Exploring the inner workings of Ibex

Connor Coutts
February 2, 2023
Data Science
Back to blogs
Ibex is an AI retention tool that is designed specifically for the iGaming industry. Its purpose is to help companies make informed decisions about how to best retain and engage their players.
At its core, Ibex works by analysing player data and making predictions about future behaviour. It then uses this information to determine the best action to take for each player at any given time.
An action, as defined by Ibex, is a combination of a communication template (such as an email, SMS or push notification) and an optional promotion. There is also a special action called the “Do Nothing” action, where no offer is given and no template is sent. This serves as the baseline against which all other actions are measured.
The goal of Ibex is to pick the next best action for each player at each point in time. The “next best” action is the one that is expected to generate the highest future income from that player. However, it is important to note that this doesn’t always mean sending an offer or promotion. Sometimes, the best action is to do nothing at all.
But what components are involved in this decision-making process? The decision-making process used by Ibex is broken down into several steps:
  1. Eligibility Matrix: This step determines which players are eligible for specific offers or promotions based on their behaviour and demographic information.
  2. Deposit Rate Model: This step predicts the likelihood of a player making a deposit in a short-term window after receiving an offer.
  3. Deposit Amount Model: This step predicts the amount a player is likely to deposit after receiving an action.
  4. GGR prediction: This step takes the estimates from the deposit rate and deposit amount models to predict the revenue a player will generate.
  5. Bonus Cost Model: This step estimates the cost of offering a bonus to a player based on the bonus mechanics.
  6. Other cost estimates: This step estimates any other costs associated with bonusing a player such as jackpot contributions, revenue share deals or game supplier fees.
  7. Lifetime value model: This step estimates the future value the player will generate (over 180 days) after an action is sent.
  8. Experimentation module: This step enables Ibex to experiment with a small number of offers daily to continuously optimise its recommendations.
  9. Control Groups: This step randomly assigns players (with a probability of ~20%) to control groups in order to asses the impact Ibex is having.
Each of these steps plays an important role in the decision-making process used by Ibex. In later blog articles, we will delve deeper into each of these steps and how they work together to help iGaming companies make the best decisions for their players.

Grow your business

Use Ibex to drive customer engagement and deliver the right messages at the right time
Start Now