This news of the largest fine since the Gambling Commission got the power to do so sends a chill through the European gambling industry. The Commission has been working with and warning all gambling companies that they have to do more, and be seen to do more, to prevent problem gamblers from operating on their online platforms.
The fine points to a number of issues that are prevalent in the gambling space, problems that we at Ravelin believe can be tackled through applying approaches that are being used today in different industries.
There is no doubt that due to the work of the Gambling Commission there has been significant improvement in honouring and tracking self-exclusion. However 888.com ruling states,
‘While 888 did have self-exclusion procedures in place, they were not robust enough and failed to protect potentially vulnerable customers.’
Graph Network and Machine Learning combine
We recently had the opportunity to present to a group of fraud experts a product for the gambling industry called Ravelin Connect.
There, we proposed that gambling is uniquely well-placed to take advantage in graph network and machine learning techniques that seek to highlight connections between accounts and to highlight those connections that might be of concern.
There are multiple use cases for this kind of technology. We have written in TotallyGaming.com about its applicability in bonus abuse. Bonus abuse is of course tightly linked with account creation abuse - and again, that is tightly connected to problem gambling. In short the capability of being able to highlight connections or links across multiple parameters helps tackle many of the most pressing issues in account management for the gambling sector.
You could argue that this is simply link analysis. And in a way you are right. But what's changed with Ravelin Connect is the speed, breadth and scale at which this analysis is done. And the ease with which the data can be interpreted in the dashboard. And the ability of Ravelin Connect to proactively highlight suspicious connections. based on the historical experience of your company. In fact, it's fairer to say that everything has changed about it.
Instead of trying to see connection across single vectors like similarity of emails, or device identification, Ravelin Connect opens up possibilities akin to multi-dimensional chess. As we have pictured below we can connect nodes on multiple vectors giving rich detailed maps of the connections in a merchant’s user database.
Note that there is no requirement here for the user’s gambling history or usage. The onsite and betting behaviour is well tracked in the gambling industry. Instead Ravelin Connect looks at the harder to spot relationships between accounts.
Recent Maltese Gambling Commission rules have mandated the elimination of duplicate accounts. This is a difficult problem to solve with a single dimensional view. But if two accounts share a device ID, phone number, physical address, IP address, voucher code or a payment instrument, then it is easier to flag. Or if any of those are then shared with a third, fourth or any number of users then it is easy to see a network of duplicate accounts. In some industries we've uncovered thousands of accounts set up with fraudulent intent, whether it is use stolen payment credentials or to abuse bonus schemes.
An interesting part of the commentary from the UK Gambling Commission was the requirement for gambling businesses to demonstrate how they were working to combat the issues of duplicate accounts and problem gambling. Ravelin Connect offers that auditing capability. In addition to flagging and/or preventing multiple accounts it is possible to show anyone relevant the validity of any account. Where issues are raised, perhaps by a gambler, it is possible to review that account and therefore allow legitimate betting activity to continue.
The fine for 888.com highlights that the complexity of online identities requires newer and more flexible tools in order to manage them especially for gambling. There are mini-industries emerging focused on bonus abuse and self-exclusion abuse and managing their threat through in one or two dimensions is either too resource-intense or ineffective. The correct application of the right technologies has the potential to release the strain on fraud and payments teams by highlighting areas of contention long before they come to the eyes of the regulators. Reputations, revenues and risk can all be managed.