Beating Addiction with Big Data

In a previous post, I wrote about the various applications of big data insights in the gambling industry, from tailored marketing initiatives to odds calculation. However, the very same data used to great commercial effect by gambling companies could have another utility: protecting customers from addiction.

Gambling brands are under increasing pressure to either relinquish their data insights to independent regulators or take internal measures to protect their customers. But what exactly is being proposed? And how precisely can big data analysis be used to combat problem gambling?

New research and innovative safeguards

A recent Wall Street Journal article explored the efforts of addiction scientists and industry consultants to spot ‘high-risk’ players through analysis of customer-tracking information. Working in collaboration with government-run casinos in the U.S., the new research “turns the industry’s own data, often used in connection with loyalty cards to identify and pamper the best customers, on its head.”

Project leader, Prof. Sarah Nelson of Harvard Medical School, presented her teams’ research at a conference in Ceasars Palace in Las Vegas. The fruit of their labours has been dubbed ‘Sports Bettor Algorithm 1.1’ (Risk Level = .134*LNfreq + 0.793*LNbpd). The algorithm (which was eight years in the making) draws on several variables – including the size and frequency of bets – to spot the hallmarks of nascent pathological behaviour in players.

Only a very small percentage of individuals can be described as ‘pathological gamblers,’ but for that minority, compulsive gambling can be highly destructive. Prof. Nelson explains that, by processing a casinos’ behavioural tracking data with Sports Bettor Algorithm 1.1, her team can accurately predict a customer’s propensity for addiction, allowing casino managers to intervene.

Invasive overkill or social responsibility?

As you would imagine, the suggestion that casinos’ own behavioural tracking data should be employed to curb excessive gambling has enjoyed a mixed reception from the industry. Some companies have flat-out rejected the idea as antithetical to their business practices and a violation of players’ personal freedoms.

“I think it’s a terrible idea,” said Gary Loveman, CEO at Caesars Entertainment Corp. and a former Harvard Business School professor, who pioneered data mining for marketing purposes in the gambling industry. “Is it McDonald’s obligation to decide you have a problem because you have a tendency to eat high-calorie lunches? You could take this to ridiculous extremes.”

Other casinos been more amenable, particularly those run by state bodies. Two government-run casinos in Saskatchewan, Canada, used a similar system – named ‘Focal’ – for seven years. As with Sports Bettor Algorithm 1.1, Focal mines variables from big data sets to identify excessive gambling on the casino floor. When the system detects a problem, it sends an alert to casino staff with the player’s location.

The system triggered about 2,900 such interactions in 2012 out of 70,000 active players-club members. However, unlike Prof. Nelson’s algorithm, Focal cannot predict addictive behaviour; only identify it in existing players.

Saskatchewan casinos stopped using Focal in 2013 but a New Zealand operator in July agreed to be the first to use it in a commercial land-based casino, in return for the government’s permission to expand its operations.

Safety online

Researchers admit that big data analysis is an unreliable effective tool for combating addiction in land-based casinos, because data sets only come from customers who opt in to loyalty cards. A major breakthrough would have to come from online casinos, which collect reams of behavioural data from millions of customers.

In this interview, addiction specialist Dr Mark Griffiths noted that preliminary efforts to enforce betting limits on ‘high risk’ internet gamblers have shown promise. “The great thing is we’ve just done some research and we got a player database of over a million online gamblers,” said Dr Griffiths. “We found that these limit setting features do work with the people that most need them i.e. the most gaming intense people.”

Swedish casino Svenka Spel has already introduced mandatory, behaviour-tracking ‘debit cards’ that must be used for transactions at its online operations. Furthermore, with online gambling slowly being reintroduced to the United States, it is possible that more internet casinos will attempt to bolster their public profile by putting big data analysis in service of ‘social responsibility.’

888 Holdings has recently launched online gambling operations in Nevada and Delaware and has intimated that it will be the first internet casino to employ algorithms to identify problem gamblers. Not far behind, Bwin.Party (which has applied to operate in Nevada and New Jersey), is planning to roll out player protections based on big data analysis. Proposed interventions include a pop-up screen to warn players when they have been gambling for too long or exceeded their typical betting limits.

There’s a very strong negative business agenda attached to problem gamblers,” said 888 CEO, Itai Frieberger. This view was echoed by Bwin’s responsible gaming manager, Joachim Haeusler, who added that such measured create more sustainable customers. “[Addicted gamblers are] bad for our reputation and bad for business. A player who gets into trouble is a lost customer.”

The house always wins?

Despite these positive noises, some industry figures are more sceptical. CEO of Castle Casino David Merry is concerned that the majority of casinos will mine big data sets to target, rather than protect, problem gamblers.

“For larger companies, it makes sense to turn away addicts – it’s basic reputation management,” Merry said. “Sadly, for every Castle and 888 there are 100 mid-tier companies seeking quick dividends. When you don’t have much of a reputation to protect, it might be tempting to market more aggressively to players who gamble heavily in the short-term, even if they have a limited shelf-life.”

It remains to be seen whether the majority of casinos will use big data analysis to identify and assist problem gamblers, or simply bombard them with marketing materials and tailored bonuses. At the very least, if online casinos hope to claw back the American market, they must make conspicuous efforts to promote their respectability. Tools like Sports Bettor Algorithm 1.1 and Focal could be their saving grace.