Everyone familiar with the sports betting industry knows that it is highly influenced by seasonality, due to the annual schedules of most sporting events. Multiple periods of inactivity leave bookmakers with lower revenues during each sport’s off-season. However, there are ways that sports betting operators can actively encourage their players to engage more, even during off-season periods.
Increasing year-round player engagement not only generates additional revenue from seasonal players, but it also enhances brand loyalty, improving the chances that these players will return to the site when betting on their favorite sport next year.
In this post, we present a step-by-step approach to identifying opportunities to entice seasonal players to bet on additional sports during the periods during which they are typically inactive.
Step 1: Discover Each Player’s Sport Betting Preferences
The first step in engaging players during their off-season is to identify their distinct sport preferences. We do this by calculating the percentage of each player’s total wagers in each of the various sports categories, and applying cluster analysis to these calculations in order to discover the distinct clusters within our customer base.
The following table illustrates an example of the outcome of this process for a particular sports betting operator:
Percent of Wagers Placed on Different Sports Categories
Step 2: Explore Cross-sport Correlations
To understand which sporting events might appeal to players outside of their regular seasonal events, we use market basket analysis, a statistical model that measures the probability that players betting on specific sports will bet on others. A market basket analysis can discover, for example, that players who bet on football have a high probability of also betting on tennis, while the probability of them betting on rugby is rather small. This helps marketers understand how to encourage players to engage more during off-peak seasons.
We do this by calculating the lift between each pair of sports. The following diagram illustrates some possible combinations of sports in our example:
We calculate the actual potential embedded in each combination using the lift formula (the actual formula is described at the end of this article):
Lift Calculation for Different Sports Wager Combinations
The lift calculations in this table show that there are specific sports categories with a higher statistical likelihood of appealing to players of other categories. For example, football fans are 7.2 times more likely to bet on tennis games, as compared to any other randomly selected player in the customer database who previously placed a bet. These same fans are also 5.5 times more likely to bet on cricket games than any other random player. This guides us to creating customer marketing campaigns that entice football bettors to try betting on the other sports categories with the highest statistical likelihood of appealing to them.
Of course, diversified operators offering additional products besides sports betting (perhaps poker, casino games or bingo) should calculate the lift between each type of sports bettor and the other games they offer, to discover additional cross-sell opportunities.
Step 3: Sending More Relevant Campaigns and Measuring the Results
Armed with this extremely valuable insight, we are able to create relevant marketing campaigns designed to entice players who have a preference for wagering in certain sports category to expand their betting horizons into additional ones.
To ensure that we’re properly measuring the success of these campaigns, it is important to run them as marketing experiments, comparing the engagement rates of test and control groups, as well as tracking metrics such as churn rates, response rates and player lifetime value.
A Holistic Approach to Player Engagement
By understanding players’ betting habits and the statistical correlations that exist between them and other sports categories, sports betting operators can increase revenues and maximize customer loyalty by encouraging their players to engage outside their normal betting seasons. When players are more loyal to a brand year-round, they are more likely to remain engaged with that brand during their primary sports seasons, and they are less likely to churn.
Appendix: Calculating Lift
The formula we use to calculate the likelihood that two factors are correlated is called lift. In the case discussed in this article, we used the lift formula to estimate how many times more likely would a randomly selected player known to bet on one sport also bet on a second sport, as compared with any randomly selected player from the entire player database:
where:
P1 = Number of players who bet on a first sport
P2 = Number of players who bet on a second sport
C = Total number of the brand’s active players (players betting on any sport)
To learn more, take a look at the Wikipedia article, Lift.
Moshe Demri leads Optimove’s global revenue team and is focused on helping clients optimize their customer retention plans and their use of the Optimove software. Moshe has vast experience consulting clients as a data scientist, analyzing their customer data and revealing actionable, data-driven marketing insights. Moshe holds a BSc in Industrial Engineering and Management, specializing in Information Systems.
