Bundling is of concern to the ACCC since Telstra took-up bundling as its key marketing concept and included Pay TV in its bundle. Both the Commission and ...
Price Bundling – New Directions John de Ridder Draft at 14 July 2003
1 - Executive summary Bundling is of concern to the ACCC since Telstra took-up bundling as its key marketing concept and included Pay TV in its bundle. Both the Commission and Telstra are still developing their understanding of the benefits and costs of bundling and this paper aims to further that process. In the next section the different forms of price bundling are described together with their benefits to consumers and providers. In particular, the use of bundling as a price discrimination tool to increase profitable sales is explained since this is overlooked in its current use as a simple discounting device. In the third section the legal and regulatory issues are examined. This suggests how demand relationships can give greater certainty as to when bundling is anti-competitive and how imputation tests should be handled in the case of bundling. These suggestions complement the current views and discussion papers published by the Commission in January 20031. The fourth section mentions some implementation issues before going into the fifth section which critiques Telstra’s current strategy and proposes a more growth oriented approach. Discount leakage is the major source of earnings dilution in the current strategy which is also defensive. The proposed approach generates 50% more incremental earnings per dollar of discount without any leakage risk and less market and regulatory risks.
2 - What is bundling? A profitable bundling strategy segments markets for revenue and profit growth based on a detailed understanding of how customers value services differently. The bundled offer is usually priced at less than the sum of the parts. But, it can involve a premium if the value proposition is, say, convenience (eg buy a set of stamps rather than search for each in the set). In the more usual case, the price is less than the sum of the parts but the pricing is more sophisticated than Telstra’s current Options and Rewards discounting as will be shown in two arithmetic examples (see Tables 3 and 4 below). Bundling is a common pricing practice and takes a number of forms. A simple example is the set-price lunch (eg “two courses and a glass of wine for $20”). This is “mixed-joint” bundling where the component services can be obtained separately as well as in the bundle. A variation of this form is “mixed-leader” bundling where a product is packaged with an existing one to attract customers. Coles Myer’s research suggests that bundles like Woolworth’s cross-selling of cut-price petrol increase sales by 1.5 to 2 percentage points2. This form is also used to introduce a new product. It is more cost effective to bestow a new service on top of an existing one (eg some EasyCall features) as an introductory offer than to promote it on its own merits. An extreme form of bundling is “pure price bundling” in which components are not available separately. FOXTEL’s “no slice and dice” position on reselling its content was a prime example (see section 3.3 below).
1
See papers at http://www.accc.gov.au/telco/fs-telecom.htm
2
SMH, 12 Dec. 2002
John de Ridder
1
2.1 – What are the benefits? Both customers and providers can benefit from price bundling. Benefits of bundling to the customer include, •
convenience (eg one bill, simplicity, one point of contact)
•
reduced marginal prices (but increased overall spend)
One bill is certainly an attractive value proposition for many customers and should generate cost savings for providers. It is not clear why Telstra can provide FOXTEL on the single bill but not BigPond (except for dial-up service). Adding more services to the bill can create “sticker shock” which is ameliorated by encouraging customers to move from quarterly to monthly billing; which also improves the provider’s cash flow. The most powerful benefit for customers after convenience as perceived by Telstra is that, “(with bundling) customers will churn less in this environment because they’ll get rewards from a financial perspective” 3. However, its Options and Rewards discounts do not change marginal prices significantly – the apparent difference between the list price and incremental service is no more than 10 per cent so it is unlikely to have a large impact on purchases. Telstra reported that after five months 100,000 FOXTEL services were signed-up for Options and Rewards of which 10 per cent were new to Pay TV 4. Some of the remainder may have come from Optus but it is likely that most came from existing FOXTEL customers. Benefits of bundling to the provider include, •
improved profits and revenues (eg cross-selling, cost savings)
•
improved customer retention (eg whole-of-business, deter entry)
2.1a – Churn reduction Telstra sees the main benefit of bundling as reductions in churn. Apart from churn-out there is churn-in and reductions in the levels of both will probably reduce some costs. But the major benefit to Telstra comes from reducing net churn-out as this increases market share (ignoring the distribution of market growth). Below, we consider reducing churn as reducing churn-out with no change in churn-in (ie a one percentage point reduction in churn is a one percentage point of revenue saved, or gained). Table 1 scopes the value of such churn reduction to Telstra.
Table 1: Value of 1% reduction in churn before cost savings Churn
Total
Revenue
Profit
Profit
Profit
Revenue
Impact
Impact-A
Impact-B
Impact-C
(per yr)
($m, 2002)
($m pa)
($m pa)
($m pa)
($m pa)
STD+ID+F2M Calls
24%
2,881
28.8
10.8
-15.1
-43.9
Mobiles
15%
3,078
30.8
14.8
-12.9
-43.7
BigPond Cable
0%
390
3.9
1.8
-1.7
-5.6
Pay TV
20%
500
5.0
1.6
1.6
1.6
3
Ted Pretty, AFR 29 Nov. 2002
4
AFR 26 May 2003
John de Ridder
2
Source: Telstra and author’s estimates
The first profit column, Impact-A, considers the impact of a 10% discount on the gross revenues in the previous column, the associated marginal costs and the opportunity cost of wholesale business forgone. The final two columns include the cost of discounts to non-churning customers assuming 10% and 20% penetration of the discounts into the product markets. The question of cost savings is considered in the next section. As an example of how the net profit column is derived, the undiscounted unit price of long distance (LD=STD+ID) calls is 15.2 cents per minute, the discount is 1.52pm and the associated servicespecific marginal cost is 2cpm. But without this net retail yield of 11.7cpm, Telstra could make 2cpm on interconnection revenues so the difference to Telstra of reducing churn (e.g. moving LD business from wholesale to retail) is only 9.7cpm. On 1% churn reduction this is worth $9.4m pa (ie LD’s share of 1% which is $14.8m scaled down by 9.7/15.2). In terms of Pay TV, the net profit number is the difference to Telstra between migrating 1% of FOXTEL revenues to Telstra Retail compared with losing them from both. The marginal cost to Telstra is very high as it is buying the FOXTEL service at the list price less 7% 5. With Telstra reselling at list price less 10% and some incremental costs in adding the customer’s FOXTEL service to the single bill, Telstra is losing money directly on migrating customers from FOXTEL. Indirectly, there is a net benefit flowing from its 50% shareholding in FOXTEL. If the customer base is below the minimum subscriber guarantee and the cost of cable and FOXTEL billing is fixed, the incremental costs must be very small; say, 20% of revenue. There is also the cost of FOXTEL’s resale discount to Telstra and an opportunity cost. The indirect equity-adjusted benefit to Telstra of reselling $5m of FOXTEL product off-sets the direct losses to produce the net profit in Table 1 above. There is no discount leakage impact for FOXTEL service as the discounts only apply to revenue moving to Telstra.
2.1b – Cost reduction But, what about the benefits of reduced costs and will these exceed the leakage from discounting? Telstra says that bundling will reduce costs in billing and marketing and advertising 6. But, this probably refers to unit costs rather than incremental costs. We have been told 7 that Telstra will spend up to $1m to equip its shops to sell Pay TV which is an incremental cost. But no cut in the P&A budget or reduction in billing costs has been announced as a result of bundling. Rather, fixed costs are being shared over more services, reducing unit costs. While it is nice to have reduced unit costs, these do not drive earnings. Only incremental costs do that. Shareholder value is created when the present value of cash-flows caused by a bundling initiative is positive. There may well be cost savings in joint supply; a firm can integrate services better than a customer can. This includes reduced costs of complexity and any incremental reductions in shared costs such as billing and marketing (economies of scope). But, unless incremental costs are negative (ie total costs fall), they will not contribute to cash-flow. Similarly, we know that increased sales over existing infrastructure increases capacity utilisation and reduces average costs (economies of scale). Again, this does not improve earnings unless the incremental costs are negative.
5
AFR 26 May 2003
6
Ted Pretty, AFR 29 Nov. 2002
7
AFR 15 Jan. 2003
John de Ridder
3
Table 2 (which ignores satellite for simplicity) shows different fixed cost structures with some carriers, including Optus, relying on some Telstra infrastructure. The different cost structures have major implications for the impact of bundling on earnings. For example, Optus can carry all services except mobiles, on its HFC network. This means that once a customer is on this network for any one service, the incremental cost of adding the other services is likely to be very small. In fact, Optus needs the customer to take more than one service to make provision of HFC network access economically viable, but it is in a good position to make this happen.
Table 2: Cost impact of bundling across platforms
Telephony
Telstra
Optus
Other
Comments
PSTN
HFC or
PSTN
PSTN owned by Telstra
GSM
GSM
All separate networks
ADSL
ADSL over Telstra PSTN
PSTN Mobiles
GSM or CDMA
Broadband
HFC or
HFC or
Internet
ADSL
ADSL
Pay TV
HFC
HFC
nil
Source: The author
Telstra provides broadband internet services both on its cable network and on its copper (ADSL) network. Note that Telstra has not allowed Options and Rewards discounts against ADSL (although it is counted as an eligible service for determining which tier of discount applies to “relevant charges”). This is because the incremental cost of providing an existing FOXTEL customer with broadband internet access on cable is much less than providing such access on ADSL. Similarly, the marginal cost of inducing a BigPond customer to take FOXTEL is much less if he or she is already on cable. Putting the resold FOXTEL service on a single Telstra bill will not generate savings in billing. Unless all FOXTEL’s customers move to Telstra, FOXTEL will continue to have its own billing system. And, adding the service to Telstra’s billing can only have increased Telstra’s incremental costs.
2.1c – Revenues increased The analysis in Table 1 applies as much to increasing market share. But, what is more interesting is how bundling can increase the size of the total market. This is described in the next section and will be contrasted with Telstra’s current approach in section 5.3 below.
2.2 - How does bundling for growth work? Bundling is an “effective means of encouraging people to spend more” 8. It works best where customers value services or bundles of services differently. Where a customer is willing to pay more than the price charged, there is a “consumer surplus”. Part of this surplus can be shifted to induce the
8
The McKinsey Quarterly, 2003 Number 1, “A help line for European telcos”
John de Ridder
4
customer to take-up another service where he or she is not prepared to pay the asking price of that additional service: “By offering different prices for customers who buy different combinations of goods, a seller can separate buyers according to their demand characteristics in a way that might not be possible with independent pricing of all goods. This discrimination allows the seller to increase profits by extracting greater revenue from those buyers who are most willing to pay. When tying is used to facilitate price discrimination, its overall effect is typically to reduce the economic efficiency cost of monopoly power”9. To take a simple example, Table 3: Reservation Prices (the most the customer is prepared to pay) Customer
Service X
Service Y
Both Services
Ian
$40
$10
$50
Tim
$10
$40
$50
Source: Adapted from “Power Pricing” by R.J.Dolan and H.Simpson
Charging $40 each for X and Y results in total sales of $80 with Ian and Tim each buying only one service. A simple 10% discount if you buy both, such as in Telstra’s Options and Rewards, will not induce either of these customers to take-up the extra service. But if the services are bundled at a price of $50, total sales will be increased by 20% to $100 as Ian and Tim will both take the bundle. Of course, this is only profitable if the marginal cost of providing the extra service to each customer is less than $20. Actually, this simple example looks almost exactly how Optus used to sell its subscription TV service. Until the FOXTEL movie channels were added to its Pay TV service, the Entertainment and Sports Packs were both priced individually at $37.95 and these were bundled into the Deluxe Pack for $49.95 per month. This was classic bundling which recognized the different valuations that customers placed on the components of the bundle and the low marginal cost of providing the extra channels. However, Optus now sells an Entertainment Pack with more channels including sports for $39.95 at the entry point with optional add-on movie packs. This change makes sense if a movie pack is a “must have” (see comment below on “complementary demand”). In Table 4, we take a slightly more complex example. Table 4: Reservation prices Customer
Service X
Service Y
Both Services
Alan
$22
$3
$25
Bob
$2
$18
$20
Carl
$16
$14
$30
David
$20
$10
$30
Source: : Adapted from “Power Pricing” by R.J.Dolan and H.Simpson
Optimizing price on each service separately, leads to prices of $16 and $10 for X and Y respectively which will result in sales of $78 (ie $16 x 3 plus $10 x 3). Optimizing on what customers are prepared to pay for both services, the bundled price is $20. At these prices everyone will take the bundle and total sales are $80. This is only slightly better than pricing services independently.
9
Weinberg quoted in NERA’s report to the ACCC on “Anticompetitive Bundling Strategies”, January 2003
John de Ridder
5
Pricing at $16 and $10 each and also providing an Options and Rewards type discount of 10% when both services are taken does induce Alan to acquire an extra service ($23.40 is less than $25). But as the discount also benefits Carl and David who are happy to take both services at their individual prices, total revenues are only $80.20 (ie $23.4 x 3 plus Bob’s $10). While this is almost the same as the simple bundled price, we can do better. We have left money on the table with Carl and David who prepared to pay much more than $20. If we price the bundle at $30 and price Services X and Y individually at $22 and $18 respectively, we can claim Carl and David’s surplus and increase total sales to $100 even though Alan and Bob buy only one service each. This is illustrated in Figure 1.
Figure 1: Reservation prices
Price of Service Y 35 30 25 20 Bob
Carl
15 10
David
5
Alan
0 0
10
20
30
40
Price of Service X
Source: : Adapted from “Power Pricing” by R.J.Dolan and H.Simpson
While bundling works when the services are independent in demand (eg Pay TV and telephony services), it works even better when they are substitutes in demand (eg mobiles and fixed services). This is because it leads to more homogeneity in valuations of the bundle than will be found in the service components separately. Bundling is not as effective in increasing profits for services that are complementary in demand because customers are inclined to purchase all the components of the bundle anyway (eg local and long-distance calls and the Optus Pay TV example above). But, this situation is where regulatory concerns should be greatest (see section 3.2).
3 - What are the regulatory issues? The ACCC published a discussion paper10 and two commissioned reports on bundling in January 2003. In its later report to the Minister it clarified and confirmed its concern that “Used anti-
10
“Bundling in Telecommunications Markets” and associated reports found at footnote 1.
John de Ridder
6
competitively, bundling may foreclose or reduce competition by enabling the carrier or supplier to leverage market power from one market to another”11. There are three sets of legal and regulatory issues in bundling, •
Tying (Section 47 of the Trade Practices Act affecting all carriers)
•
Pricing (the Telstra conduct regime under Part XIB of the TPA)
•
Exclusivity (the Telstra access regime under Part XIC of the TPA)
3.1 – Tying For both Telstra and Optus, the Pay TV service (the “tying” service) is only provided together with fixed telephone service (the “tied” service). This form of bundling is also known as third line forcing when one service is provided by another supplier (e.g. FOXTEL, even though it is 50% owned by Telstra). All companies have to notify such conduct to the ACCC which has 14 days to object. Although Telstra notified the ACCC of its intention to bundle FOXTEL into Options and Rewards as early as July 2002, it did not launch until it obtained the ACCC’s advice on 13 November 2002 that it would not block it (“at this stage”). A related legal issue is full line forcing which occurs where the tied services are acquired from the same provider. Again, companies must notify the ACCC but immunity from court action is obtained immediately upon lodgement of the notification. And, again, the immunity gained by lodging a notification remains in force unless, and until, the Commission issues a notice revoking immunity. In economic terms, NERA noted12 that tying “could be considered anti-competitive if it allows the seller with market power in one market [for the tying product, say, Pay TV] to ‘leverage’ that power into a second market in which it would otherwise face competition” [the tied product such as long-distance phone calls]. A factor not considered explicitly by NERA is that the impact of bundling on competition depends upon how the services are correlated in demand (see section 2.2 above). If they are substitutes in demand (ie negatively correlated), there should be less regulatory concern as single service providers can still compete. This applies even if there is a near monopoly of demand for one of the services in the bundle. This is one reason why mobile companies can compete with Telstra’s bundling of fixed and mobile telephony on a single bill. There is also little reason for concern if the services are independent in demand; “a sufficient proportion of consumers in the market for the tied product [long-distance] must also be willing to purchase the tying product [Pay TV], otherwise the tie-in will be unlikely to capture a substantial enough proportion of rivals’ sales to have a harmful effect” 13. FOXTEL and telephony appear to be independent in demand but several of Telstra’s telephony competitors (excluding Optus) disagree. That is, they think selling Pay TV helps sell fixed telephony. The ACCC has indicated it is keeping a watching brief in this ultimately empirical issue. It is seeking data on the uptake of bundled offers and their cost from Telstra. There is, however, a very justified regulatory concern where two services are complementary in demand (ie positively correlated) and one service (the “tying” service) is a near monopoly. First, as noted in section 2.2, bundling is unlikely to improve profits. Second, and more importantly, it creates a hurdle for competitors because they now have to supply both services to satisfy customers. But, as
11
Section 8.2.1 of “Emerging Market Structures in the Communications Sector, June 2003
12
“Anticompetitive Bundling Strategies”, NERA, January 2003
13
Ibid.
John de Ridder
7
discussed above, it is not yet clear if Telstra’s competitors have to sell both Pay TV and fixed telephony in order to be competitive. However, even when there is complementary demand there can be mitigating factors or remedies which can allow bundling to proceed: •
Possibly, the ACCC could be satisfied that there is a cost basis for any price differences between the bundled service and its unbundled parts. But, this may be difficult to establish. The new, tougher accounting separation requirements will not help as they apply only to the traditional fixed line network (i.e. they exclude BigPond, ISDN, mobiles and Pay-TV).
•
Or, the access regime effectively eliminates the monopoly so that competitors could construct similar bundles (even if they choose not to) which is discussed in 3.3 below
•
Or, the bundling is seen to be necessary for safety or to control quality; which does not appear to be the case with any of the bundling in question.
3.2 – Pricing The underlying concern of smaller competitors is “related to fears that Telstra would set wholesale prices at a level that would squeeze ‘single service’ providers that rely on Telstra out of the market” 14 The second of the two NERA reports15 for the ACCC says “there are three necessary conditions for an anti-competitive price squeeze to be a rational and viable strategy for an integrated firm: •
two markets must be vertically related and the upstream product must be a necessary input into producing the downstream product
•
at least one firm must be vertically integrated and possess market power in both the upstream and downstream markets; and
•
the downstream market must be open to competition from rival, non-(vertically) integrated firms.
The bundling of FOXTEL with fixed telephony, which is the focus of most angst, does not fit into this set of conditions. Neither is an essential input for the other and it is debateable how much market power Telstra has in the call market. NERA itself says often that they are in separate markets. Despite this, NERA says that if the ACCC wants to perform an imputation test (ie retail price > access price + marginal cost) to detect a price squeeze where one element, Pay TV, is not in the same market as fixed telephony (and not provided by Telstra’s competitors), the (discounted) price of Pay TV can be subtracted from the price of the bundle before imputing wholesale prices and retail costs to the inputs of the fixed telephony service. Once the subtraction has been done, this looks just like the conventional imputation testing the ACCC has performed on Telstra’s fixed telephony business over a number of years. The ACCC looked at the bundling of Pay TV and telephony this way and concluded that “Telstra’s proposed pricing does not indicate a price squeeze based on information provided to the ACCC” (13 Nov. 2002). Given the subtraction approach, another 5% to 10% discount by Telstra on calls to some of the market is consistent with the competitive process. However, while this approach may work in this instance it is not going to work generally. If we revisit the simple example in Table 3 and add some assumptions about access prices and marginal costs as shown in Figure 5, each service passes the imputation test when sold separately (40 > 20 + 10). But, the bundle does not pass the recommended test. Subtracting service Y from the bundle as recommended (50 – 40) < 20 + 10.
14
ACCC, 13 November 2002
15
“Imputation Tests for Bundles Services”, January 2003
John de Ridder
8
Table 5: Imputation Testing Customer
Service X
Service Y
Bundle
Prices
$40
$40
$50
Access price
$20
$20
$40
Marginal cost
$10
$10
$20
Source: Table 3 and author
This is not an efficient rule as deprives consumers of additional consumption. An alternative recently suggested by Tom Hird16 is to pass the imputation test the bundle price > opportunity cost of the bundle. In Table 5 the opportunity cost is the margin on service X sold on its own (ie 30=40-10) plus the marginal cost of supplying service Y (ie 10). This is less than the bundled price of 50 so the test is passed. But, it has an important corollary: access seekers who wish to provide the same bundle have to have access prices reset from 40 to 10 (ie bundle price less opportunity cost) for both services.
3.3 – Exclusivity AAPT has suggested that there are “risks of providing Telstra and Optus with a near exclusive ability to bundle Pay TV with a range of telecommunications services”17. The jury is still out on whether it is necessary to offer Pay TV in order to be competitive in fixed telephony. If it is, some might see a precedent in the decision to “declare” Telstra’s local call service which was rationalized in the context of bundling. The ACCC saw this declaration as enabling Telstra’s competitors to supply an integrated package of services (local, long-distance and international) all on a single bill to a customer18. While the single bill is an undiscounted bundle, it has been a very attractive convenience proposition for customers. The declaration of local calls removed regulatory concerns about bundling across local (monopoly) and long-distance (competitive) call services. Following this approach, the ACCC has secured voluntary section 87B undertakings from the three parties to the content sharing arrangements that would allow the resale of a Pay TV service. But, other resellers of FOXTEL19 will not get as good a deal as Telstra and will not enjoy the indirect benefits Telstra obtains through its part ownership of FOXTEL20. The voluntary 87B undertakings given by Telstra and Optus Vision with respect to their HFC cable networks concern only television content and distribution21. That is, Telstra and Optus remain the major providers of broadband internet over cable. It is unclear whether ISPs consider the ability of Telstra and Optus to sell both voice and internet services against them is a major problem for them. It was a major issue in making broadband services over wholesale ADSL pay (ie the business models looked weak without telephony); but this is no longer the case as “Line sharing will enable broadband
16
Tom Hird (NERA) “Competition and Price Discrimination in Network Industries: Access Pricing and Imputation Tests in the Presence of Bundles” (ITS Conference, Perth June 2003) 17
Former CEO, David Bedford, AFR 15 November 2002
18
ACCC December 1998
19
AAPT and Harvey Norman have agency agreements to resell FOXTEL (AFR 26 May 2003)
20
See discussion around Table 1
21
But the ACCC warnd that declared access to HFC “may need inquiry by the Commission…particularly if Telstra retains ownership of its HFC” p41 “Emerging Markets” report, June 2003
John de Ridder
9
providers to focus on providing high-speed data services to consumers, such as the Internet, video-ondemand and other multi-media applications, without needing to take on the responsibility and cost of providing voice services” 22 Note also that FOXTEL’s undertaking to resell “Sports Plus” (Fox footy and ESPN) goes some way to resolving the ACCC’s concern about FOXTEL’s monopoly of Pay TV content.
4 – What are the implementation issues? Even if regulatory concerns are assuaged and there is a market opportunity to increase profits through bundling, the benefits can be lost through poor execution. There are a number of implementation issues as illustrated below. Figure 2: Customers needs and behaviours - links to different levers
Brands Carried/ Product Mix
Pricing and Terms
(
Sales Effort/ Resources Allocation
Sales Support/ Reseller Service
Telstra Levers )
Product Design and Features
FOXTEL Levers
Retail Pricing
)
Advertising and Promotion
Advertising
Post-Sale Customer Support
Initial Need Arousal
(
Reseller Support and Service
Brand/Resell er Awareness & Image
Reseller Visits Information Search
Reseller consideration Set
Choice
Delivery and Installation
Post-Sale Experience
Understanding customer thoughts and actions is essential to understanding which are the most important levers to pull
Source: Corporate Value Associates
Note that price is only one factor among many. There are a number of examples where the only obvious difference between success and failure has been, for example, advertising and promotion. Issues for Telstra include cross “product-silo” management coordination and the division of costs and benefits between the internal and external units involved. With the current very simple form of bundling these issues are not too large and the organization restructure at Telstra announced in December 2002 eliminates the tensions between OnAir and Telstra Retail on customer ownership at the stroke of a pen. A major internal difficulty in developing bundled mobile-fixed telephony offers would have been that while OnAir saw Telstra Retail as simply an alternative sales channel to its dealerships, Telstra Retail saw OnAir simply as a product-house. Such tensions will no longer arise now that OnAir has been absorbed by Telstra Retail. It may not have been so easy with FOXTEL. It was reported that “FOXTEL’s shareholders spent much of (2001) haggling over plans to introduce a bundled Pay TV – telephony service” 23. Agreement with FOXTEL on “customer ownership” may have been Telstra’s pay-off for agreeing to the content-sharing deal with Optus. But, that agreement was based on Options and Rewards. A more sophisticated
22
Professor Allan Fels, 2 December 2002
23
The Age, 6 March 2002
John de Ridder
10
growth-oriented bundling offer could need further cooperation from FOXTEL; but not the option put forward in section 5.3.
5 – Will bundling succeed? Telstra clearly sees bundling as an important part of its business model. It made the issue of bundling across FOXTEL, telephony and BigPond a show-stopper in the FOXTEL-Optus content-sharing arrangements. Now that it has the waiver it sought for “third-line forcing” Telstra has “all the tools we need not only to protect our business but (also) to recover market share, to reduce churn and deepen the relationship we have with our customers” 24. Telstra already bundled across services under its “Options and Rewards” program. This provides a 5% discount where a customer registers the purchase of two eligible services (eg fixed and mobile telephony) and 10% if three eligible services are purchased (eg fixed and mobile telephony and Big Pond)25. The “eligible services” and “relevant charges” for discounting are set out in Table 6. Table 6: Options and Rewards
Eligible Service
Relevant charges for discounting [1]
Comments
Fixed Service, ie
Call costs (excluding services below)
Must take access and local calls and be preselected to Telstra to be have an Options & Rewards discount
All HomeLine Services except HomeLine Part All BusinessLine except Businessline Part
Not HomeLine Plus calls
Part services do not require preselection to Telstra (used in resale)
Not BusinessLine Plus calls
All ISDN Services except OnRamp Express Mobile Service except prepaid
Excluding roaming and calls under some plans
BigPond Home and Business (ie dial-up)
All monthly and usage charges
But not BusNet 10 nor BigPond Direct
BigPond Cable
All monthly and usage charges
Not on single bill
BigPond ADSL
Not discounted
Not on single bill
BigPond Satellite
All monthly and usage charges
FOXTEL
All monthly charges
Acquired from Telstra
[1] See Telstra’s terms and conditions for various exceptions and exclusions.
24
Ziggy Switkowski, AFR 14 November, 2002
25
There is also a 15% anniversary discount applied after the O&R discount in the 12th month which would bring the average annual discount to just over 11%. But, for simplicity it is ignored in this paper’s analysis.
John de Ridder
11
Source: Telstra and the author
5.1 - How does Telstra’s bundle compare with Optus? The following bundle of services could be acquired under Options and Rewards with an effective discount of 8.6% (mobiles could be included).
Table 7: Telstra bundle for a valuable residential customer (illustrative)
$ per month $21.90
HomeLine Complete (access not discounted)
$15.36
Local (22 cent) and Neighbourhood (15 cent) calls
$20.00
STD, ID and Fixed-to-Mobile calls
$39.95
FOXTEL Basic Package on cable
$54.95
300MB plan
----------$152.16
Total before discounts
$139.13
after 10% discount (excl. PSTN access); effective 8.6% discount
Source: Telstra and the author
This is more like price discounting than bundling because it has a product focus; add another service and get a better discount. Options and Rewards do not reflect how customers value each of the services; each component in the bundle is considered identical. It is likely to lead to margin erosion. This simple discounting does not reflect the relative product margins and price elasticity of demand needed for profitable pricing. Fortunately, the discount is capped at 10% and does not apply to all services (eg PSTN and ADSL line rentals are excluded). Without FOXTEL or a Telstra mobile service, the discount would be only 5% as only two eligible services (fixed telephony and BigPond) are taken. The cost of adding FOXTEL is not just the 10% discount on FOXTEL of $4 pm but also the increase in the discount on the other services from 5% to 10% which is worth another $4.52 pm. Optus has been bundling across telephony, Pay TV and cable internet for a bit longer than Telstra. At March 2003, 60 per cent of customers on its HFC network and 22 per cent of its off-net customers took two or more services (one of which has to be fixed telephony)26. There are 5 “Choice” bundles to choose from with different mixes of services (see Table 8 below). There does not appear to be any incentive to migrate up to a higher value bundle. The fixed telephony component of each bundle and the PayTV components (which appear in 3 of the bundles) cost exactly the same.
26
Martin Dalgleish, Optus MD Consumer and Multi-Media, 23 June 2003
John de Ridder
12
Table 8: Optus Choices
Optus Choices:
Fixed Telephony Pay TV
1
2
3
4
5
■ ■
■
■ ■ ■
■
■ ■
■
■
■
Dial-up Internet Cable Internet Source: Optus and the author
The nearest corresponding Optus bundle to the Telstra bundle in Table 7 is the “Choice 5” package which has a number of options but could consist of the following elements in Table 9. Note that the “normal” prices are not used in the other Choice bundles.
Table 9: Optus Choice 5 (illustrative)
$25.00
Line rental (normally $29.90)
$0
Local calls at 15 cents
$20.00
Long distance and Fixed-to-Mobile calls
$39.95
Entertainment (Pay TV) Pack (add-on packs available)
$54.95
Lite (550MB) OptusNet Cable (normally $64.95pm)
(first 100 per month free)
---------$154.80
normal price without Choice 5
$139.90
with Choice 5; effective discount is 9.6%
Source: Optus and the author
5.2 – What is the earnings impact of Telstra’s current bundling? As discussed earlier, Telstra’s current bundling strategy is defensive and concedes discounts where they may not be necessary. To scope the earnings impact of the current strategy, assume optimistically that it results in a 4 percentage point reduction in churn levels across all services. This is very optimistic in FOXTEL’s case because this is what its CEO is hoping to achieve over 3 years27.
27
Kim Williams, The Age 6 March 2002 which is also consistent with later comment that “we will not rest until we it first below 15% and then lower than that” (AFR 26 May 2003).
John de Ridder
13
And, it is optimistic for all services because the way we interpret churn it is synonymous with an increase in market share (see section 2.1a above). Applying the analysis developed in section 2 to this scenario yields the results shown in Figure 3.
Figure 3: Churn reduction of 4% (Hypothetical)
Preselectable calls Mobiles
$m pa 140.00
BigPond
120.00
FOXTEL
Less associated discounts to 10% of base 20% of base
100.00 80.00 60.00 40.00 20.00 0.00 -20.00
Revenue
Profit-A
Profit-B
Profit-C
Source: Author’s estimates
This is Table 1 scaled up to 4% and so again the result is sensitive to discount leakage. Profit-B based on 10% leakage is probably the minimum one would expect from the vigorous marketing of Options and Rewards. It was reported that “about 17% of Telstra’s customers now buy two or more services in a bundle, up from around 10% at the end of last year”28. Until November 2002, most of the promotion was “below the line” (eg bill inserts) but since then it has been advertised heavily. The results in Figure 3 look better than those of Table 1 because the discount leakage scenarios are unchanged while the churn benefits are four times larger. But, a larger increase in market share leads to a greater risk of price war. So, it is doubtful that such market share gains could be achieved because competitors would lower their own prices too. And, at what point do market share gains trigger a regulatory response which might stop bundling?
5.3 – How could bundling benefit Telstra more? An alternative strategy is to use bundling to cross-sell products using the principles described in section 2. Telstra says it is trying to cross-sell: “It’s about upselling, and when you’ve got the database you can use CRM [customer relationship management] tools to target specific customers” 29. This was
28
AFR, 23 September 2002
29
AFR, 15 January 2003
John de Ridder
14
said in the context of selling FOXTEL to Telstra customers but a more profitable opportunity is to upsell FOXTEL customers to BigPond cable. This scenario also consistent with and supports Telstra’s growth strategy. Telstra plans to get the number of broadband internet customers up to 1 million (includes business and wholesale customers); most of which will be ADSL.. At the same time, FOXTEL’s aims to double its penetration rate (i.e. another 0.8m customers). Given FOXTEL’s growth aspirations and the low take-up of BigPond Cable among existing customers, there is an opportunity to increase BigPond cable penetration. If we start with FOXTEL as part of any bundle, Telstra’s addressable market comprises the 2.5m households passed by its cable. About 0.5m of these take FOXTEL on cable (the rest are on satellite) and there are about 0.1m customers on BigPond Cable30. It is not known what overlap already occurs and it is doubtful whether Telstra knows either as customers for BigPond Cable and FOXTEL are on different billing and customer management systems. However, customers will identify the over laps for Telstra when they select eligible services for Options and Rewards. Assuming there is no overlap between BigPond Cable and FOXTEL we have the following:
Figure 4: Telstra cable customer segments
BigPond Buy
Not Buy
Buy
0
0.5m
Not Buy
0.1m
2.0m
FOXTEL
Source: The author’s assumptions
The largest and hardest market segment to target comprises the 2m households passed by cable that do not take either FOXTEL or BigPond Cable. Some of these will take-up one or both services as cable penetration levels increase. The current 22% penetration levels for Pay TV are low relative to the 40% level attained by BSkyB in the UK due to anti-siphoning laws protecting free-to-air sports in Australia and due to BSkyB going digital in 1999. By comparison, BSkyB’s non-digital rival in the UK, NTL, has a penetration level of only 28% due also to poorer content. FOXTEL has the content, is planning to go digital in 2004 and is aiming for 35-40 per cent penetration by 200831.
30
At the end of March 2003 the ACCC reports 196,000 broadband cable customers (AFR 19 June 2003) and Optus reports 96,000 broadband cable customers (AFR 14 May 2003) 31
AFR 26 May 2003
John de Ridder
15
In the next largest segment are the existing 0.5m FOXTEL customers who might be persuaded to buy BigPond Cable. This seems like a good match given the circumstances that lead to profitable bundling discussed in section 2. That is, it seems likely that the cable TV and broadband cable services are negatively correlated (ie substitutes). In crude terms, the “couch potatoes” and “computer nerds” are in different market segments; but these can be brought together in a bundle. Hypothetically, in Table 10 we rework an earlier example. Table 10: Reservation prices, $ per month (hypothetical) Customer
FOXTEL
BigPond
Both
Ian
$50
$30
$80
Tim
$20
$60
$80
List price
$40
$55
Source: The author
It is likely that customer segments value these services as differently as shown in this hypothetical market research. If the willingness to buy both services is only $80, a 10% discount on the existing prices of $40 and $55 is still above the reservation budget and will not stimulate extra sales. But, suppose instead that the bundled price is set at the size of the available wallet. At $80 both customers will purchase both services. The incremental revenue from selling BigPond Cable to an existing FOXTEL customer would be $40 which is a discount of 27% relative to the list price of $55. Assuming, as in earlier calculations, that the marginal cost is one third of the list price then this leads to profitable growth. It is unclear whether this would pass the current imputation by subtraction test discussed in section 3.2 above. Subtracting BigPond from the bundle seems to fail the test: (80-55) < 40 + 032 . Starting at the other end and subtracting FOXTEL from the bundle is not possible as the access price for BigPond cable is not defined. However, since FOXTEL can be resold and Telstra does not have bottleneck power over high-speed internet (eg unbundled local loop and line sharing are declared services), the test seems unnecessary. Note that it is not possible to offer this kind of marginal pricing on ADSL not only because the marginal costs would be too high as discussed earlier, but also because the retail versus wholesale price relativities are keenly balanced. Wholesale ADSL prices have already been adjusted down as a result of a Competition Notice finding that retail prices were too low (due to an imputation test based on monthly downloads). But, this might still be accommodated by the innovative approach suggested by Tom Hird (see section 3.2). But how significant could this scenario be for earnings and how does it compare with the current strategy? Assume 100,000 sales of BigPond Cable to FOXTEL cable customers (i.e. 20% of a growing base). The results in Figure 5 are compared with the aggregated results of a 4% churn reduction (from Figure 3).
32
Access price of FOXTEL is 40 less a small discount or commission on resale and even if marginal cost of FOXTEL is zero, test still fails
John de Ridder
16
Figure 5: Upsell 100k FOXTEL cable to BigPond Cable (Illustrative)
$m pa
300.00
Churn reduction
250.00 200.00
Less associated discounts to 10% of base 20% of base
Cross-sell
150.00 100.00 50.00 0.00
Revenue
Profit-A
Profit-B
Profit-C
Source: The author
This suggested price bundling approach minimises discount leakage because discounts are only applied where they increase incremental profits. Discount leakage is the major source of earnings dilution in the current strategy. Under the current strategy, one dollar of discounts generates one dollar of incremental earnings (with 10% of base discounted an extra 10%) or none (with leakage to 20% of base). The proposed approach generates 50% more incremental earnings per dollar of discount without any leakage risk and less price war risk. The proposed approach also minimises regulatory risk as: •
in the current approach, the monopolistic “tying” product that concerns competitors and regulators is FOXTEL while here it is cable Internet which is subject to competition, particularly from ADSL.
•
the target customer base is narrow (FOXTEL customers) lessening the impact on, say, ADSL services and with no impact on telephony markets
As noted, the proposed approach dovetails Telstra and FOXTEL’s growth strategies and it will also achieve some churn reduction; but less than current strategy. Another benefit not factored-in is that apart from up-selling FOXTEL customers to cable, it can also target FOXTEL’s non-Telstra customers for other Telstra services.
6 - Conclusion Telstra’s current approach is a defensive strategy whose churn reduction benefits are largely off-set by discount leakage and which carries market (e.g. price war) and regulatory (e.g. divest or cease and desist) risks. Claims of cost savings may confuse unit cost reduction with incremental cost reduction where only the latter benefits earnings. Options and Rewards may have been an easy way to incorporate FOXTEL into the Telstra offer quickly, but it is now time to develop a more profitable bundling pricing strategy including FOXTEL that assists the roll-out of broadband services on cable.
John de Ridder
17
The Optus Choices could be revamped to increase the uptake of higher value packages. At the moment they are simply lifestyle choices. Optus should have low marginal costs where it can use its HFC infrastructure to sell service. On the evidence available, the major carriers have some way to go in making bundling work profitably. Their current strategic purpose is defensive, which may commoditise services and reduce margins.
DO NOT PUBLISH THE WORKING NOTES BELOW Working notes Table 1: Value of 1 percentage point reduction in churn Revenue
Churn
- 1% Churn
- 1% Churn
($m, 2002)
(per yr)
($m pa gross)
($m pa net)
STD+ID+FTM calls
2,881 [1]
24%[2]
28.8
10.8 [3]
Mobiles
3,078
15%[4]
30.8
14.8 [4]
BigPond Cable
390 [5]
n.a.
3.9
1.8 [5]
Pay TV
500 [6]
20%[6]
5.0
1.5 [7]
Source: Telstra and author’s estimates
Notes to Table 1: [1] Telstra’s retail telephony 2002: Revenues
Minutes
$m
m
Line Rentals
2,394
Local Calls
1,525
10799 (no.)
Fixed-to-Mobile
1,401
3691
National Long Distance
1,161
8946
International Calls
319
Value Added Services
143
819
6,943 [2] Productivity Commission report on Competition and Regulation (21 September 2001) says “the number of customers churning between competing providers (incl Telstra) for their preselected long distance traffic has increased from around 130k pm in mid-1997 to 200k pm by early 2000”. The preselection basket includes international and calls to mobiles. 24%= (200k * 12) / 10m lines [3] The gross figure is overstated because (a) there is the cost of the O&R discount (10%), (b) there are marginal costs and (c) there are opportunity costs, ie some of what Telstra loses (wins) at the retail level it gains (loses) as wholesale business.
John de Ridder
18
In the case of long distance (STD+ID), the retail yield is 15.2 cents pm (from note [1]) compared with a marginal cost of, say, 2 cents pm. But, the net gain is not 13.2cpm because there is the cost of the discount (1.52cpm) and there is also an opportunity cost of foregone interconnect at, say, 2cpm (actually more as they pay the ADC too. The ADC does not have to be counted in the marginal cost because it is funded out of the contribution left after mc is taken out). So, the net gain from holding the customer is 9.68 cpm and the revenue-weighted net gain compared with wholesale is $9.42mpa LD = $9.42m = ((1480/2881)*28.8)*(9.68/15.2) For F2M calls, the retail unit price is 38 cents pm but 53% (MRE’s Aug 2002 estimate of mobiles market share) of calls terminate on non-Telstra networks at, say, 23cpm. So, the net retail yield is (38cpm*90%)(23cpm*53%)=22cpm. But, if these calls were preselected to another carrier Telstra would avoid the cost of the discount (3.8cpm) and would collect 10.8cpm (ie 23cpm on 47% ) on FTM calls which terminated on its own mobile networks. The difference between keeping and losing the calls is 11.2cpm= 22-10.8 F2M = $4.13m = ((1401/2881)*28.8)*(11.2/38) or $13.55 mpa for LD and F2M together. But there are, say, 5 cpm of shared avoidable costs (eg billing) or $2.71mpa = 28.8*(5/(15.2+38)) bringing the net benefit to shareholders to $10.84mpa (= 9.42+4.13-2.71). [4] Based on 1.5% pm churn in Q4 2002 quoted by Telstra to analysts. The gross benefit of reduced churn is overstated because the 47.3cpm retail revenue ($2734m in access plus calls over 5,780m minutes – both include prepaid) is off-set by direct costs of, say, 10 cpm (eg P&A, dealerships, subsidies), the cost of the O&R discount (4.73cpm), costs in interconnecting calls to other mobile networks and the opportunity cost of wholesale business forgone in calls to Telstra’s fixed (2 cpm on 60% of calls) and mobiles (23 cpm on 40%*47% of calls) networks. So, the net retail yield of 27.69 (=47.3-4.73-10-(23*40%*53%)) less opportunity cost of 4.92 = 60%*1cpm (one end)+ (40%*47%*23)) is 22.77cpm. $14.83 mpa = 30.8 * ( 22.77 /47.3) [5] Annual Report shows internet revenues as $55m but this includes ADSL and bus and w’sale so assume 70% cable. Churn-out from BigPond broadband Internet is probably not much of a problem yet; the customer base is small and there have not been alternatives. If a customer leaves, there may be some marginal costs saved in data carriage and customer care reducing the net loss. If the customer leaves to take-up an alternative ISP service, there is probably some wholesale business to reduce the net loss even more. Overall, the net avoidable loss is, say, one third of revenue allowing also for the O&R discount. [6] $500m estimated from ARPU of $55pm (MRE, Aug 2002, p113) and 0.8m customers. Kim Williams, reported 20.5% churn in The Age 6 March 2002. [7] See spreadsheet.
Table 6: Telstra bundle of Pay TV + Telephony + Cable Internet
$21.90
HomeLine Complete (access not discounted) Chosen over Home Line Plus whose calls are not discounted
$15.36
Local (22 cent) and Neighbourhood (15 cent) calls Assumes 80 local calls pm split 60:40
$20.00
STD, ID and Fixed-to-Mobile calls Assume $20 pm across all these
John de Ridder
19
$70.00
Mobile service
$39.95
FOXTEL Basic Package on cable ($43.16 on satellite) The average customer yield is $55 due to extra channels bought
$54.95
300MB The next is $64.95 for 1GB plan but neither mirror Optus
----------$222.16
Total before discounts
$202.13
after 10% discount (excl. PSTN access); effective 9% discount
Source: Telstra and author’s selections and estimates
Table 8: Optus Choice 5
$25.00
Line rental (normally $29.90)
$0
Local calls at 15 cents (first 100 free every month) At 80 calls pm, this is slightly better than $20 line with 20c calls (first 50 free each month)
$20.00
Long distance and Fixed-to-Mobile calls Assuming $20 for both Optus and Telstra simplifies a lot
$39.95
Entertainment Pack (add-on packs available)
$54.95
Lite (550MB) OptusNet Cable (normally $64.95 pm) The next is Standard 3G at $69.95 (normally $79.95)
---------$154.80
normal price without Choice 5
$139.90
with Choice 5; effective discount is 9%
Source: Optus and author’s selections
Table xx: Recurrent revenues - Omit mobiles Service
Revenues
Customers
ARPU pm
Margins
PSTN
$2,394m [5]
8.74m [1]
Mobiles (Post paid)
$3,078m [4]
4.062m [4]
BigPond
$390m[6]
1.276m [3]
Medium
FOXTEL
$400m [2]
0.8m [2]
Low
High $63 (post-paid)
High
Source: Telstra 2002 Annual Report
John de Ridder
20
[1] Of which, 6.35m were residential [2] Of which, 0.487m on cable (rest satellite). $400m estimated. [3] Of which, 0.168m are broadband. Includes business and wholesale customers. [4] $63 x 4m (total reported $3,475m) Post paid 71% of 5.7m (excl inactive prepaid). Telstra estimates mobile penetration as 63% of population. [5] See Table 1, note [1] above [6] Table 10 Annual Report shows $553m for Internet and ISP. Assume 70% Res.
Table xx : Target cabled residential market for bundling with FOXTEL (author’s estimates) Service
Revenues
Customers
$1,350m [1]
4.5m [2]
High
BigPond
$70m[3]
0.1m [4]
Medium
FOXTEL
$240m [5]
0.5m [6]
Low
PSTN
Margins
Source:
[1] $25 x 12 x 4.5m [2] Addressable market (existing 0.8m+0.2m cable customers = 22% penetration) [3] $55pm x 12 x 0.1m [4] At 31 March 2002, total b’band by cable is 146,800 (AFR 13 August) say Big Pond has 70%. Also, ACCC (2 Dec 2002) reports 158k cable customers at Sept ’02; 70%=110k (mostly Res). [5] (0.5/0.8) * $400m=$250m and $40 x 12 x 0.5m = $240m [6] I understand (from where ?) that 0.5m of FOXTEL’s customers are on cable.
John de Ridder
21
