Acquisition by Satellite Information Services Limited of Racing Data Plc

No. ME/1649/02

A report under section 125(4) of the Fair Trading Act 1973 on the advice of the Director General of Fair Trading, given on 18 February 2003, to the Secretary of State for Trade and Industry under section 76 of the Act

ASSESSMENT

Jurisdiction

The merger satisfies the share of supply test in respect of the provision of processed text services to licensed betting offices (LBOs). The ECMR does not apply.

The parties

Satellite Information Services Limited (SIS) provides LBOs with: live pictures and commentary on horse and dog racing (FACTS service); processed text service giving information and odds on racing; raw data on racing for those customers (eg large chains of LBOs) wishing to format their own text information service; and associated equipment. In the year ended 31 March 2002 it reported pre-tax profit of £6.6m on turnover of £93m and gross assets of £27m. Racing Data provided LBOs with a processed text service. In the year ended 31 December 2001 it reported pre-tax profit of £22,405 on turnover of £313,274 and gross assets of £407,246.

Relevant markets

The parties overlap in the supply of processed text to LBOs. A text information service provides LBOs with information – runners, riders (for horses), trainers, form etc – and betting odds on horse and dog races. A text service is seen by LBOs as an essential part of their service and a complement to the FACTS picture service supplied by SIS (and taken by all LBOs).

Supplying a text service involves 'processing' raw data, which can be sourced either from SIS or the Press Association (PA). SIS provides raw data to larger chains of LBOs, which they then process internally into their own text service, and to a company called Alphameric (see below) which processes the data as agent for its LBO customers. SIS also supplies a generic processed text service which it makes available to individual LBOs. Racing Data receives raw text from the Press Association (PA) which it processes and transmits to individual LBOs. For small chains of LBOs or independent LBOs there is no good substitute for a finished processed text service; processing raw data would not be commercially viable for such operators and SIS contracts prevent LBOs co-operating to purchase raw data and studio facilities centrally. We regard the supply of processed text services as the relevant product market in this case.

Small chains of LBOs or independents can purchase a processed text service from SIS, Alphameric or Racing Data. Alphameric processes data sourced from SIS which is then delivered to individual LBOs via SIS distribution. However, LBOs using Alphameric must also have an agreement with SIS for the sourcing of the data. Alphameric also provides software solutions to larger chains processing text in-house.

The geographic market is likely to be the UK as available suppliers are all located in the UK.

Horizontal issues

For LBOs taking a finished processed text service, the parties' combined share of supply in the UK (by number of shops) has been estimated as 87.8 per cent (increment 6.9 per cent) (see note 1). Alphameric's share of supply is the remaining 12.2 per cent. However SIS's share of supply to LBOs taking a finished processed text service is 100 per cent (increment 6.9 per cent) post-merger in terms of the sourcing of the data (from which the text service is derived), as Alphameric's customers must have contracts with SIS. SIS has estimated its combined share of supply of processed text across all LBOs, including those that process data in-house, as 31.8 per cent (increment 1.9 per cent) with Alphameric at 26.1 per cent and the remaining 42.1 per cent represented by LBOs processing data in-house. However, this last figure may be an under-estimate as it is possible that some LBOs counted by SIS as customers of Alphameric are actually using Alphameric systems for their own in-house processing rather than having a text service provided by Alphameric.

Whatever basis is taken, the increments in share of supply attributable to the acquisition of Racing Data are small. At the time of the merger, Racing Data had just 170 customers (out of a total of over 8,000 LBOs). While its original business plan was to attract 500 customers, numbers peaked at 300. One reason given by third parties for the overall decline in the number of customers was increasing customer dissatisfaction with Racing Data's services. In particular, it appears that ante-post odds sourced from PA data will sometimes be inconsistent with the odds appearing on SIS's FACTS picture service which can lead to confusion in the LBO as to, for example, the odds available on a particular race. These problems and an alleged lower service quality meant that Racing Data appears to have been seen as a budget alternative to SIS. However, over time even this attractiveness was eroded as prices for the Racing Data and SIS processed text services converged. At the time of the merger, Racing Data's price per LBO was around £500 per quarter and SIS's was £1,923 per annum. Further, smaller independent LBOs are increasingly being bought by the larger chains and thus switching to their in-house text service, thereby reducing the available customer base for Racing Data. These factors tend to indicate that any competitive constraint on the SIS text service provided by Racing Data, particularly by the time of the merger, was limited.

Some third parties considered that Racing Data still provided some competitive constraint on SIS and were concerned that SIS would increase prices post-merger. (SIS has given Racing Data customers a written guarantee it will not increase prices for two years.) In this context, two issues were considered:

• There was a suggestion that, in response to Racing Data's launch in 1994, SIS had introduced a 40 per cent price reduction to processed text customers using equipment more than five years old. However, documents provided by SIS show that this discount was introduced in 1992 (on SIS's fifth anniversary). So it was not, as has been suggested, a competitive response to Racing Data's entry.
• Even if Racing Data had provided some competitive constraint when it was launched, customer churn figures since then suggest that this constraint had diminished over time. In each year since 1998 Racing Data lost more customers than it gained and only 4-10 new customers were gained each year from 1999 to 2002. Although the fact that Racing Data won some new customers each year suggests that it could compete at some level with SIS, the small number of new customers (some of whom were gained through acquisition by existing Racing Data customers) indicates that any competitive constraint was limited.

The evidence that Racing Data was still providing a substantial competitive constraint upon SIS in the supply of text services is therefore weak.

Barriers to entry to providing a processed text service appear to be considerable. Sunk costs appear to be high in relative terms. Entry costs would include setting up a studio for processing raw data and developing a delivery system for processed text. A source of raw data would also be required. Currently only SIS and PA provide raw data and continued provision by PA cannot be guaranteed. A competing provider of a processed text service would, moreover, require compatibility with a picture service where SIS currently has an exclusive deal for the provision of pictures and text for afternoon greyhound racing (see note 2) and some horseracing. The decline of independent LBOs (with an increasing number being bought up by large chains) reduces the size of the potential customer base for which a new entrant might compete. Since the major LBO chains are shareholders in SIS (see note 3) they appear unlikely to switch from SIS.

Buyer power appears to be low as the only customers of a processed text service are small LBO chains that buy independently.

Vertical issues

Vertical issues in relation to raw data collection and the provision of picture services have been considered above. The shareholdings by the major chains of bookmakers in SIS also raise the possibility of SIS being able to take steps which would impact on the downstream bookmaking industry and adversely affect the smaller chains and independents. The question at hand is whether the potential for such adverse effects is increased by the merger. As already indicated, the competitive constraint provided by Racing Data pre-merger appears to have been rather limited.

Other issues

SIS has claimed that Racing Data was a failing firm, unable to meet the necessary investment cost of converting its service to digital. While this factor may have contributed to the diminution of Racing Data's competitive impact, an examination of Racing Data's accounts and information provided by the parties on other possible bidders has indicated that the criteria for a failing firm defence (see note 4) do not all appear to be met. Also, SIS's decision to pay a premium (of £(see note 5)) over book value (of £272,000, shown in management accounts dated 30 September 2002) for Racing Data's assets casts some doubt on whether Racing Data was a failing firm, though the asset value was not large. SIS contended that had Racing Data simply ceased to trade, the change-over from Racing Data to SIS text services could not have been achieved overnight and that without the merger Racing Data customers – who are also customers for the SIS FACTS picture service – would have been left without a text service and unable to operate. SIS suggested that for some smaller LBOs this could have led to them exiting the market. On the basis of the information received on this point it is not possible to reach a clear view whether this could be considered a clear and quantifiable customer benefit (see note 6).

Third party views

These were mixed. While some believed Racing Data to be imposing a competitive constraint on SIS, others felt that the Racing Data service was never a viable option.

Conclusion

SIS undoubtedly has a strong position in respect of the supply of information to LBOs, including racing pictures and commentary, processed text services and the gathering of raw data. Racing Data, which SIS has acquired, supplied LBOs with a processed text service. However, by the time of the merger, Racing Data had a small and declining number of customers. It exerted little competitive constraint in the market for the processing of data about racing (an activity which many LBO chains carry out in-house in any event). The merger therefore appears unlikely to result in a substantial lessening of competition.

I therefore conclude and recommend that you do not refer this merger to the CC.

NOTES

2. BAGS - Bookmakers Afternoon Greyhound Racing.

3. Ladbrokes – 23 per cent, William Hill Organisation – 19 per cent, Tote – six per cent.

4. See discussion of failing firm defence in the OFT consultation paper Mergers: substantive assessment (pdf 208 kb).

5. Figure deleted at the request of SIS.

6. See discussion of customer benefits in the above consultation paper.

Back to top