HC 1554 Gambling

Written evidence submitted by Featurespace (GA 087)


In Featurespace’s view the Gambling Act has not done enough to protect vulnerable people from the adverse effects of gambling. In particular:

· The 2010 Prevalence Study does not provide a detailed understanding of problem gambling in the United Kingdom.

· There have been a number of significant enhancements, including both the volume of gambling data captured and advancements in player behaviour modelling technology, that allow our understanding of problem gambling to be increased.

· Existing requirements around player protection are generally reactive rather than proactive.

· The Gambling Act should encourage a proactive approach to player protection which, in the online environment, aims to feed appropriate information back to the player and in general reduce self-exclusion and rates of problem gambling.

About Featurespace

Featurespace was a founded in 2005, a spin-out from the University of Cambridge that develops software products which provide its clients with highly accurate insight into the behaviour of their online customers. Using the power of Bayesian inference, its award-winning products are used to predict behaviour based on fine grained individual models. The insights it creates are used across multiple business applications, including in the prevention of fraud and understanding customer lifetime value.

Featurespace was founded by CEO, David Excell, and William Fitzgerald, Professor of Applied Statistics at the University of Cambridge. Today, the company employs some of the UK’s brightest young software engineers and data analysts and has grown by 125% over the last year.

Recently, David Excell was named ‘ITC Enterprise Young Entrepreneur 2011’ at the ITC Entrepreneurs’ Forum Enterprise Awards. At the EGR B2B Awards 2011, Featurespace won the ‘Services Rising Star of the Year’ award and was highly commended within the ‘Fraud & Compliance Solution of the Year’ category. David has also won the Growing Business ‘2011 Young Guns’ award.


Q: How effective the Act has been in its core objectives to:

- ensure that gambling is maintained crime-free and conducted in an open and fair manner,

- protect children and vulnerable people from the adverse effects of gambling,

- update the legislative framework with regards to online gambling;

Since the implementation of the Act, the 2010 Prevalence Study commissioned by the Gambling Commission showed that there has been an increase in both the number of people participating in gambling activities and number of people who scored highly for problem gambling. The statistical confidence levels in the problem gambling results are not significant enough to conclude that overall the rate of problem gambling has increased within the United Kingdom. It is still unclear if the Gambling Act has achieved or failed in one of its core objectives in protecting vulnerable people from the adverse effects of gambling.

Over the last four years Featurespace has worked almost exclusively within the online gambling industry gaining significant insight into how players gamble and interact with online sites. Featurespace has also reviewed how a majority of operators implement player protection functionality within their products.

As required by existing license conditions player protection controls implemented by operators are predominantly reactive. Players are able to specify limits (such as time, or amounts) or exclude themselves before an operator will restrict their gambling. Proactive problem gambling controls are limited to awareness, education and ‘self-excluders’ who attempt to create a new account on a site which they’ve previously excluded themselves from.

Featurespace believes that these controls can be significantly enhanced to further protect vulnerable people from the adverse effects of gambling. Vast volumes of gambling data is captured and analysed on a regular basis within most online gambling environments but almost no effort within this process is put towards the protection of players.

Behavioural analytics is a rapidly emerging technology that learns and predicts a player’s activity highlighting specific patterns or abnormal actions. Specific patterns are learnt from past customer behaviours and abnormal actions are identified by comparing the current player activity to a statistical representation of their past behaviour. Featurespace has implemented its own behavioural analytics platform which is able to model behaviour at an individual level. Traditionally analytics learns the patterns that occur within large groups or segments of customers. To accurately identify players at risk of problem gambling Featurespace believes individual models are necessary as they allow abnormal or at risk play to be highlight by comparing the current player’s activity with their own past activity.

For an effective proactive player protection solution knowing which players are at risk is only half the problem. The output from the individual player models must be used to coordinate the messaging which is displayed to the customer or control the prominence of particular gambling features. For example, for a low risk customer, it may be appropriate to re-enforce the presence of deposit limits, whereas for a high risk customer, enforcing session limits may be appropriate.

Industry wide support is still required for on-going research to validate which individual player changes truly represent early indicators of problem play. The industry also needs to develop a code of practice on how and when players at risk of problem gambling should be communicated with.

We believe that it is in both the government and operators best interests to substantially improve player protection functionality. For governments it reduces the cost of rehabilitating players who have developed gambling problems and for operators it builds a larger base of long-term, profitable customers.

Q: the financial impact of the Act on the UK gambling industry;

Featurespace has no view on this question.

Q: the effectiveness of the Gambling Commission since its establishment, and whether it represents good value for money;

Featurespace has no view on this question.

Q: the impact of the proliferation of off-shore online gambling operators on the UK gambling sector and what effect the Act has had on this

Featurespace has no view on this question.

Q: why the Act has not resulted in any new licences for casinos or "super" casinos;

Featurespace has no view on this question.

Q: the effectiveness of the classification and regulation of gaming machines under the Act;

Featurespace has no view on this question.

Q: what impact the Act has had on levels of problem gambling.

Featurespace has no view on the Acts effect on the level of problem gambling. Featurespace believes that the Act could do more to reduce problem gambling as stated in our response to the first question.

October 2011

Prepared 20th October 2011