Tuesday, 2 February 2010

Why turnover is an irrelevant metric

Ralph Topping, whose latest diatribe I blogged about earlier today, is not the first person to have quoted Betfair's "turnover" in arguing, bogusly, about Betfair's relative tax or levy payments. The whole of the gambling industry is taxed (and levied) on gross profit; and gross profit = margin x turnover. Therefore, an operator with a turnover of 200 and a margin of 10 will be taxed on the profit of 20; just as an operator with a turnover of 1000 and a margin of 2 will also be taxed on the profit of 20.

However, because it comes up so often, and is still being wheeled out for current debates, I thought some might be interested in learning more about what our 'turnover' number actually means. For those who don't want the long version below, the short version is "not a lot".

Q. How does Betfair measure turnover?

A. The most commonly used metric for turnover is the ‘total matched’ amount which can be found on the site for every market. This figure represents the sum total of the backers’ stakes, doubled. A back bet of £100 makes £200 in total matched turnover, irrespective of the odds.

At even money (decimal odds of 2.0), it is therefore the case that the stated turnover is the same as the amount being wagered by each party, but where odds are extreme, this far from the case: if the odds are 10/1 (11.0 in decimals), then a punter wishing to win £1,000 need only bet £100, and turnover would still be expressed as £200 – double the backer’s stake. If the odds are 1/10 (1.10 in decimals), then a backer wishing to win £1,000 would have to bet £10,000 – making the turnover for the same win figure £20,000.

Applying this turnover metric to an individual user, that user’s total matched turnover on a given bet will be the backer’s stake doubled, irrespective of whether the user in question is the backer or the layer.

Q. What makes Betfair’s (and its users’) numbers so big?

A. Betfair charges its users a margin (commission) based on a user’s net profit per event. In contrast, a traditional bookmaker builds his margin into the prices he offers the punter. This means that a punter is paying the bookmaker’s margin on each and every bet he places with that bookmaker.

Therefore a Betfair punter can make several bets on a market, but will only pay a margin on his overall net profit position on the market in question. This encourages the Betfair punter to make multiple bets on a single market.

A punter with £100 in his account could lay a horse at odds of 6.0 for £20, thus taking up all the funds in his account. However he can now lay another horse in the market for the same odds and same stake, because he is allowed to keep betting as long as his exposure on the market has not gone over the £100 he has in his account. In this instance he is increasing the turnover on the exchange but doesn’t have to use more funds to do so. In contrast, every bet with a traditional bookmaker not only incurs a margin for the punter, but also requires him to fund each bet separately.

Separately, the ability to back for and against (or ‘lay’) every outcome, means that an exchange punter can bet on the movement in a price (so called ‘trading’). On the limited occasions that the ability to back for or against an outcome is available with a traditional bookmaker, the punter again pays a margin on each bet (this is also the case with spread betting) which makes ‘trading’ a price less attractive.

Betting on the movement in the price of an outcome allows a user to generate a large turnover, whilst taking very little risk. We once had a situation where a customer 'turned over' £246,000 on a rugby match without ever having more than £400 at risk. He ended up winning £150, which would have been roughly what he'd have got if he had placed a single bet on the right team to win.

Q. What relationship does this total matched turnover amount have to Betfair’s profits?

A. Betfair’s profit on an event is generally a very small percentage of this turnover figure, but it varies depending on the nature of the event. Two extremes are: a 5 day test cricket match which will typically generate significant turnover, but a very small yield for Betfair; and a 30 second greyhound race which will give a much higher yield based on a much lower turnover amount. The cricket match by its nature allows users to move in and out of positions as the game gradually develops, whereas the market on a typical greyhound race will only form in the minutes before the race and the race, will not be covered ‘in-running’.

Q: Is any metric of turnover on an exchange comparable with what a traditional bookmaker would consider to be his ‘turnover’?

A. Not really. A traditional bookmaker’s turnover is the sum total of all the back bets (whether backing for or backing against an outcome) that he accepts. The respective pricing mechanisms of the traditional and exchange platforms (punters being charged on a per bet versus per market basis), together with activity on an exchange which could not be done through a traditional bookmaker, means that any comparison between a turnover figure from a traditional bookmaker and an exchange, is like comparing apples and oranges.

Q. Does an analysis of turnover in some form have any benefits?

A. It is difficult to see any. The arbitrary nature of betting exchange turnover, and the flaws in treating it as a meaningful number are highlighted by the following:

1. Different exchanges calculate the ‘total matched amount’ differently. As described above, Betfair doubles the backer’s stake. However, Betdaq takes the sum of the backer’s and layer’s stakes. £100 bet at digital odds of 11.0 gives £200 of turnover on Betfair, but £1,100 of turnover with Betdaq.

2. A small punter can generate huge turnover with just a nominal amount in his account. A Betfair punter needs to collateralise his maximum liability on a market. However, if he never puts himself in a position where this liability is more than say, £10, then irrespective of the number of bets he places and the turnover he generates, he needs no more than £10 of funds in his account.

3. If a user wants to cancel an unmatched bet on Betfair he can do one of two things: he can click the ‘cancel’ button; or he could self-match – i.e. he could accept his own offer, giving him no exposure on the market and freeing up the original funds which had until then been unmatched. Cancelling his bet obviously results in no turnover, but self-matching will cause a total matched turnover of double the punter’s back stakes. Logically a punter will cancel rather than self-match, but the fact that an alternative which achieves the same result can have such a radically different impact on the turnover number, further undermines that relevance of that number.

Turnover, however defined, only generates profit for the operator if it provides punter loss. Betfair believes that the best way to maximise this over the lifetime value of the customer is to offer a low margin offering which does not penalize the punter for making multiple bets on a market, or otherwise increasing the range of betting opportunities for the punter.

The only meaningful metric is total punter loss. This provides the bookmaker’s (exchange or traditional) profits, and as a consequence the betting duty yield.


  1. Mark

    As far as I was aware, self matching does not increase the total matched figure on a Betfair market. I just tested to confirm, and there was no increase in the total matched figures.

    But I get your point.


  2. You're quite right, TG. It used to be included, be changed it because in the rare cases where people self-matched it was artificially inflating the number (as described here) and also people were messing up the price graphs by self matching at daft prices when a market opened. I copied and pasted this from a past briefing note because it was relevant to the comments made by Ralph, and I didn't notice on a quick skim-through that it was written before we made the change. The rest of it still stands, though.