Here are the key terms used in Amplifier AI. We'll be updating this as new features are added.
Anything missing? Please let us know with this form.
The total number of active users who have placed at least one bet in the selected time period. These are users who have placed bets before the selected time period.
This reason code indicates the percentage of users who were predicted churn due to losing a near win bet. A bad beat is defined as a lost single bet with odds lower than 1.25 or a lost accumulator that has more than 3 selections with only one losing selection.
Bets/ Stake/ Wager
Daily Bets: median daily player stake on a game
Total Bets: sum of stakes
Bonus spins are assigned to players to encourage more spins. It is the same as Promotional Free spins
This reason code indicates the total number of churn predictions that were sent to the operator via the API in the selected time period.
Churn Prediction % - Active Users
This indicates the percentage of users who were at risk of churning in comparison to all the users who have placed bets in the selected time frame.
Churn Prediction % - Settled Bets
This indicates the percentage of users at risk of churning from the total number of users who had a settled bet and no active bets.
Future Anthem splits players into groups using behavioural clustering (unsupervised learning models). Player behaviour is assessed across a variety of attributes including daily game sessions, total spins, total stakes, bonus stakes, and session length.
Anthem uses this method to cluster players by game, by day, to provide a consistent framework, and to compare games.
This reason code indicates the percentage of users who have three or more consecutive losses.
Daily Active Players (DAP)
Daily distinct player count
Gross Gaming Revenue
Formula: Total Stakes – Total Payouts
High-Value players do not usually play as many spins as High-Value Multi Session clusters, however, they place bets of similar average size. They contribute the most to total GGR and NGR due to their sheer player volume, and usually have second highest RTP
High-Value Multi-Session Clusters
High-Value Multi Session players tend to have the highest number of rounds, sessions, highest bet value, and highest RTP. They are termed multi-session for playing multiple sessions of the same game per day
Formula: (spins with payout > 0)/ total spins
Key Performance Indicator
A quantifiable measure of performance over time for a specific objective. For example, revenue.
Low-value players have the lowest contribution to all the metrics compared to other clusters
Mid-Value Multi-Session Players
Mid Value Multi-session players play fewer rounds than High-Value clusters, place smaller daily bets, have smaller session lengths and less RTP than the High-Value clusters and play multiple sessions of one or more games.
Mid-Value players play even fewer rounds than Mid-Value Multi-Session players, place smaller bets, or have similar session lengths but usually play one session for the respective game.
Monthly Active Players (MAP)
Monthly distinct player count
Month Till Date (MTD)
Month Till Date (MTD) stands for values on the latest day of the current month. For example, a column for GGR Month Till Date will show the GGR values accumulated until the current day of the month.
Most Player Increases
This visualisation in Optimise lists top 5 operators in terms of increase in daily average players on a week on week basis, for the last 7 days.
Formula: GGR – Promotional Free Spins Payouts
New Active Users
Total number of new users who placed their first bet in the selected time period
New Players: players within their 10 days of tenure (current date – registration date)
Retained Players: all the rest
New Registered Users
Total number of users who have registered in the time period. These are users with newly created accounts who did not place any bets in the selected time period.
This page in Optimise allows studios to see the total bets placed for each game, for each operator. Values are only shown for games that have been released.
Other (reason code)
The percentage of users for which churn was predicted and no clear reason code could be identified for churn.
There are many different classifications of players, the is usually unique players as opposed to daily average players.
Player Retention Rate
Retention rates are derived from a weekly cohort analysis. The analysis considers the return of players on the same game, from the first time the player plays a game.
Standard retention KPIs shown in the report are:
- W1 retention: average retention rate 1 week across all cohorts
- W4 retention: average retention rate 4 weeks across all cohorts
- W8 retention: average retention rate 8 weeks across all cohorts
Promotional Free Spins
Total promotional free spins stakes
Real Money Bets
Total Bets placed without considering bonus bets.
Total spins without counting bonus spins.
Stakes placed without bonus stakes.
Every churn prediction provides a contextual reason code to enable marketing teams to build tailored player messages, they include:
- Reduction in spend
- Reduction in frequency
- Reduction in activity
- Significant loss
- Bad beat
- Consecutive losses
Reduction in Activity
This reason code indicates the percentage of users who were predicted churn due to a reduction in their betting activity. This reason code is triggered when the number of bets in the last 7 days is lower than the number of bets in the 7 days prior to that (day 7 - day 14)
Reduction in Frequency
This reason code indicates the percentage of users who were predicted churn due to a reduction in their betting frequency. The reason code is triggered when the number of active days (days bet placed) in the last 7 days is less than the number of active days in the 7 days prior to that (day 7 - day 4)
Reduction in Spend
This reason code indicates the percentage of users who were predicted churn due to a reduction in their spending. Their total stakes in the last 24 hours were less than their average daily stake.
Daily Rounds: median daily player spins on a game
Total Rounds: total number of spins
Rounds per Player
This is calculated using using ‘daily average players’. At present this value is a median.
Return to Player (actual)
Formula: Total Payouts /Total Stakes
Game-session, i.e. if a player plays three different games consecutively we classify them as three different sessions. For the purpose of reporting, the sessions end after 30 minutes of inactivity.
Daily sessions: median daily player sessions on a game
Median player session length
This reason code indicates the percentage of users who were predicted churn due to a significant loss. A significant loss churn prediction is triggered when the amount lost by the user in his last bet is 50% or higher compared to the maximum of the 5 previous losses.
In Optimise, we observe certain KPIs by their volume within specified ranges of spins, or bands of spins. For example, total spins can be arranged into the following bands: <5. 5-20. 20-50. Spin banding allows us to observe how much stakes were placed for these bands for each game.
Same Period Last Month (SPLM)
Same period last month. For example, the GGR SPLM column will show GGR values on the same day of the month, but last month.
In Optimise, we observe certain KPIs by their volume within specified ranges of stakes, or bands of stakes. For example, total stakes can be arranged into the following bands: < £5, £5-20, £20-50. Stake banding allows us to observe KPIs, such as the volume of players who placed bets, within these ranges.
Tourists contribute the least in terms of NGR and GGR, playing on average one session, placing the smallest daily bets, and tending to have the lowest RTP. But tend to be slightly higher than Low-Value players.
Turnover per player
This is calculated using ‘daily average players’. At present this value is a median.
Users with Settled Bets
The total number of users who had at least one bet settled in the selected time period and had no active bets after the bet settlement.
Player distinct count
Formula: (spins with payout > stake)/ total spins