19) assert that the important point for cluster development is whether some firms emerge out of the class of small entrepreneurial start-ups to become major players in world markets. These leading firms, it is argued, “will become one of the sources of increasing returns for the continuous growth of the cluster, in the form of training for potential spinouts, development of managerial and technical competencies, along with various forms of backward and forward linkages” (Bresnahan et al, 2001, p.19). By implication, therefore, it is important to understand the firm-building and company development strategies employed by the leading firms in the cluster, in order to understand the process of cluster development14.

Finally, on the question of explaining the origins and initial establishment of a cluster, it is important to highlight the potentially important supporting role to be played by State development agencies and semi-State institutions. The ‘habitat’ of specialised suppliers

16

and service providers and supporting institutions that underpin a successful technology cluster such as Silicon Valley (Barr and Tessler, 1999) will not be present in an embryonic cluster. The case of the Israeli high technology cluster, for example, suggests that the State can play an important role in underpinning the development of a cluster. The Israeli government provided a range of supports for high tech entrepreneurs including incubator facilities and generous R&D grants (Roper and Frenkel, 2001; De Fontenay and Carmel, forthcoming).

The arguments outlined in the preceding discussion have been synthesised and summarised in the form of a two-phase explanatory framework of cluster development (Figure 1). In the remainder of the paper, this explanatory framework is used as the basis for an explanatory account of the development of the IISI.

EXPLAINING THE ORIGINS AND INITIAL ESTABLISHMENT OF THE INDIGENOUS SOFTWARE INDUSTRY

A window of opportunity

The United States, particularly Silicon Valley, is the undisputed leader of the global ICT industry. Other G7 nations also have well-developed ICT sectors. However, the last decade or so has seen the emergence of a number of new clusters of ICT-related growth outside the G7 nations, including those in Israel, India, Scandinavia, Taiwan and, of course, Ireland (Bresnahan et al, 2001)15. Although there are significant differences between these locations, they have all exhibited a significant acceleration in the production of ICT during the 1990s (ibid). The ICT clusters in India, Israel and Ireland which focus primarily on software and services - have experienced annual double-digit growth in the number of new firms, in ICT revenues and employment, and in exports (ibid). Taiwan has shown a similar pattern of growth in the manufacturing of PCs and related businesses. In Scandinavia (specifically Sweden and Finland) ICT growth was triggered by the emergence of the wireless hardware market (ibid). The key point here is

17

that the emergence of the Irish indigenous software industry is not a unique phenomenon and must be seen within the context of some wider structural changes in the global ICT industry. It is important to identify these structural changes before moving on to consider the specific factors and processes that underpinned the development of the Irish indigenous software industry.

The primary structural factor has been the sustained increase in global demand for ICT products and services over the last decade, primarily driven by the US market. As Bresnahan et al (2001) have observed, the volume and variety of this demand exceeded initial expectations. Successive new technologies (such as the PC, associated hardware and software, and more recently the Internet) turned out to have significantly larger markets than was first anticipated. Newer technologies have also tended to complement existing technologies. For example, the commercialisation of the Internet also raised the level of demand for PCs, Mainframe computers, telecommunications infrastructure and services, and of course software.

In the specific instance of the software industry, Arora et al (2001: p.2) identify four enabling structural factors that opened a window of opportunity for “countries well below the technology frontier but rich in human capital” – such as Ireland, Israel and India. First, software was de-coupled from hardware in the 1980s and this created a burgeoning independent software industry, with the possibility for new firms to emerge and challenge the established industry leaders. Second, rapid improvements in data communications (and communications more generally) allowed for the possibility that software production could be undertaken in locations that were geographically remote from the US market (the dominant source of software demand). Third, the geographical spread of markets for software and the range of possible locations for the industry were widened by the increasing globalization of economic activity. Fourth, the fact that the software industry is highly skilled-labour intensive meant that software activities could be attracted to locations that were rich in specific IT skills. This point was accentuated by the fact that the established centres of ICT production (especially Silicon Valley) were facing increasing pressures on labour and land (Bresnahan et al, 2001, p.25).

18

Favourable factor conditions

Ireland has been able to seize upon the structural opportunity described above to develop a rapidly emerging indigenous software industry (alongside countries such as India and Israel) whereas other countries and regions have not capitalised to the same extent. To understand why this is the case we must first look to the specific range of favourable factor conditions (after Porter, 1999) or location factors (after classical economic geography) that were present in Ireland. As we shall see, the Irish State played an important role in creating some of these factor conditions.

Abundance of appropriate skilled labour

For most commentators, the single most important factor in the initial emergence of the indigenous software industry was the early abundance in Ireland of appropriate skilled labour (O’Gorman et al, 1997; Ó Riain, 1997 and 1998; Arora et al, 2001). The indigenous software industry has a large requirement for specific skills since it employs a high proportion of third-level education graduates, primarily holders of degrees in computer science and related subjects. For example, a 1992 survey by the National Software Directorate found that 75% of employees in indigenous software firms were third-level graduates and a 1996 survey by O’Gorman et al (1997) found that four-fifths of indigenous companies had workforces comprising at least 70% graduates. By implication, the third-level education system is the most important supplier of appropriate skilled labour to the indigenous software industry.

Ireland’s third-level education system rates highly in international comparisons and has for some time been relatively focused on courses relevant to the software industry. Even back in 1992 - in the early stages of the software industry’s development - degrees in ‘mathematics and computing’ amounted to 5 per cent of all degrees awarded “which was the fifth highest percentage among all OECD countries” (O’Gorman et al, 1997, p.18). This focus on software-related subjects has been enhanced during the expansion of the

19

third-level education system in the 1990s16. By 1998, degrees in mathematics and computer science accounted for over 10% of all third-level degrees awarded compared to approximately 6% in the UK and 5% in France and Germany (Source: Key Education Statistics, Department of Education and Science).

Ireland’s relative abundance in skills relevant to the software industry can be traced back to the 1970s, when the Irish government decided to increase the supply of graduates in technical subjects in order to attract foreign direct investment in high-tech sectors (O’Gorman et al, 1997; Ó Riain, 1997). Consequently, Ireland already had a relatively abundant supply of technical labour by the mid-1980s (Ó Riain, 2001, p.31). However, the major expansion in the supply of computing graduates occurred in the early 1990s. The increase came both from an increase in student admissions to existing programmes and also from a greater variety of computing courses - including joint honours degrees with a computing component and postgraduate conversion courses (Arora et al, 2001, p.15). The Regional Technology Colleges (now known as ‘Institutes of Technology’) played a major role in the expansion and accounted for around a half of all computing students by the mid-1990s (O’Gorman et al, 1997, p.18). By 1996, the annual output of computing graduates from the Irish third-level education system had reached over 1,800, comprising over 600 single honours computing degrees plus a variety of joint honours degrees, computing diplomas and computing certificates (McIver Consulting, 1998). It is widely accepted that this massive expansion in skilled labour supply resulted in a surplus of computing graduates, relative to local demand, throughout the first half of the 1990s. Unlike many other countries, Ireland did not experience a significant IT skills shortage until at least 1995 (O’Gorman et al, 1997).

In addition to the abundant supply of third-level graduates, there are other labour-related factors that enhanced Ireland’s potential for developing an indigenous software industry. Firstly, the early indigenous software firms are said to have benefited from the experience gained by managers and technical staff in previous employment with multinationals operating in Ireland (O’Gorman et al, 1997; Ó Riain, 1998). In particular, multinationals are likely to have been important in introducing technically competent Irish computing

20

graduates to the world of international business, and perhaps in developing their managerial abilities. Work experience with multinationals may also have been important in developing the advanced but rather generalised technical skills of computing graduates into more specialised skills, suitable for use in specific sub-sectors of the software industry (O’Gorman et al, 1997, p.21).

The second additional labour-related consideration concerns the possible role of return migration of computing graduates. As Ó Riain (1997) and Arora et al (2001) note, when the Irish economy was in the doldrums in the early 1980s, many of the new wave of graduates emigrated, particularly to the United States. This may ultimately have been a blessing in disguise for Ireland since many of these migrants returned to Ireland in the mid-to-late 1990s. The trend of net emigration from Ireland in the 1980s was reversed in the mid-1990s when net immigration changed from -400 in 1993 to +22,800 in 1998. Over half of the immigrants to Ireland in this period were Irish returning emigrants (statistics cited in Arora et al, 2001, p.20). Ó Riain (1997, p.28) suggests that return migration of people who had worked in multinationals abroad may have been a contributory factor in the development of some indigenous software firms. Arora et al (2001, p.20) suggest returning migrants brought with them “an important background of working experience in advanced technological and business environments” as well as “personal linkages with the international business community”. The evidence in support of these assertions is rather vague at present, and the question certainly requires further investigation, however there is evidence that this process of ‘brain circulation’ has played a role in the emergence of nascent ICT clusters in India, Israel and Taiwan (Bresnahan et al, 2001)17.

The third and final additional point concerning labour supply for the Irish indigenous software industry concerns the opportunity cost of labour in Ireland. Arora et al (2001) argue that Ireland’s traditionally weak industrial base, with low employment opportunities for engineers and professionals, meant that “the growing software industry could readily claim not only new graduates with the required skills, but could also lure away human capital from other sectors of the economy” (p.16).

21

Other factor conditions

Another frequently cited factor, and arguably an essential pre-condition for the development of the IISI, is Ireland’s telecommunications infrastructure. The telecoms network received over IR£2 billion of investment in the 1980s, some of which came from EU programmes such as STAR and TELEMATIQUE (Coe, 1997; O’Gorman et al, 1997). Consequently, the initial establishment period of the IISI was supported by a stateof-the-art telecoms infrastructure, including a 100% digital transmission network and international fibre-optic and satellite links (Grimes, 1999)18. Finally, an additional advantage for Ireland was the English language, which assisted Irish firms in their quest to penetrate the dominant US market (as well as the geographically proximate UK market).

Spark of entrepreneurship

The factor conditions outlined above were vital in preparing the Irish economy for the emerging opportunity in the global software industry but they are not in themselves sufficient to explain the emergence and establishment of the IISI (Ó Riain, 1997 and 1998). It was suggested earlier that, in order to get a cluster started, an initial spark of entrepreneurship is required, followed by years of firm-building and marketing-building efforts by the ‘pioneers’ of the cluster. Regarding the first part of the equation it is useful to identify the origins of the early software start-ups (i.e. find out where the entrepreneurs came from). According to Ó Riain (1997) the early Irish indigenous software firms emerged by three main routes.

Services to products. According to Ó Riain (1997, p.29) many indigenous software product companies began by providing ‘bespoke’ or custom services to businesses, then expanded this business into making consultancy kits and subsequently into products, gradually expanding into export markets. Early customers within Ireland - who commissioned various IT development projects - are said to have provided an important

22

catalyst. Among these customers were MNC subsidiaries in various industries and it has been argued that Ireland’s general success in attracting FDI was important in creating additional local demand for fledgling software firms (O’Gorman et al, 1997; Ó Riain, 1997). Ultimately, however, the limited size of the Irish market may have contributed to the success of the indigenous sector by forcing it to become export-oriented from an early stage (Grimes, 1999).

Spin-outs from larger firms. The second route by which indigenous software firms are said to have emerged is through various types of ‘spin out’ from larger firms. According to Ó Riain (1997) some indigenous software firms were created when firms in other industries, such as telecommunications or computer hardware, spun off their software divisions. Other firms emerged from divisions of MNCs, semi-state bodies and Irish firms. Ó Riain (1997) argues the key dynamic in this process was local in that “domestic managers created new competencies and business for their divisions and convinced the management of the parent company to support their projects” (p.30). A further type of spin out occurred when users of software - in vertical markets such as banking and training – applied their detailed knowledge of these markets to open up opportunities for new software businesses (Ó Riain, 1997, p.30).

Firms based on academic research. The third main group of indigenous firms are those that emerged almost directly from academia, being set up by professors and graduate students based on their on-campus research (Ó Riain, 1997, p.30). Whilst the university spin-off/start-up route is probably not the most numerically significant this should not hide the fact that some of the most successful indigenous software companies originated from a university environment (e.g. IONA Technologies, Baltimore Technologies, Trintech, Piercom). It is widely acknowledged that university spin-offs are among the most technically-sophisticated and fastest-growing firms in the indigenous industry, operating in areas such as development tools, system software, computer-based training and telecommunications (Arora et al, 2001, p.8).

23

Another interesting point of note on this subject is the apparent lack of indigenous software development firms originating as spin-offs from software MNCs operating in Ireland (certainly there are few high profile examples). Given the lack of core development work carried out by the packaged software MNCs in Ireland this may not be entirely surprising. Nevertheless, with many of the leading software MNCs having been established in Ireland for fifteen years, and a significant number of software development centres within the overseas sector of the industry, this apparently low return might be seen as disappointing by policy-makers.

Finally on this subject it is interesting to examine the origins of some of the leading firms in the indigenous industry (in many ways the ‘pioneers’ of the IISI). The first observation is that these pioneering firms emerged from a variety of sources. SmartForce (previously known as CBT Systems), the e-learning/computer-based training specialist and Ireland’s first Nasdaq IPO, is said to have emerged on the back of an Irish government-led ‘courseware’ initiative in the mid-1980s (Ó Riain, 1999; HotOrigin Ltd, 2002). Riverdeep, also in e-learning and listed on the Nasdaq, is a spin-off from SmartForce. Datalex, a provider of IT solutions to the airline and travel industry, was established when the in-house IT development team at Aer Lingus spun-off in 1985 (HotOrigin Ltd, 2002). Parthus, which develops software for semiconductor design, was established in 1993 when Digital Equipment Corporation closed its Irish R&D operations and the core development team stayed together to form a new company (HotOrigin Ltd, 2002). The most common origin among the leading indigenous firms, however, is the academic research route (IONA, Baltimore and Trintech). IONA Technologies, for example, was founded in 1991 on the back of research at Trinity College Dublin on distributed computing, which had been supported by the EU ESPRIT programme and Ireland’s National Board for Science and Technology (Ó Riain, 1999; O’Neill, 2001).

Firm-building and market-building strategies

The leading firms in the IISI are the products of concerted firm and market-building efforts by pioneers of the cluster, in some instances spanning 10-15 years. In terms of

24

firm building, it is possible to identify a number of important strategies that were adopted by many of the most successful indigenous software firms. First, they focused on developing software products for niche markets (rather than providing bespoke services) and, second, they targeted export markets from an early stage, especially the dominant US market. A notable strategy employed by many early indigenous software firms was to begin by providing software services (usually to domestic customers) and then use the income stream generated to fund product development efforts (O’Gorman et al, 1997). This strategy – known as ‘bootstrapping’ – was particularly useful in the early stages of the industry’s development when external sources of finance were extremely scarce.

The product-oriented development path was, to a large extent, defined by the various State agencies and institutions (Ó Riain, 2000). Most notably, the National Software Directorate was clear in identifying software products as the preferred route for the indigenous industry, and in emphasising the need for software product firms to export from an early stage due to the small size of the domestic Irish market (NSD, 1992). This strategic guidance to the industry was also backed up by practical interventions. For example, Forbairt (the predecessor of Enterprise Ireland as the indigenous industry development agency) was apparently more sympathetic to grant applications from software product companies (Ó Riain, 2000, p.36). Some specific policies were also devised to help young product companies, such as assistance with marketing to overcome the problems of exporting (ibid). In advocating the product-oriented route, the State agencies “realized that the greatest value-added within the industry and the clearest way to avoid the low-cost competition trap was through a focus on software products rather than services or international sub-contracting for coding and programming” (ibid).

A further point about the focus of indigenous firms on products is that they tend to be specialised in niche markets (defined by customer type, product type or technical expertise). The advantage here is that Irish firms can avoid direct competition with the leading US multinationals who have advantages of economies of scale and oligopoly dominance of the general-application product segments (O’Gorman et al, 1997, p.37). This strategy of focusing on product spaces that are complementary to the main sources

25

of ICT demand (Silicon Valley and the US) rather than competing directly with them has also been adopted by other emergent ICT clusters such as the Israeli and Indian software industries (Bresnahan et al, 2001, p.20).

A third strategy employed by indigenous software firms has been to internationalise their operations from an early stage. For example, many firms have opened overseas offices in the US, and other key markets such as the UK. Firms primarily use these overseas offices to sell and market their products from within export markets. It has been claimed it is almost essential for non-US software companies to have a US office if they are to secure contracts with major US customers. Ó Riain (1997) cites the example of an Irish Trade Board consultant in the US telling Irish companies that they must “become, look and feel like an American company” in order to succeed in the US market. A second more recent aspect of company internationalisation is merger and acquisition activity. In recent years the leading Irish indigenous companies have shown an inclination to acquire companies in the US, UK and other countries (as discussed earlier). The suggested motives for these takeovers are to access new/complementary technologies and to increase the customer base/market share (although this requires further investigation).

A fourth aspect of company strategy employed by many Irish indigenous software firms is the use of international strategic alliances and OEM agreements. One of the chief motivations for such strategic alliances is that they can allow small Irish indigenous software firms to access the marketing/distribution networks of larger, more established firms. This strategy allows small firms to quickly expand their reach into target markets and ensure sufficient levels of after-sales support, which they would not otherwise be able to achieve by themselves (PWC, 1999). Ó Riain (1999) has shown a link between exporting and international alliances, noting “integration into international markets appears to inevitably bring integration into international inter-firm relationships” (p.231). Interestingly, government agencies and trade associations may have prompted many of these alliances. For example, the Irish Software Association and the National Software Directorate have led trade missions to Massachusetts and Silicon Valley in order to build up these kinds of links (Ó Riain, 1997). A similar alternative to the strategic alliance is

26

the OEM customer agreement. In this scenario, the smaller indigenous software firm secures a, sometimes exclusive, agreement with a leading industry player or OEM (original equipment manufacturer). The highest profile example of this in the IISI is the early agreement between IONA Technologies and Sun Microsystems whereby Sun agreed to incorporate IONA’s middleware product on its Sun workstations in exchanges for a 20% equity stake (O’Neill, 2001). This agreement is said to have been a critical factor in IONA’s success. Among other effects, an OEM agreement allows the Irish software firm to capitalise on the reputation and profile of the OEM.

The fifth strategy of note, which has been employed by only a handful of the leading indigenous firms to date, is to undergo an initial public offering (IPO) on the Nasdaq stock exchange. Although an IPO has traditionally been viewed merely as an exit strategy for venture capitalists and other investors, Teubal et al (2000, p.42) have convincingly argued that an IPO should also be seen as a strategy for company development. The rationale here is that an IPO, with its associated processes, generates new tangible and intangible assets, which can then be exploited to help the company progress to the next stage of its development (Teubal et al, 2000). The tangible assets generated by an IPO include cash and shares, which the company can use to fund expansion. The intangible assets include credibility and reputation (useful for attracting investors and customers), social capital (newly-formed personal and company links which help facilitate future sales) and enhanced managerial capabilities (since the IPO process requires the assembly of a team of experienced managers and the appointment of a Chief Financial Officer).

The implication of the above arguments is that those Irish software firms that have undergone an IPO on the Nasdaq accumulated various assets during the process which have since enabled them to accelerate their development, regardless of whether the IPO was initially undertaken with such clear strategic intentions. Although this argument makes the IPO option sound highly appealing, the IPO process is not without its demands and risks. To be successful on Nasdaq, companies must satisfy some very demanding criteria including quarterly financial reporting and adhering to the business norms favoured by US investors (Ó Riain, 1999). The alleged short-term priorities of US

27

investors, which might not accord with management objectives, have been a source of concern for some Irish business commentators19.

As a final point on firm-building strategies it is worth noting that there may have been an element of learning from past mistakes within the indigenous software industry. The first generation of Irish indigenous software firms from the 1980s (including firms such as Aldus, Workhorse and Glockenspiel) was not as successful as the ‘IONA generation’. Several of these firms fell victim to asset-stripping acquisition activity by foreign firms. Ó Riain (1999) attributes their ‘failure’ to a lack of business savvy (their founders were usually technologists who had no experience of building technology companies or marketing technology products) and to the absence of appropriate supporting institutions, rather than to technical shortcomings (firms such as Glockenspiel were highly innovative). These early firms may have played an important role in the subsequent development of the indigenous industry by fostering technical skills and giving business experience to key individuals within the Irish industry. For example, IONA recruited several former employees of Glockenspiel, who added valuable technical and commercial experience to the core research team from Trinity College (Ó Riain, 1999). Ó Riain (1999) also argues that “seeing the mistakes made by companies like Glockenspiel enabled those who came after to avoid them as best they could” (p.251).

Supporting role of State and semi-State institutions

In the early stages of development of the IISI, Ireland did not have a well-developed supporting ‘habitat’ of the kind that underpins the competitiveness of successful technology regions such as Silicon Valley (a situation to be expected of a fledgling cluster). Here the Irish State played an important role in the early development of the industry by attempting to compensate for this lack of specialised suppliers (e.g. venture capitalists), service providers (e.g. IPR lawyers, specialised marketing consultants) and other supporting institutions. This intervention took two main forms: first there was direct assistance from the State development agencies (both financial and non-financial) and second the State established a number of specialised semi-State institutions.

28

Ireland’s State development agencies played an important role in supporting the firm- and market-building efforts of the cluster ‘pioneers’ by providing early-stage finance and other forms of ‘softer’ assistance. In the late 1980s and early and mid-1990s, when there was no significant private venture capital industry in Ireland, the State agencies were the dominant external supplier of finance to indigenous software firms. For example, fourfifths of the indigenous software firms surveyed by O’Gorman et al (1997) had received some form of State financial assistance. State agencies Forbairt and An Bord Trachtala (predecessors of Enterprise Ireland) offered a range of employment, capital and R&D grants. According to Ó Riain (2000, p.33) a typical package of grant aid to an early-stage indigenous software company in the late 1980s consisted of IR£100,000 spread over 2-3 years. Total grant payments by state agencies to indigenous software companies are said to have amounted to IR£3-3.5 million per annum in 1988-90, and over IR£5 million per annum in 1991 and 1992 (Clarke, 1995, p.143). One notable feature of this grant regime was that finance was often tied to various aspects of company development, in particular marketing and management development (Ó Riain, 2000, p.37).

As well as offering grant aid, the State development agencies have made equity investments in early-stage software companies. By 1992, the State agencies had equity investments totalling IR£2.7 million in 48 indigenous software companies and by 1997 this was reported to have increased to nearly 100 firms (Ó Riain, 2000, p.34). In the most publicised example, Forbairt took an equity stake in IONA Technologies in 1994. When IONA subsequently went public on the Nasdaq in 1997, the state agencies realised a very significant return on their small initial investment. One-third of the equity stake was shrewdly sold-off in 2000, prior to the technology crash, for a reported return of IR£27 million, with the profits being earmarked for reinvestment in other start-up companies20.

The State development agencies also supported the firm- and market-building efforts of the cluster ‘pioneers’ by providing ‘softer’ forms of assistance, including help with marketing, management development and training. This regime of support reflected an increasing emphasis within the State agencies on developing indigenous industry and was

29

part of a strategy described by Ó Riain (2000, p.32) as ‘making winners’. Thus, Forbairt designed a suite of policies aimed at the various stage of company development including feasibility grants for start-ups, employment grants for the early growth phase, R&D funding for product development, and training and management development for the company development phase (Ó Riain, 2000, p.39). An Bord Trachtala became involved in the marketing side with the aim of promoting exporting. The State agencies also provided a kind of consultancy service, supplying technical and market information and helping with the development of business plans (O’Gorman et al, 1997, p.24). Three fifths of the indigenous firm surveyed by O’Gorman et al (1997) had received such assistance. Finally, there was also a mentoring scheme whereby small companies were put in touch with experienced industry figures that, in return for a small sum, offered guidance and sometimes became company directors. This ongoing scheme is said to have involved 12 software companies between 1989 and 1992, and 53 between 1993 and 1996 (Ó Riain, 2000, p.42).

The second important contribution of the State to the development of the software habitat was in the creation of a specialised set of supporting institutions. Perhaps the most important of these have been the National Software Directorate (NSD) and Centre for Software Engineering. These institutions are said to have helped to develop a common sense of purpose between the State agencies and the ‘pioneers’ of the industry and they also promoted business and technical learning within the industry. The NSD was formed in 1991 with the support of EU funding in response to demands from the emerging software industry for better representation within government. It drew its staff from the State agencies and from the industry. Its two directors have been highly respected figures within the industry. The first, Barry Murphy, had success in the 1980s as the Managing Director of Insight Software (a leading indigenous firm). The second director, Jennifer Condon, was a managing director of ICL in Ireland. Although the NSD was a State body, it occupied the middle ground between the State and the industry. The main role of the NSD was to represent the needs of the industry to the development agencies and the government and provide a strategic perspective on the development of the software industry in Ireland. As noted earlier, the NSD has produced influential strategic guidance

30

to the Irish industry (NSD, 1992). It also monitored the development of the industry through its regular surveys. The NSD has also successfully lobbied the government on behalf of the industry on issues such as IT skills supply, access to capital and the telecoms infrastructure.

The Centre for Software Engineering (CSE), located at Dublin City University, was founded in 1991 as one of three software-related centres within the government’s Programme in Advanced Technologies. Of the three centres, the CSE has been the most influential. Its role has been mainly promotional and informational. In particular the CSE has played an important role in promoting best practice in software development within the Irish industry (e.g. the use of quality assurance procedures). The CSE also runs an array of technical and managerial training courses for software industry professionals and offers consultancy services to software SMEs, and in doing fulfils a technology transfer role within the Irish industry. The CSE’s eleven full-time staff, led by founding director Robert Cochran, includes experienced industry professionals. The CSE’s funding came initially from the Irish government and EU programmes but it is increasingly self-funded (commercial revenues from training and consultancy services). It is run as a not-for-profit organisation. The CSE has an active contact group of about 130 companies and deals with about 90 firms in a typical year (including up to 50 consultancy assignments plus attendees at training courses) but it has dealt with several hundred companies during its lifespan21. Thus, it seems likely that the CSE has been very influential within the IISI.

EXPLAINING HOW THE DEVELOPMENT OF THE INDIGENOUS SOFTWARE INDUSTRY HAS BEEN SUSTAINED

An improved environment for entrepreneurship and firm-building

After about 1997, there was a marked acceleration in the growth of the IISI. NSD statistics show that employment, the number of companies, revenues and exports all grew more rapidly during 1997-99 than in the preceding period (Tables 1 and 2). HotOrigin

31

Ltd (2001, p.5) also note a marked increase in the new firm formation rate in the industry in this period, claiming that half of all indigenous software product companies in the industry in 2001 had been founded since the beginning of 1999. Another notable fact is the vast majority of new start-ups in the indigenous software product industry now begin as spin-offs from established indigenous software firms (HotOrigin Ltd, 2002). Together this evidence suggests that the cluster may have reached a stage of maturity where success is breeding further success (i.e. the industry seems to have entered a phase of sustained development). A possible explanation for this phase-shift is that there was, by the late 1990s, an improved environment for entrepreneurship and firm-building in Ireland, as compared with the early 1990s. It is proposed here that this improved environment can be understood as a function of the operation of (1) agglomeration economies and (2) collective learning (or knowledge dissemination) processes within the emerging cluster, which were previously either absent or present to a much lesser extent.

Agglomeration economies

Recall from earlier that agglomeration economies are usually understood as benefits (cost savings) that accrue to a firm as a result of its location within a geographical concentration of similar or related firms. The three main sources of agglomeration economies are a shared pool of skilled labour (a ‘thick’ local labour market), specialised local supply industries (which develop in response to a local concentration of demand), and localised knowledge spillovers. The first two effects are discussed in this section and knowledge spillovers are considered in the subsequent discussion of collective learning processes.

Shared Pool of Skilled Labour

On the first type of agglomeration economy, it is clear that Ireland has now developed a thick technical labour market in software-related skills, particularly in the Dublin region. The prolific output of computing graduates from the third-level education institutions since the mid-1980s has been supplemented by various labour market processes. First,

32

there has been over decade of on-the-job learning within both the overseas and indigenous software sectors, and related sectors such as telecoms and ICT hardware manufacturing. Second, there was a wave of return migration to Ireland in the late 1990s, including many experienced technical and managerial personnel (as discussed earlier). Third, software firms invest heavily in employee training by comparison with other sectors of the economy (O’Gorman et al, 1997). Taken together these facts convey a clear advantage (agglomeration economies) on today’s indigenous software firms compared with their predecessors, namely, there is a much greater range of experienced managerial and technical labour available to growing firms.

Specialised Local Supply Industries

On the second type of agglomeration economy, there is clear evidence of the development of various specialised suppliers and service providers within the cluster, in response to the growing local concentration of demand from indigenous software firms. The most obvious example of this is the emergence over the last five or so years of a sizeable venture capital industry in Ireland, including the increasing involvement of international VC investors. As Ó Riain (2000, p.32) notes, it was only in 1998/99 that private investment capital became abundant in Irish software. Since 1999, numerous emerging indigenous software firms have received first and second round venture funding, ranging from €0.5 to €30 million (Table 11). In fact, venture capital is now the dominant source of external finance for ‘build-phase’ indigenous software companies and is said to be in abundant supply for good business propositions (PWC, 1999; HotOrigin Ltd, 2001 and 2002). My preliminary analysis of news coverage of the industry suggests that venture capitalists invested at least €400 million in indigenous software companies between 1999 and 2001. In fact such is the abundance of VC in Ireland, that PWC (1999) claim the current finance weaknesses of Irish software relate to the demand rather supply side22. Specifically, PWC (1999) claim that some indigenous companies lack the ability to market themselves sufficiently in order to attract funding. Some founders are also said to be reluctant to relinquish equity stakes in their companies (PWC, 1999, p.112).

33

One interesting aspect of the emergence of the VC industry in Ireland is the role of the State in kick-starting the industry. Ó Riain (2000, p.32) notes that the expansion of the VC industry “was initially largely underwritten by the state”. EU funding has also played a vital role, with one particular initiative (the EU Seed and Venture Capital Measure of the Operational Programme for Industrial Development 1994 –1999) being particularly important. This measure had the objective of establishing VC funds to provide equity capital for growth-oriented SMEs in Ireland. The Programme allocated ECU 43.9 million, which was used to leverage private sector commitments, resulting around ECU 90 million being available for investment over the period of the programme. The programme is administered by Enterprise Ireland which has, through its Investment Services Directorate, implemented a policy of forming partnerships with private sector institutions, corporates and venture capitalists to establish new venture capital funds to invest this funding in Irish SMEs (Enterprise Ireland, 2000). All of the available funding had been committed to investments by the end of 2000, through 15 different funds, 13 of which had made investments in software firms (Table 12). In fact, software was by far the leading destination sector for investments, with 98 investments totaling IR£44 million (66% of the total value invested). Managers from the private sector run the individual funds, with guidance from investment committees. From the launch of the Measure in 1996 to the end of December 2000 a total of 214 investments had been made in 101 different companies, totaling IR£66.7 million (Enterprise Ireland, 2000). Seventy-three per cent of the value of these investments went to Dublin-based companies.

A final important dimension of the increasing involvement of VC firms is their potential influence on the management and business strategies of indigenous software firms. It is common practice for VC firms to appoint a non-executive director to the board of companies they invest in and research has shown that VC-appointed non-execs have played an important coaching role to entrepreneurial start-ups in Silicon Valley (e.g. Hellman, 2000; Hellman and Puri, 2002).

There are also signs that a second type of specialised supply industry is developing within the cluster, namely, service firms with a specific focus on the software/technology sector,

34

particularly in Dublin. One interesting example of this development is HotOrigin, an early-stage venture capital and specialist consultancy company. HotOrigin, which is privately-owned, was founded in March 2000 in Dublin by an experienced team from corporate, consulting and start-up backgrounds. Its CEO was formerly Head of Accenture’s Strategy practice in Ireland. One of its aims is to facilitate the growth of young software companies during critical stages of development, by providing essential services such as management team building, business strategy development, technology strategy, strategic alliance partnerships and venture financing.

Another example of the development of specialised business services is the decision of major Boston-based US law firm Brown Rudnick Berlack Israels to open an office in Dublin in 2002. The Dublin office advises Irish technology firms doing (or planning to do) business in the United States about issues such as employment legislation, legal disclosure, product licensing, intellectual property protection, raising venture capital and setting up a US office. The firm is already working with nearly 20 Irish technology companies, including Eurologic Systems and Fineos23. The development of a critical mass of export-oriented software companies is clearly crucial to the attraction of such firms, which should in turn facilitate further success for the IISI.

Localised collective learning/knowledge dissemination processes

Recall that the collective learning approach recognises the potentially crucial role of localised knowledge spillovers in sustaining clusters of knowledge-intensive industry and focuses on the identification of processes by which knowledge is disseminated through the cluster. The ensuing discussion highlights examples from the IISI of some of the key collective learning/knowledge dissemination processes identified by Pinch and Henry (1999) and Keeble and Nachum (2001) (Table 10).

Knowledge circulation due to movement of key individuals

35

According to Keeble and Nachum (2001), localised collective learning can be promoted by flows of professionals and “embodied expertise” through the local labour market. Similarly, Pinch and Henry (1999) note that a high turnover of personnel within a cluster can disseminate knowledge because “when such personnel move, they take with them crucial information about how things are done”. In the case of the IISI, Ó Riain (1998) argues that its particular structure (i.e. small firms with flat hierarchies) encourages a steady turnover of personnel since individual employees often have to move firms to get promoted. Here, the career profiles of five key individuals (‘cluster pioneers’) are used to illustrate the operation of this process within the IISI.

1. Dr Chris Horn. As co-founder, Chairman and former CEO of IONA Technologies, Chris Horn is perhaps the best known, and certainly one of the most successful, entrepreneurs in the IISI (refer Figure 2). Prior to co-founding IONA (with Annrai O'Toole and Dr Sean Baker), he was a faculty member at the Department of Computer Science at Trinity College Dublin. Significantly, he was involved in a programme of research on distributed computing at Trinity, which was funded by the EU ESPRIT programme between 1984-89 and subsequently supported by Ireland’s National Board for Science and Technology (NBST). Key elements of this research were commercialised with the foundation of IONA Technologies in 1991. As joint CEO-Chairman of IONA between 1991-2000, Chris Horn oversaw the sustained growth of the company, including its 1997 IPO on the Nasdaq (IONA was Ireland’s second Nasdaq IPO). Some of Horn’s accumulated expertise (embodied knowledge) has been disseminated through his prominent role in industry affairs. Interestingly there is evidence that this dissemination process has accelerated since Horn resigned the CEO position at IONA in May 2000. In September 2000, Horn was appointed chairman of special industry advisory committee to ITEQ, the new market for Irish technology companies on the Irish Stock Exchange. Then in 2001, Horn was appointed to two non-executive directorships, with Sepro (a Dublin-based billing solutions specialist) and CR2 (a Dublin-based channel banking solutions provider and an emerging star of the IISI). It is likely that, by appointing Horn to their boards, the

36

two firms are hoping to tap-in to his knowledge and experience of the industry to assist them in developing their businesses.

2. Dr Jim Mountjoy. Best known as the founder of Euristix (a successful indigenous telecoms software firm), Jim Mountjoy has a PhD in Telecoms Engineering and his work experience includes spells at the Department of Posts and Telegraphs and the NBST (in the telecommunications area). Complementing these public-sector roles, Mountjoy was Managing Director of Baltimore Technologies between 1984-90. He left Baltimore to found his own company, Euristix, allegedly after an internal dispute at Baltimore, and built Euristix into one of the leading indigenous software firms before it was acquired by US company FORE Systems for US$80m in March 1999. Like Chris Horn, Mountjoy, as a pioneer of the IISI, is now re-applying his knowledge on the board of a number of emerging Irish technology firms (Figure 3). He is Chairman of both Headway Software (a young fast-growing builder of software development tools) and Tsunami Photonics (a spin-off from University College Dublin specialising in laser optics for the telecoms industry). He is also a nonexecutive director of Corvil Networks (a telecoms bandwidth utilisation specialist), Aspect Software (a provider of CRM applications for the banking and insurance industry) and Duolog Technologies (a specialist in next generation wireless technologies). In the latter case Duolog’s CEO, noted the value of Mountjoy’s “experience in growing a successful high tech company” in the press release announcing his appointment.

3. Fran Rooney. Well known as the CEO of Baltimore Technologies, Fran Rooney led Baltimore though a merger with UK firm Zergo Holdings and a Nasdaq floatation in 1999, and oversaw the rapid growth of the company to over 700 employees and annual revenues of US$40 million. Like Horn and Mountjoy, Rooney has also gone on to a string of board appointments (Figure 4). In September 2000, while he was still at Baltimore, he was appointed as a non-executive director by Norkom Technologies (a Dublin-based CRM software specialist). Rooney resigned as Baltimore CEO in July 2001 following the collapse of the company’s share price. Later in 2001 he

37

became Chairman of Addoceo (a developer of e-learning products) and non-executive director of both WestGlobal (a specialist in web service management systems) and Epionet (a Dublin-based web services development specialist). In the latter case, Epionet specified in a press release that Rooney’s experience at the helm of Baltimore would help it target key international markets and guide it though its expansion strategy.

4. Barry Murphy. An Electrical Engineering degree and a Masters degree in Engineering Science from University College Cork began Barry Murphy’s distinguished 25-year career in the Irish software industry, during which he has held various high profile positions (Figure 5). Murphy made his name as MD of Insight Software in the 1980s. Insight was Ireland’s largest indigenous software company when UK firm Hoskyns acquired it in 1988 (Ó Riain, 1999). Murphy’s next role was as the influential first head of the National Software Directorate (1988-96). Subsequently he was CEO of Cullinane Group Ireland, an organisation founded by US software industry veteran John Cullinane to invest in Irish software companies. During this time he was involved in writing a strategy document for the Irish Software Association (the leading industry association). Murphy’s pioneer role differs from those of Horn, Mounjoy and Rooney in that he spent a large spell building some of the key supporting institutions for the IISI. The importance of this type of entrepreneurship has been acknowledged by Bresnahan et al (2001) in their study of Silicon Valley’s Imitators. Bringing things full-circle, Murphy re-entered the private sector of the industry in 1999 when he became the first CEO of Openet Telecom (a fast growing specialist in internet protocol mediation software for global telecoms companies).

5. Robert Cochran. The fifth and final cluster pioneer considered here has over 30 years experience in the IISI in a variety of roles straddling State, semi-State, private sector organisations (Figure 6). Cochran was a member of the NBST in the 1980s (in the area of software/data processing/microelectronics). At the same time he managed his own software company Generics, a technically sophisticated operation which went out of business in 1990 (Ó Riain, 1999). Since 1991, Cochran has been Director of

38

the Centre for Software Engineering, establishing it as the national focal point for promoting best practice in software development and a key node for knowledge dissemination within the cluster. Cochran has also passed on his expertise as an expert advisor to several government agencies and as an occasional member of the executive committee of the Irish Software Association. He maintains links with international best practice in software engineering through his membership of the Industrial Advisory Board for IEEE Software magazine. He retains a personal stake in the IISI through his company Catalyst Software (which specialises in software development tools).

To summarise, the key point which these examples illustrate is that, through their involvement with numerous private firms, State agencies and supporting institutions, these key individuals (and others like them) have, during the course of their careers, accumulated embodied knowledge (about technologies, products, markets and ways of doing business) and disseminated it through the IISI cluster.

Localised spin-offs of new firms from existing businesses

Another important process by which knowledge can be disseminated within a cluster is through “high rates of localised entrepreneurship and spin-off of new firms from existing businesses” (Keeble and Nachum, 2001 p.10). Pinch and Henry (1999) argue that, in knowledge-intensive industries “it is relatively easy for individuals to leave existing companies and set up new companies. Typically, these individuals have spotted new solutions to problems and hope to exploit the market niches thus created” (p.824). As noted previously, spin-offs from existing indigenous technology firms are now the most common source of new software start-ups (HotOrigin Ltd, 2002). Here two recent examples of the spin-off process within the IISI are discussed by way of illustration. Example 1: Macalla Software from Quay Financial Software24. This firm was founded in 1998 when six key executives and technologists left Quay Financial Software after it was taken over by CSK. It specialises in CORBA-based financial frameworks for trading and

39

investment banking, exploiting the opportunities offered by open standards-based middleware. It has raised two rounds of venture funding; IR£1 million March 1999 (valuing the company at IR£3 million) and €8.6 million in September 2000. This funding has been used to build the team, accelerate product development activities and expand worldwide sales and marketing operations. Macalla has won a number of awards and was recently identified by Tornado Insider as one of Europe’s Top 100 emerging private technology companies (June 2002). Some key expertise (embodied knowledge) was transferred in the spin-off, including the former Technical Director’s experience of developing the US market for Quay and technical knowledge accumulated by the technologists in their related work at Quay.

Example 2: Cape Clear Software from IONA Technologies. This a young, fast-growing, Dublin-based firm specialises in XML-based Internet infrastructure products for business-to-business applications. Three former IONA Technologies executives established the firm in August 1999 (David Clarke, Hugh Grant and John McGuire)25. The firm’s CapeConnect middleware systems are built to the CORBA operating standard. CORBA is an industry standards body of which IONA has been an influential member. Cape Clear received US$2m in funding from ACT Venture Capital, Enterprise Ireland and two private US investors in April 2000 in return for 20% of equity. The money was to be used to expand the firm’s US operations26. CapeClear is one of the rising stars of the IISI and has been tipped as a future IPO candidate. Key experience transferred in the spin-off included the founders’ extensive experience in developing and marketing component middleware and Internet products in their previous employment with IONA27. The strong links between Cape Clear and IONA have been reinforced with further defections from IONA by Annrai O’Toole (a co-founder and former CTO of IONA), who was appointed as Executive Chairman, and Colin Newman (also a founding director of IONA) who was appointed Head of Marketing in November 2000. Both men acquired a substantial shareholding in Cape Clear and joined the board as Executive Directors. IONA also has an equity stake in Cape Clear28. Then, in May 2002, Cape Clear appointed Robert M Walmsley, previously VP International Sales during three years at IONA, as President and Chief Operating Officer29. This wave of defections to Cape Clear

40

eventually appears to have resulted in some hostility between the two companies. IONA launched an (ultimately unsuccessful) lawsuit in early 2002 to prevent the appointment of its former sales director by Cape Clear, prompting the resignation of Annrai O’Toole from his board position at IONA30.

Reconfiguration of staff in new firms

Pinch and Henry (1999) also suggest that knowledge can be disseminated within a cluster when experienced personnel are reconfigured in new firms. The case of Headway Software presents an excellent example of this process operating within the IISI (refer Figure 7)31. Headway, a software development tools supplier with unique visualisation technology, was founded in 1999. It has a development centre in Waterford, European Sales Office in Dublin, and US HQ in Boston, MA. In addition to its directors it has 15 employees, including 10 software engineers. It launched its first product (Headway reView) in May 2001 and its second (Headway reOrganize) in September 2001. Sales are on a license model basis. Headway had raised funding of over US$1.5 million from Enterprise Ireland, Delta Partners, key executives and private investors by May 2001 (company valued at US$5m) and has long-term aspirations of a public floatation. The firm won the Irish Software Association’s Technical Innovation Award in 2001.

A closer look at the composition of Headway’s board of directors illustrates how a group of individuals with different experiences within the software industry, including some outside Ireland, can be reconfigured to form of a new start-up company with a betterthan-average chance of success. Headway’s co-founders were Chris Chedgey (the firm’s Chief Technical Officer) and Brendan O’Reilly (the CEO). Chedgey had 18 years previous software engineering experience in Ireland, Canada and the USA and had worked on various complex software development projects, including work for NASA. (Chedgey may be an Irish national and returning Irish emigrant). O’Reilly was a cofounder of Sybase Ireland in 1990 and established the services business of Sybase under the name Aras Consulting. Previously he had five years in sales with Hewlett-Packard. These founders have assembled an experienced board of directors, with complementary

41

expertise, in order to achieve their objective of rapid growth and success. Tom Sullivan (VP of Marketing) was previously Technical Director Asia-Pacific for Cyrano Inc. He had also been an IT manager and Architect in Prudential Assurance and a pre-sales consultant in Sybase. Paul Hickey (VP of Strategic Alliances and Manager of the Boston office) had 15 years experience as a software engineer for Generics (Robert Cochran’s company), Lotus Development and IONA (6 years experience including a period as VP US operations).

As a recipient of venture funding, Headway also has a representative of the major investor, Delta Partners, on its board (Shay Garvey). Finally, Headway has also appointed two very experienced non-executive directors: Jim Mountjoy, who we have discussed above as the founder of Euristix and a former MD of Baltimore, and Colin Newman, who was an IONA co-founder and, as noted above, is the current VP of Marketing at IONA spin-off CapeClear. Thus the case of Headway Software shows how start-ups in the sustained development phase of the IISI are advantaged by their ability to draw on an increasingly experienced localised pool of managerial talent (a clear external effect of agglomeration). By tapping into the local ‘knowledge pool’ this new generation of startups is better prepared for success than were their predecessors.

Formal and informal networking by managers and professionals

Another process of collective learning identified by Keeble and Nachum (2001) concerns “knowledge exchange and development through informal and formal networking, collaboration and personal interaction by professionals and managers of cluster firms” (p.10). Similarly, Pinch and Henry (1999) have shown the role of informal collaboration, extensive personal contact networks of people working in the industry and gossip (in workplace and social settings) in disseminating knowledge through a cluster of knowledge-intensive industry. Methodologically, it is difficult to generate evidence to support the operation of these processes within the IISI from secondary data sources. The discussion here is therefore limited to a summary of some of the findings of research by Ó Riain on this issue.

42

On the subject of informal networking and information exchange within the IISI, Ó Riain (2000, p.19) notes “informal communication, structured by customers/users and technologies, is part of the taken for granted structure of everyday life in the industry”. He also notes that “in many cases managers are barely conscious of the extent to which developers in their firm constantly share information with developers in other firms (Ó Riain, 2000, p.19). Ó Riain (2000, p.20) also argues that informal networking has played an important role in the dissemination of management/business know-how within the cluster, especially for smaller firms who commonly “shared information about managerial and business issues with other firms - usually in the form of information about how ‘the system’ works and in the form of ‘war stories’”.

Informal networking within the IISI is clearly made possible by the relatively small size of the industry and its geographical concentration in and around Dublin. However, Ó Riain also argues that networking with the IISI has particular social and cultural underpinnings. For example, he notes that employees in different firms often “know each other from their county of origin and local neighbourhood (particularly from rural and urban middle class neighbourhoods), from schools and colleges and from inter-firm career patterns” (Ó Riain, 1997, p.31). Of particular importance, it is argued, is the highly class-stratified educational and social system in Ireland, especially in Dublin, which means “residential, educational and work networks often overlap so that developers are likely to have multiple weak ties to others in the industry” (Ó Riain, 2000, p.21).

Demonstration effects

The final mode of collective learning considered here concerns ‘demonstration effects’, which can be viewed as a kind of indirect knowledge spillover that results from geographical clustering. Demonstration effects within the IISI relate to both firms and key individuals. In the case of Israel, Teubal et al (2000, p.50) have argued that the leading data security software firms (e.g. Checkpoint, Aladdin) made an enormous contribution to other companies in the Israeli high-tech cluster “through the reputation

43

effects they generated and the business models which they offered to these companies”. Similarly, as Arora et al (2001, p.18) have observed, successful Irish software firms like Baltimore and IONA have acted as ‘role models’ and created a ‘follow-the-leader effect’ within the IISI. For example, by making it to the Nasdaq the leading indigenous firms have demonstrated a successful business model to the next generation of indigenous software start-ups, and made it easier for other Irish firms to follow in their footsteps. Second, key individuals such as Chris Horn, Jim Mountjoy and Fran Rooney have acted as ‘role models’ to the next generation of Irish entrepreneurs, demonstrating that Irish entrepreneurs can succeed in the global software industry. The relatively ‘small world’ of the Irish industry, with its geographical concentration in greater Dublin, has meant that the word has spread quickly, through both keynote speeches given to industry events and though word of mouth. The potential significance of demonstration effects in a country that is a relative newcomer to the world of high-tech entrepreneurship cannot be overstated.

CONCLUSION

Summary of findings

This paper has sought to explain the development of a cluster of high growth, knowledgeintensive businesses (i.e. software firms) in what has traditionally been regarded as a less favoured region (i.e. the Republic of Ireland). I have argued that the Irish indigenous software phenomenon is the product of a complex range of factors and processes that can be best interpreted using a two-phase model of cluster development. The origins and initial establishment of the cluster can be attributed to the presence in Ireland of some favourable factor conditions (notably an abundance of appropriate skilled labour) and to the firm and market-building efforts of pioneering entrepreneurs, who employed a range of astute company development strategies. The Irish State played an important role both by creating some of the key factor conditions and by supporting the firm and marketbuilding efforts of pioneering entrepreneurs with various direct and indirect measures.

44

More recently, the IISI appears to have entered what might be called the ‘sustained development phase’. In this phase, I have argued, the development of the industry has been sustained by the operation of various agglomeration economies and localised collective learning (or knowledge dissemination) processes.

Wider lessons

By way of conclusion, I attempt to draw out a number of wider lessons from the case study of the IISI, including lessons for policy makers in other regions. I make four main points relating to (1) the role of the State in cluster development, particularly for knowledge-intensive industries; (2) the relationship between inward investment and indigenous development; (3) the role of international or global ties in cluster development; and (4) the origins of entrepreneurial start-ups.

First, this paper has shown the important role played by the Irish State and EU in the development of the IISI. For example, the State laid the foundations for the industry by expanding third-level education in technical subjects (including computer science) and by upgrading the telecoms infrastructure to international standards (although there is some irony that original motivation was to attract FDI rather than stimulate indigenous development). The State also provided a range of direct (financial and non-financial) and indirect assistance to support the firm and market-building efforts of entrepreneurs and establish various supporting institutions. There is surely a lesson for policy makers in other regions that the State can play an important role in supporting indigenous economic development, particularly in so-called less favoured regions. One important point to remember here, though, is that, aside for targeting software and providing some strategic guidance, at no stage did the Irish development agencies attempt to promote specific products or technologies or interfere with company decision-making.

The second point of wider relevance concerns the relationship between inward investment and indigenous development. Upon initial inspection, the parallel of emergence of the overseas and indigenous segments of the Irish software industry may

45

imply some causal link from inward investment to indigenous development. However, the available empirical evidence suggests a note of caution regarding this interpretation. A primary consideration is the lack of any significant direct trading relationships between the leading indigenous software firms and software MNCs operating in Ireland. The majority of leading indigenous software firms derive most of their revenues from exporting software products, rather than entering into sub-contracting relationships with MNCs in Ireland, hence Ó Riain’s (1997, p.27) assertion that the IISI “emerged largely from a set of local dynamics”. It is also the case that software MNCs in Ireland have been relatively unimportant as a source of new firm spin-offs (Ó Riain, 1997). The one area where the presence of software MNCs in Ireland does seem to have been a significant factor is though indirect spillovers relating to the labour market. Thus, O’Malley and O’Gorman (2001) argue that software MNCs “helped to develop labour skills which have proved useful in the indigenous software industry” (p.315). MNCs are also said to have acted as a training ground for founding entrepreneurs of indigenous software firms “two thirds of them had gained experience working in a foreign-owned MNC in Ireland at some stage in their careers” (O’Malley and O’Gorman, 1997, p.316). This evidence should moderate the expectations of policy makers in relation to the anticipated knock-on effects of attracting inward investment in sector such as software.

The third point of wider relevance from the case of the IISI concerns the role of international or global ties in the development of indigenous clusters of KII. My arguments in this paper are largely supportive of Ó Riain’s (1997, p.27) assertion that the IISI “emerged largely from a set of local dynamics”. For example, I have emphasised the role of firm-building by Irish entrepreneurs, various State supports to the industry and latterly various agglomeration economies and localised collective learning/knowledge dissemination processes. However, I feel it is important to re-state the important role played by various international or global ties in the development of the IISI. The most obvious point here is the overriding orientation of the leading indigenous firms to international markets, notably the US. The rise of the IISI has been to a great extent contingent on the surging global demand for software, driven by the US market. Other important international dimensions to the development of the IISI include the increasing

46

involvement of international venture capitalists in Ireland and the floatation of leading Irish firms on international stock exchanges, including the Nasdaq, the participation of indigenous firms in international strategic alliances, and the possible role of return migration of skilled/experienced Irish emigrants. This evidence confirms the importance of being aware of the wider global context, as well as local dynamics, when studying industry clusters. In terms of policy, one lesson is that development agencies should encourage firms to be outward looking from the outset, possibly assisting them in exporting and forming international alliances for example.

A fourth and final point of wider relevance concerns the origins of entrepreneurial startups. In the case of the IISI, the majority of early software product firms originated as spin-offs from larger firms in other sectors, as start-ups targeting the emerging local demand for IT services and spin-offs from academic research. More recently, the majority of firms have started as spin-offs from established indigenous software firms and also as academic spin-offs. One implication is that high-tech entrepreneurship is cumulative, with an established cluster of KII naturally generating new spin-offs, which help to grow the cluster. In terms of policy, the evidence suggests the importance of adequately funding university research and other forms of R&D and promoting and assisting its commercialisation. Another important option for policy-makers in the case of software is implementing initiatives that encourage and assist services firms to convert their service activities into marketable products (which tend to yield higher returns than service work).

Issues for further research

Finally, to end the paper, I will outline some areas/issues for further research, both in the specific case of the IISI and for the study of cluster development in knowledge-intensive industry (KII) more generally:

1. Pinch and Henry (1999) were pioneers in illustrating the potential value of a focus on knowledge dissemination and collective learning as an explanation for clustering in

47

KIIs. In this study I have attempted to apply these ideas to a different empirical example, albeit in an exploratory way. I agree that the approach seems to have considerable utility and suggest there is a need for further empirical research on the operation of collective learning processes, both in the context of the IISI and elsewhere.

2. Bresnahan et al (2001) have argued that firm and market-building efforts by cluster pioneers have played a crucial role in the establishment of ICT clusters in various location around the world. Teubal et al (2000) have also shown the utility of studying company development strategies (in the case of the Israeli data security software industry). The case of the IISI also shows the importance of examining the strategies of the leading firms in a successful cluster. Hence there is a case for further research on the company development strategies employed by entrepreneurial start-ups. In the specific case of the IISI there is a need for further attention to the role played by international strategic alliances and the IPO process in the development of leading companies.

3. Bresnahan et al (2001) suggest that emigration and subsequent return migration of managerial and technical personnel may have played an important role in the development of ICT clusters in India, Israel and Taiwan. Saxenian (1999) refers to this process as one of ‘brain circulation’ (as opposed to the traditional ‘brain drain’). There are suggestions that this process has also played a role in the development of the IISI but the existing evidence is rather weak and indirect. There is therefore a need for further research on the process of brain circulation and its role in cluster development.

4. Finally, the new evidence presented in this paper on collective learning/knowledge dissemination processes within the IISI has highlighted the potentially significant role of experienced non-executive directors in fast-growing entrepreneurial start-ups. There is a strong case for further research on the role played by non-execs, including experienced industry pioneers and representatives of venture capital investors.

48

Acknowledgements - This paper report was produced as part of NIERC’s core research programme on ‘Competitiveness in Tradeable Services’ (see www.qub.ac.uk/nierc). The Department of Finance and Personnel (DFP) and the Department of Enterprise, Trade and Investment (DETI) fund the programme. Additional support was received from the Northern Ireland ICT Directorate. Unless otherwise stated the report expresses the views of the author. Chris Allen provided valuable research assistance.

49

REFERENCES

Amin A, Thrift N (1992) Neo-Marshallian nodes in global networks, International Journal of Urban and Regional Research, Vol.16, pp.571-587. Arora A, Gambardella A, Torrisi S (2001) In the Footsteps of Silicon Valley? Indian and Irish Software in the International Division of Labour. Stanford Institute for Economic Policy Research (SIEPR) Discussion Paper No. 00-41, Stanford University, California. Barr A, Tessler S (1999) Software Entrepreneurship in Korea (Final Report). Stanford Computer Industry Project (published on-line at: www.aldo.com/papers/). Bresnahan T, Gambardella A, Saxenian A, Wallsten S (2001) ‘Old Economy’ Inputs for ‘New Economy’ Outcomes: Cluster Formation in the New Silicon Valleys. Stanford Institute for Economic Policy Research (SIEPR) Discussion Paper No. 00-43, Stanford University, California. Bresnahan T, Gambardella A, Saxenian A (Eds) (2002) Silicon Valley and its Imitators. McGraw Hill. Camagni R (1991) Local ‘milieu’, uncertainty and innovation networks: towards a new dynamic theory of economic space in Camagni R. ed Innovation networks: spatial perspectives, Belhaven Press, London, pp.121-142. Camagni R (1995) The concept of innovative milieu and its relevance for public policies in European lagging regions, Papers in Regional Science, 74, pp. 317–340. Coe N M (1997) US transnationals and the Irish software industry: assessing the nature, quality and stability of a new wave of foreign direct investment, European Urban and Regional Studies, Vol.4, No.3, 211- 230. Coe N M (1999) Emulating the Celtic tiger? A comparison of the software industries of Singapore and Ireland, Singapore Journal of Tropical Geography, Vol.20, No.1, 36-55. Cooke P, Davies C, Wilson R (2001) Innovation Advantages of Cities: From Knowledge to Equity in Five Basic Steps. Paper presented to ESRC Cities Competitiveness and Cohesion Programme International Research Workshop on “Innovation and Competitive Cities in the Global Economy”, Worcester College, Oxford University, 28–30 March. Crone M (2001) Software and Computer Services Employment in the UK Regions and Ireland, 1995-99: A Review of the New Annual Business Inquiry Data for Great Britain and Equivalent Data for Northern Ireland and Republic of Ireland. NIERC Report Series No.19, NIERC, 46-48 University Road, Belfast, BT7 1NJ. Crone M (2002) A Profile of the Irish Software Industry. Consultation draft published online at: www.qub.ac.uk/nierc. De Fontenay C, Carmel E (forthcoming) Israel’s Silicon Wadi: The Forces Behind Cluster Formation, in Bresnahan T, Gambardella A and Saxenian A (Eds) Silicon Valley and its Imitators. McGraw Hill. (Expected to appear in 2002).

50

Enterprise Ireland (2000) 2000 Report: Seed and Venture Capital Measure of the Operational Programme, 1994-99. Enterprise Ireland, Glasnevin, Dublin 9. Expert Group on Future Skills Needs (1999) Ireland’s Information Technology Skill Requirements. Forfas, Dublin. Gordon I R, McCann P (2000) Industrial clusters: complexes, agglomeration and/or social networks? Urban Studies, Vol.37, No.3, 513-532. Grimes S (1999) The Information Society in Ireland – a Late Developer Catches the Wave. Paper presented to European Regional Science Association Conference, University College Dublin, 23-27 August, 1999. Heeks R (1996) India’s Software Industry: State Policy, Liberalisation and Industrial Development. Sage Publications, London. Hellman T (2000) Venture Capitalists: the Coaches of Silicon Valley. Unpublished manuscript, Graduate School of Business, Stanford University (obtained via web-site). Hellman T, Puri M (2002) Venture Capital and the Professionalisation of Start-up Firms: Empirical Evidence, The Journal of Finance, Vol.LVII, No.1, pp.169-197. Henry N, Pinch S (2000) Spatialising knowledge: placing the knowledge community of Motor Sport Valley, Geoforum, Vol.31, 191-208. HotOrigin Ltd (2001) Ireland’s Emerging Software Cluster: A Hothouse of Future Stars (Report Overview). HotOrigin Ltd, 64 Lower Mount Street, Dublin 2, Ireland (www.hotorigin.com). HotOrigin Ltd (2002) Ireland’s Software Cluster: Innovation – the Fuel for International Success. HotOrigin Ltd, 64 Lower Mount Street, Dublin 2, Ireland. Keeble D, Nachum L (2001) Why do business service firms cluster? Small consultancies, clustering and decentralisation in London and Southern England. Working Paper No.194, ESRC Centre for Business Research, University of Cambridge. Krugman P (1991) Strategic Trade Policy and the New International Economics. MIT Press, Cambridge, MA. Krugman P (1991) Geography and Trade. MIT Press, Cambridge, MA. Marshall A (1890) Principles of Economics. Macmillan, London. McIver Consulting (1998) Manpower, Education and Training Study of the Software Sector. FAS, Dublin. Muir A (2001) Broadband Initiatives (Presentation of Interim Report). Commissioned study for ScotlandIS by Mason Communications Ltd, 5 Exchange Quay, Manchester, England, M5 3EF (published on line at: www.scotsoft.org.uk/projects/index.html). National Software Directorate (1992) The Software Industry in Ireland: A Strategic Review. National Software Directorate, Dublin, Ireland. National Software Directorate (1997) Ireland as an Offshore Software Location. National Software Directorate, Dublin, Ireland.

51

Ó Riain S (1997) An offshore Silicon Valley? The emerging Irish software industry, Competition and Change, Vol.2, 175-212. Ó Riain S (1998) A Tale of Two Globalisations: the Irish Software Industry and the Global Economy. Working Paper No. 101, ESRI, Dublin. Ó Riain S (1999) Remaking the Developmental State: The Irish Software Industry in the Global Economy. Unpublished dissertation, Department of Sociology, University of California, Berkeley. Ó Riain S (2000a) The flexible developmental state: globalization, information technology and the ‘Celtic Tiger’, Politics and Society, Vol.28, No.2, pp.157-193. Ó Riain S (2000b) Mobilizing the Region: The Growth of the Indigenous Irish Software Industry. Unpublished manuscript. O’Gorman C, O’Malley E, Mooney J (1997) The Irish Indigenous Software Industry - An Application Of Porter’s Cluster Analysis. Research Series Paper No 3, National Economic and Social Council, Dublin. O’Malley E, O’Gorman C (2001) Competitive advantage in the Irish indigenous software industry and the role of inward foreign direct investment, European Planning Studies, Vol.9, No.3, 303-321. O’Neill E (2001) From Campus to Nasdaq. Paper presented to conference ‘Stimulating Technology Commercialisation’, Spires Centre, Belfast, 8 March (Author is Director of Innovation Services at Trinity College Dublin). Pinch S, Henry N (1999) Paul Krugman’s geographical economics, industrial clustering and the British motor sport industry, Regional Studies, Vol.33, No.9, 815-827. PriceWaterhouseCoopers (1999) Strategic Development of Internationally Traded Service Industries Throughout Ireland - Sector Report on ‘Software’. Consultancy report prepared for Enterprise Ireland. Roper S, Frenkel A (2000) Different paths to success? The electronics industry in Israel and Ireland, Environment and Planning C, Vol.18 , pp.651-665. Saxenian A (1994) Regional Advantage: Culture and Competition in Silicon Valley and Route 128. Harvard University Press, Cambridge, MA. Saxenian A (1999) The Silicon Valley-Hsinchu Connection: Technical Communities and Industrial Upgrading. Stanford Institute for Economic Policy Research Discussion Paper No. 99-10. (Published on-line: http://www.sims.berkeley.edu/~anno/writings/). Saxenian A (2002) Silicon Valley’s New Immigrant High-Growth Entrepreneurs, Economic Development Quarterly, Vol.16, No.1, February, 20-31. Sheridan C, Sterne J (1999) The Geography of Irish Software. National Software Directorate, Dublin (Published on-line at: www.nsd.ie/news/98_f/270199F.html). Storper M (1995) The resurgence of regional economies, ten years later: the region as a nexus of untraded interdependencies, European Urban and Regional Studies, Vol.2, No.3, 191-221.

52

Swann P, Prevezer M (1996) A comparison of the dynamics of industrial clustering in computing and biotechnology, Research Policy, Vol.25, No.7, 1139-1157. Teubal M, Avnimelech G, Gayego A (2000) Globalization and Firm Dynamics in the Israeli Software Industry: A Case Study of Data Security. Unpublished manuscript (available on-line at: http://atar.mscc.huji.ac.il/~economics/facultye/teubal/papers.html).

53

FIGURES AND TABLES

54

Table 1: National Software Directorate Statistics on Irish Software Industry in 1991-99 1991

1993

1995

1997

1998

1999*

CAGR %91-95

CAGR %95-99

CAGR %91-99

3,801 NA 3,992 7,793

4,495 NA 4,448 8,943

5,773 NA 6,011 11,784

9,200 NA 9,100 17,300

9,250 1,730 10,650 21,630

11,100 2,076 11,715 24,891

10.04%

24.00%

16.81%

10.77% 10.89%

18.15% 20.56%

14.40% 15.62%

291 NA 74 365

336 NA 81 417

390 NA 93 483

561 NA 108 669

630 11 120 761

690 12 120 822

7.60%

15.83%

11.64%

5.88% 7.25%

6.58% 14.22%

6.23% 10.68%

Revenue (IR£m) Indigenous Takeover Foreign Total Indigenous as a % of total

150 NA 1,580 1,730 8.7

236 NA 1,756 1,992 11.9

386 NA 2,611 2,997 12.9

528 NA 3,933 4,461 10.5

719 145 4,346 5,210 16.6

1,006 174 4,824 6,004 19.7

26.66%

32.23%

29.41%

13.38% 14.73%

16.59% 18.97%

14.97% 16.83%

Exports (IR£m) Indigenous Takeover Foreign Total Indigenous as a % of total

61 NA 1,548 1,609 3.8

116 NA 1,726 1,842 6.3

226 NA 2,585 2,811 8.0

365 NA 3,854 4,219 8.7

446 126 4,042 4,614 12.4

624 145 4,365 5,134 15.0

38.74%

35.82%

37.27%

13.68% 14.97%

13.99% 16.25%

13.84% 15.61%

Employment Indigenous Takeover Foreign Total Number of companies Indigenous Takeover Foreign Total

Notes: 1991 data from NSD (1992); 1993, 1995 and 1997 data downloaded from NSD web-site in July 1999; 1998 and 1999 data supplied by NSD to author in January 2001; * 1999 figures are NSD estimates based on a number of indicators, mainly previous growth rates; ‘Takeover’ companies are firms that were originally established in Ireland and owned by Irish nationals but were subsequently acquired by a foreign entity – this category was not separately distinguished from the ‘indigenous’ category prior to 1998; percentage increases are compound annual growth rates.

55

Table 2: Key Ratios for Irish Software Industry in 1991-99 1991

1993

1995

1997

1998

1999

Exports as % of revenue Indigenous Takeover Foreign Total

40.7 NA 98.0 93.0

49.2 NA 98.3 92.5

58.5 NA 99.0 93.8

69.1 NA 98.0 94.6

62.0 86.9 93.0 88.6

62.0 83.3 90.5 85.5

Employees per company Indigenous Takeover Foreign Total

13.1 NA 53.9 21.4

13.4 NA 54.9 21.4

14.8 NA 64.6 24.4

16.4 NA 84.3 25.9

14.7 157.3 88.8 28.4

16.1 173.0 97.6 30.3

Revenue (IR£m) per employee Indigenous Takeover Foreign Total

39,463 NA 395,792 221,994

52,503 NA 394,784 222,744

66,863 NA 434,370 254,328

57,391 NA 432,198 257,861

77,730 83,815 408,075 240,869

90,631 83,815 411,780 241,212

Revenue (IR£m) per company Indigenous Takeover Foreign Total

515,464 NA 21,351,351 4,739,726

702,381 NA 21,679,012 4,776,978

989,744 NA 28,075,269 6,204,969

941,176 NA 36,416,667 6,668,161

1,141,270 13,181,818 36,216,667 6,846,255

1,457,971 14,500,000 40,200,000 7,304,136

Note: ratios calculated by author from NSD statistics

56

Table 3: Leading Irish Indigenous Software Firms Company Name Baltimore Technologies # IONA Technologies # Horizon Technology Group # SmartForce # Datalex Communications # Parthus # Eurologic Systems VISION Consulting IT Alliance FINEOS CR2 Trintech Group # Eontec Norkom Technologies Educational Multimedia Group Performix Technologies CardBASE Technologies Intuition Publishing Riverdeep Interactive Learning # WBT Systems

Expertise Secure systems for e-business and mobile commerce Middleware, systems software, webserver software Holds a range of consulting and ASP companies Computer-based training, educational software IT solutions to the airline and travel industry. Semiconductor design software Communications/data security software e-business and technology consultants Outsourcing, testing, consulting and support services Component-based solutions for life and pensions industry. Internet banking, channel management, card payment solutions Secure payment software products for e-commerce Multi-channel banking software solutions Intelligent customer interaction software (i.e. eCRM) Computer-based training/e-learning software Contact centre solutions Smartcard, Internet payment and e-commerce solutions Web-based banking and financial markets training software Interactive learning solutions Web-based training and server software developers

Established 1976 1991 ? 1983 1985 1993 1988 1984 1997 1993 1996 ? 1994 1998 1998 ? ? ? ? ?

Employment 1200 800 650 500 462 370 360 350 280 270 250 225 200 200 170 145 135 100 100 100

Notes: companies marked # are publicly listed; ‘Established’ is the year of establishment; ‘Employment’ is the highest noted employment figure during 20002001 (i.e. before the downturn) and should be seen as indicative only; employment figures are for the whole company not just the Irish operations. Source: Compiled by author using data from National Software Directorate web-site company directory (www.nsd.ie), Kompass.ie and TechWatch.ie

57

Table 4: Employment in Top 20 Overseas Software Firms by Activity Activity

Number of firms 5 5 6 2 2 20

Localisation, manufacturing and distribution Software supporting subcontractor Software development centre Acquired former indigenous companies Other (e.g. consultancy) Top 20

Employees

% of Top 20

3,940 2,770 2,790 980 620 11,100

35.5 25.0 25.1 8.8 5.6 100.0

Source: analysis of company database compiled by author from overseas software company list on IDA Ireland web-site (www.ida.ie/docs/issw.html) and other sources including Kompass.ie.

Table 5: Indigenous Software Industry - Employment by Company Size in 1998 Company size (by employment) 100+ 50–100 25–49 11–24 1-10 Unknown Total

No. of companies

% of Companies

Employment

10 24 58 112 316 110 630

1.9% 4.6% 11.2% 21.5% 60.8% 100%

2,170 1,814 2,057 1,769 1,445 9,255

Source: supplied by NSD to author in January 2001.

58

% of Total Industry Employment 23% 20% 22% 19% 16% 100%

Table 6: Indigenous Software Industry - Revenue and Exports by Company Size in 1998 Company size by employment 100+ 50–100 25–49 11–24 1-10 Total

Revenue (IR£m) 263 143 137 99 77 719

Share of revenues 36% 20% 19% 14% 11% 100%

Revenue/ employee 121,900 79,000 67,000 56,000 53,000

Exports (IR£m) 213 90 75 40 28 446

Share of exports 48% 20% 17% 9% 6% 100%

Share of revenue from exports 81% 63% 55% 40% 36% 62%

Source: supplied by NSD to author in January 2001.

Table 7: Specialisation of Indigenous Software Firms in 1999

Bespoke software (customised development services) Software products (i.e. applications) Internet-related services (on-line services, Internet-access services, development of portals) Multimedia services (e.g. outsourced video, desktop publishing, CBT) Other services (professional services including customer support) Localization System software/tools (e.g. CASE tools, C++ or Java compilers)

N 300 227 135 77 76 70 54

Note: firms can be included under more than one category (total population of approx. 515 firms). Source: Arora et al (2001) - elaborations on the NSD database and company web-sites information.

59

% 58% 44% 26% 15% 15% 14% 10%

Table 8: Export Markets of Indigenous Software Companies by Geographic Region UK Western Europe North America Far East Pacific Rim Other

21% 28% 43% 1% 3% 4% 100%

Source: data downloaded from NSD web-site in July 1999. Note: reference year not specified but probably 1997.

Table 9: Geography of Leading Indigenous Software Firms in 2001 Location Dublin Cork Limerick/Shannon Galway Other Unknown Total

Companies n % 96 76.2% 7 5.6% 7 5.6% 8 6.3% 4 3.2% 4 3.2% 126 100.0%

Employment N % 8,424 87.0% 469 4.8% 275 2.8% 226 2.3% 134 1.4% 140 1.4% 9,688 100.0%

Source: analysis of list of Top 126 indigenous firms compiled by author.

60

Average size 87.8 67.0 39.3 28.3 35.0 33.5 76.9

Table 10: Examples of Collective Learning/Knowledge Dissemination Processes Keeble and Nachum (2001) Flows of professionals and “embodied expertise” through the local labour market High rates of localised entrepreneurship and spinoff of new firms from existing businesses Informal and formal networking, collaboration and personal interaction by professionals and managers of cluster firms N/A

Pinch and Henry (1999) High turnover of personnel, including key designers and engineers Firm failure/new firm formation leading to reconfiguration of staff in new firms Spin-offs from existing firms Informal collaboration, observation and gossip (in workplace and social settings) Extensive personal contact networks of people in the industry ‘Information leakage’ though links between firms and their suppliers

Source: summary of Keeble and Nachum (2001, p.10) and Pinch and Henry (1999, p.824-5).

61

Table 11: Selected VC Investments in Emerging Irish Indigenous Software Firms, 1999-2001 Date

Recipient

Investor(s)

Amount

30/03/01 25/02/00 14/04/00 10/03/00 02/06/00 07/04/00 03/11/00

Am Beo Buytel Cape Clear cardBase Celtech Software CR2 CR2

IR£3M IR£2M US$2M US£8M US$5M US$10.5M US$8M

23/10/01 28/04/00 14/11/01 18/08/00 17/11/00 15/09/00

CR2 Eontec Eontec eware JLS Software Macalla

10/12/99 20/10/00 01/12/00 27/08/99 21/01/00 17/03/00 07/07/00 21/01/00 03/08/01 10/12/99 04/05/01 24/11/00 07/09/01

Massana Massana Network 365 Norkom Technologies Norkom Technologies Openet Telecom Piercom Qumus Rococo Software Sepro Softco Softech Vordel

ACT VC Trinity VC ACT VC ICC, Enterprise Ireland, private investors ICC VC Alpinvest and ACT VC GIMV, NIB Capital and ACT VC, Goodbody Stockbrokers and AIB Corporate Banking. GIMV, NIB Capital, ACT VC ICC Software Partners Warburg Pincus, ICC VC 3i Group and ICC VC Trinity VC Reuters Greenhouse Fund, 3i Group and Guinness Ulster Bank Equity Fund. ACT VC Dresdner Kleinwort Benson Amadeus Capital Partners, Trinity VC Trinity VC Trinity VC XATF ACT VC ACT VC Trinity VC Agincourt Dolmen Corporate Finance ACT VC Elderstreet, Dresdner Kleinwort Wasserstein, ICC, Intel 64 Fund and Powerscourt 3i Group Delta Venture Capital of Ireland and Vertex Venture Capital of Singapore ADD Partners, Vertex Manangement, Delta Partners and Enterprise Ireland

18/08/00 Wolfe Group 23/09/99 XIAM 27/04/01 XIAM

US$7.5M US$10.4M US$25M US$10M US$5M €8.6M US$7M US$16.5M IR£15M US$2M US$3M IR£2M IR£2.5M US$3M IR£2.5M IR£3.3M €4M IR£11.5M US$10M IR£5M IR£2M €6.75M

Source: author’s analysis of news stories from TIU TechWatch news archive (www.techwatch.ie)

62

Table 12: Summary of VC Investments Made by Funds Established Under the EU Seed and Venture Capital Measure

Source: Enterprise Ireland (2000, p.5)

63

Figure 1: Explanatory Framework (Development of Knowledge-intensive Industry Clusters)

Two Phase Explanatory Framework Sustaining Cluster Development

Origins & Initial Establishment


Window of opportunity




Favourable factor conditions




Spark of entrepreneurship




Firm-building and marketbuilding strategies




Supporting role of State and semi-State institutions




Improved environment for entrepreneurship/firm-building...




Agglomeration economies:  Local pool of skilled labour  Specialised supply industries (e.g. VC, support services)




Collective learning/knowledge dissemination processes:  Movement of key individuals through the cluster  Spin-offs/new firm formation  Formal/informal networking

Source: devised by author based on review of relevant literature.

64

Figure 2: Simplified Career History for Dr Chris Horn

Dr Chris Horn Sept 2000: Appointed chairman of ITEQ advisory committee

Chaired on Future Skills’ Prominent role in industry affairs

1984-89: EU Esprit research Comp Sci Trinity CD

1991: Co-founder of IONA Technologies

May 2000: Resigns as IONA CEO (stays on as Chairman)

1991-2000: Chairman and CEO of IONA Technologies

1997: IONA floats on Nasdaq

Sept 2001: appointed nonexec. director by Sepro

Nov 2001: appointed nonexec. director by CR2

Source: compiled by author using various secondary data sources.

65

Figure 3: Simplified Career History for Dr Jim Mountjoy

Sept 2000: appointed nonexec. director of Duolog Technologies

Dr Jim Mountjoy

July 2001: appointed nonexec. director of Aspect Software

Key work experience: Background: Ph.D. in Telecoms Engineering

Dept of Posts & Telegraphs National Board for Science & Technology

1990-99: founds Euristix and builds it to 150 e’ees

1984-90: Managing Dir of Baltimore Technologies

1999: Euristix sold to FORE Systems of USA for $80m (Mountjoy leaves the company in July 2000)

Source: compiled by author using various secondary data sources.

66

Non-executive director of Corvil Networks Current chairman of Headway Software Current chairman of Tsunami Photonics

Figure 4: Simplified Career History for Fran Rooney

Fran Rooney appointed Sept 2000: < fill > non-exec. director of Norkom Technologies

1990s: CEO of Baltimore Technologies Oversees growth to 700+ e’ees & annual revenues of US$40m

July 2001: Resigns as Baltimore CEO after collapse in share price

Sept 2001: appointed Chairman of Addoceo

October 2001: appointed nonexec. director by Epionet

October 2001: appointed nonexec. director by WestGlobal

1999: Baltimore merges with Zergo Holdings (UK) then floats on Nasdaq

Source: compiled by author using various secondary data sources.

67

Figure 5: Simplified Career History for Barry Murphy

Barry Murphy

Early 1970s: University College Cork B.E. Electrical Engineering & M.Eng.Sc.

1980s: Managing Director of Insight (leading Irish software co)

1988-96: First head of Ireland’s National Software Directorate

1996-99: CEO of Cullinane Group Ireland

Involved in writing strategy document for Irish Software Association

1988: Insight acquired by Hoskyns (CapGemini)

Source: compiled by author using various secondary data sources.

68

Since 1996: CEO of Openet Telecom

Figure 6: Simplified Career History for Robert Cochran

Robert Cochran Advisor to Irish govt & European Commission < fill > Board of IEEE Software & Irish Software Assoc. Member of National Board for Science & Technology

1980s Managed his own software company Generics

1991-present: Director of Centre for Software Engineering at Univ. Coll. Dublin Established NSD as the national focal point for best practice in software development

Consultant for his own development tools company Catalyst Software

Source: compiled by author using various secondary data sources.

69

Figure 7: Diagramatic Representation of Antecedents of Headway Software

Reconfiguration of staff in new firms Founder of Euristix and Former MD of Baltimore Tech.

18 years software eng’g experience in Ireland, Canada, USA (incl. NASA)

Cyrano Asia-Pacific, IT man’gr Prudential, consultant in Sybase

Tom Sullivan (VP Marketing) Jim Mountjoy (non-exec)

5 yrs in sales with Hewlett-Packard & established Sybase Ireland in 1990s

Experience as an engineer and m’gr with Lotus, IONA

Representative of venture capitalist

Chris Chedgey (Co-founder/CTO)

IONA co-founder, now VP Marketing at CapeClear

Brendan O’Reilly (Co-founder/CEO)

Paul Hickey (VP Strat. All.)

Shay Garvey (Delta Partners)

Colin Newman (non-exec)

Board of Directors

HEADWAY SOFTWARE Key Facts • • • • •

Software development tools supplier with unique visualisation technology Founded in 1999. Board plus 15 employees (including 10 software engineers) Development centre in Waterford; European Sales Office Dublin; US HQ Boston First product launched in May 2001 (sales on license model basis). Over US$1.5m raised from Enterprise Ireland, Delta Partners, key execs, etc

Source: compiled by author using various secondary data sources.

70

Appendix 1: National Software Directorate Definition of Software Industry Includes companies undertaking the following activities:


Development of software products/systems for subsequent sale to end users




Development of systems software or software development tools for sale to vendors




Services directly related to the design and/or development of software systems




Localisation of either own organisations products or third-party products




Development of programmes or systems for subsequent incorporation into hardware




Technical training in the area of systems analysis, design and programming




Provision of ‘hot-site’/disaster recovery facilities

The definition does not include:


Personnel of Irish companies based in overseas offices




Data entry or data processing or bureau services for local or overseas organisations




Training services for standard packaged software




Software development by in-house IT departments




Sale of hardware or third party software packages

Source: O’Gorman et al (1997, p.2)

71

NOTES 1

For a descriptive profile of the overseas software industry in Ireland see Crone (2002).

2

The software industry is a unique sector within Ireland’s ‘Celtic Tiger’ economy in exhibiting a parallel

development of indigenous industry alongside an influx of inward investment (O’Gorman et al, 1997). Among the other leading industrial sectors in the Celtic Tiger economy are ICT hardware manufacturing, pharmaceuticals and medical instruments, which are all dominated by overseas multinationals. 3

Perhaps the highest profile example is Saxenian’s (1994) study of ‘Silicon Valley’ and ‘Route 128’ in the

United States. Within the UK, notable examples include Amin and Thrift (1992) on the City of London’s financial services industry, Keeble and Nachum (2001) on small consultancy firms in London and southeast England, and Pinch and Henry (1999) on ‘Motor Sport Valley’ in central/southern England. 4

When reviewing this data it is important to bear in mind the NSD definition of the software industry (see

Appendix 1). 5

This vendor base consists of both Irish-owned and overseas-owned supplier companies.

6

Table 5 of the Census of Industrial Production 1998 gives total persons engaged in industry as 63,518 in

the Dublin region against a total of 260,357 in Ireland (24.4%), and the number of industrial units as 1,333 in the Dublin region against a total of 4,932 in Ireland (27.0%). 7

Crone (2002) gives a more detailed descriptive account of the overseas software industry.

8

Coe (1997) cites EITO data for 1994 which values the domestic Irish IT market at ECU 763 million out of

an EU total of ECU 129,576 million (i.e. 0.59%). 9

These Top 20 firms account for a sizeable part of the overseas industry.

10

Baltimore Technologies de-listed from the Nasdaq in September 2001 but is still listed on the London

Stock Exchange. Parthus has a dual listing on the Nasdaq and London Stock Exchanges and Trintech has a dual listing on the Nasdaq and Neue Markt. 11

Storper’s (1995) notion of ‘untraded interdependencies’ might also be grouped with this alternative

approach. 12

Pinch and Henry (1999) reach this conclusion in their empirical study of the British motor sport cluster

(‘Motor Sport Valley’) and Keeble and Nachum (2001) do so from their research on the clustering of small consultancy firms in London and South East England. 13

This presents a problem in terms of advising policy-makers in less favoured regions who wish to promote

the development of clusters of knowledge-intensive industry where none currently exist. 14

Teubal et al (2000) employ a similar rationale in their study of the Israeli data security software cluster.

They argue that a focus on the ‘dynamics of growth’ of very successful companies within the cluster (i.e. building growth profiles of the firms and highlighting key elements of company strategy) is an important first step in uncovering the wider story of the development of the Israeli high tech cluster. 15

Bresnahan et al (2001) previews the findings of a forthcoming book ‘Silicon Valley and its Immitators’

(to be published in 2002 by McGraw Hill).

72

16

Total student enrolments in Ireland’s third-level education institutions rose dramatically from 64,137 in

1989/90 to 115,696 in 1999/2000 (Source: Key Education Statistics, Department of Education & Science). 17

Saxenian (1999) has documented the important role played by return migrants (to/from Silicon Valley) in

the development of the electronics industry in Taiwan’s Hsinchu region. Saxenian (2002) highlights the role played by immigrant entrepreneurs (of Chinese and Indian origin) within Silicon Valley. 18

By the late 1990s, however, Ireland was beginning to fall behind the best international standards in

telecoms infrastructure. Whilst Dublin still compares favourably in an international context, this is not the case for Ireland as a whole (Muir, 2001). In particular, there has been a failure to the roll-out high-grade broadband infrastructure to the Irish regions (Grimes, 1999; Muir, 2001). Attempts by the Irish government to address this situation, as part of the National Development Plan for 2000-06, using public-private partnerships model have fallen behind schedule due to a lack of interest from the private sector and financial problems in the global telecoms sector. There are also growing concerns over the high cost of telecoms services in Ireland (HotOrigin, 2001, p.7). Many of these problems have been attributed to the slow pace of telecoms deregulation in Ireland (Grimes, 1999; HotOrigin, 2001; Muir, 2001). 19

These pressures were illustrated most graphically in 2001 and 2002 when Baltimore Technologies and

Datalex decided to de-list from the Nasdaq, whilst retaining a public listing in Europe. 20

Source: IEP Professional Ireland, No.246, April 19, 2000.

21

Source: personal interview with Robert Cochran conducted by the author (25/01/01).

22

The one remaining apparent problem on the issue of availability of finance in Ireland is said to be a lack

of seed-stage investment funding for start-up companies (HotOrigin Ltd, 2001, p.6). 23

Source: IEP Professional Ireland, No.346, May 1, 2002.

24

Source: TIU TechWatch news archive (various dates) unless otherwise stated.

25

IONA Technologies has produced numerous other spin-offs including Prediction Dynamics, Rococo

Software, TSL, Zenark, Wilde and Sabadille Systems (TIU TechWatch, 08/05/02) 26

Source: IEP Professional Ireland, No.246, April 19, 2000.

27

Source: Tornado Insider feature article on Irish start-ups (www.tornado-insider.com/irish.asp).

28

Source: TIU TechWatch news archive (article dated 10/11/00).

29

Source: IEP Professional Ireland, No.346, May 1, 2002

30

Source: TIU TechWatch news archive (article dated 08/05/02).

31

Source: TIU TechWatch news archive and company web-site.

73