Responsible gambling affects lives everywhere, but operators can’t mitigate the problem without the right tech in place, says Motti Colman, Senior Director of Sales at Optimove. How do predictive models work?
According to research by gov.uk, an estimated 0.5 per cent of the UK population reached the threshold to be considered problem gamblers, and an estimated 3.8 per cent of the population are classified as at-risk gamblers. Potentially, that’s more than 2.5 million individuals.
“The evidence suggests that harmful gambling should be considered a public health issue,” the research concludes, “because it is associated with harms to individuals, their families, close associates, and wider society.”
Gambling-related harm affects people, families, livelihoods, and life itself and has considerable cost to society, likely “in excess of £1.27bn [a year],” the research claims. In recent years, operators in the UK have experienced tightening regulations and enforcement, resulting in social responsibility fines served to operators spiked by more than 100 per cent and surpassing £40m annually.
So the problem gambling issue is here to stay, and, at least from a regulatory standpoint, will be felt more so than in Europe compared to the decentralised US industry, which will inevitably, eventually, catch on. But what does the future hold? Of course, the issue must be raised, knowledge gaps addressed, and awareness boosted, but will it suffice? To get the job done, operators need the right tools to help them meet the challenge. Soon, having a predictive model in place will become a necessity.
In the coming weeks, The UK Gambling Commission will expect operators to “take timely action where indicators of vulnerability are identified.” In other words, operators are expected to put in place predictive models. Can tech play an integral part in fighting the problem?
Predictive models and how they work
A common misconception about responsible gambling is that operators can only react to it, not predict it. That is false, and we should know. Predictive models have been a part of our offering for years.
Predictive behaviour modelling is the science of applying mathematical and statistical techniques to historical and transactional data to predict customers’ future behaviour. The benefits of it are significant. By predicting at-risk players early on, operators can:
- Create daily lists of potential at-risk players to have monitored by account managers or support teams.
- Leverage player’s risk potential to optimise marketing initiatives and reduce the number of players who develop unhealthy behaviours.
- Generate periodic reports and monitor trends.
Responsible gambling predictive models vary but usually include these basic broad lines:
Defining at-risk players – By exploring historical data and developing a definition based on trends. For example, an operator defined their at-risk players by assigning a weighted score, from 1 to 10, to several player activities, such as time spent on site and bonus usage. The higher the weighted average, the more at-risk the player is.
Understanding the data – The more data available for a machine learning algorithm, the more accurate the results.
Choosing the variables – Once the dataset is balanced, the attribute selection process can begin. Selecting the correct variables is crucial for the accuracy and success of the prediction model.
By clustering players into groups based on the gross amount or trendline slope of daily bets made, operators can separate players into two groups – those predicted to become at risk and those with a low likelihood.
Creating the model and analysing results – machine learning algorithm will run and identify players who are predicted to become at-risk. Models can be self-optimising, allowing for changes in player preferences and industry trends to influence the model creation.
Utilizing the at-risk predictive model – segment players into three tiers – low, medium, and high, based on their likelihood to become at-risk. Players with a low-risk level can be given the occasional promotional campaign. In contrast, players in the medium-risk level can receive 30 per cent of the promotional campaigns as the low-risk group received and players from the high-risk level can receive only informative and educational campaigns.
Seatbelts and responsible gambling
Did you know? Until 1966, cars in the UK were often made without seat belts. Many manufacturers offered seat belts as extras to the vehicle. What about the seat belt rule? That came into force only in 1983, yet it seems so logical, natural even. Putting on a seatbelt is a second nature.
The same applies to responsible gambling in the UK and Europe (the US will take its time, but it will get there), and the supporting factors are all in place. A bespoke machine learning algorithm can empower gaming operators to better understand their player base, their trends, and the behaviours players exhibit before becoming at risk. These insights can optimise marketing strategies and improve long-term player retention.
The industry is increasingly aware of the potential damage to its players and business and willingly or otherwise will act. When they do, technology will be ready to play a significant role in tackling the issue.
