Predictive segments

Bonus Efficiency:
Spend the Bonus Budget Where It Changes Play

Many bonuses pay for play that would have happened anyway. Smartico's Bonus Eligibility model checks every player each day and says whether another bonus makes sense. Segment on that answer, and spend your bonus budget on the players it actually helps.

One verdict, four reasons

Every day Bonus Eligibility gives each player one of three answers: Eligible (recent behavior supports another bonus), Suppress Bonus (another bonus is not the right move now) or Unknown (not enough data yet). The answer comes from four signals, each with its own weight:

  • Profitability (35%): does the player win or lose over time? A player who keeps winning costs you margin, so this signal leans to Suppress.
  • Bonus Effectiveness (30%): do bonuses bring extra real-money play, or only replace money the player would have spent anyway?
  • Bonus-to-Deposit Ratio (25%): does the player get more bonuses per deposit than other players on your brand?
  • Bonus Share (10%): over the last 30 days, is the player Self Funded or Bonus Dependent?

A verdict needs at least three signals with enough data. Each signal is also a condition of its own. Add Value Score to see what is at stake, and favorite games to choose what to offer instead of cash.

The models on this page
ModelWhat it answers hereRole
Bonus EligibilityDoes another bonus make sense for this player now?Decides who
Its four signalsWhy: margin, effect of past bonuses, bonus per deposit, own money or bonus money.Decides who
Value ScoreHow much is at stake: the player's worth compared with your other players.Decides who
Favorite gamesWhich game or tournament to offer instead of a bonus.Shapes the offer
Recommended amountsThe player's usual bet: free spins or a free bet at a value that fits them.Shapes the offer

The segments to build

Four audiences cover every player. Make them branches of one journey, in this order: a player takes the first branch that fits. Each is a real-time segment, so the branch is chosen from the verdict at the moment the player reaches it.

  1. Self-funded players of high value

    Who

    Valuable players who play with their own money.

    Bonus Share: Self FundedValue Score (lifetime): Top or High

    What they get

    Status and service, not cash: early access to tournaments, a store item, a personal contact.

  2. A bonus that pays off

    Who

    Players for whom a bonus pays off.

    Bonus Eligibility: EligibleFavorite gameRecommended casino bet

    What they get

    The bonus as planned, on their favorite game. Free spins or a free bet can match their usual bet.

  3. Suppress: keep the contact, drop the bonus

    Who

    Players for whom a bonus is not worth it.

    Bonus Eligibility: Suppress BonusBonus Share

    What they get

    The same journey without the bonus: a free tournament, a mission or news about their favorite games.

  4. Unknown: your usual rules

    Who

    Players the model cannot judge yet, such as new players.

    Bonus Eligibility: no verdict

    What they get

    Your usual bonus rules, as before.

How to know it worked

Fewer bonuses is easy. Fewer bonuses with the same play is the goal. So you measure two things together: what the bonuses cost and what the players did.

  • Give the campaign a control group: Smartico holds back a set percentage of the eligible players when they enter and compares both groups on a conversion rule, such as "deposit approved". A sticky control group keeps the same players out on repeat entries.
  • Compare bonus cost per player and real-money activity (deposits, bets, players still active a month later) in both groups, over the same weeks.
  • Never judge on the verdict itself. More players marked Eligible is not a result.

It works when bonus spend goes down and activity stays flat.

Campaign group vs control group
MeasureWhat you want to see
Bonus costLower in the campaign group
Deposits and betsAbout the same in both groups
Players still activeAbout the same in both groups

Common mistakes

Most bonus-segment problems come from reading the verdict as more than it is.

  1. Treating it as fraud detection.Bonus Eligibility asks whether a bonus pays off for an honest player. Multi-accounting and bonus fraud belong to your risk and payment checks.
  2. Reading Unknown as no.Blocking every player without a verdict removes new and quiet players from offers they would normally get. Let them follow your usual rules.
  3. Cutting the contact with the bonus.A Suppress Bonus player still gets news, tournaments and missions. Only the cash step goes.
  4. Counting only the savings.Lower bonus cost with lower activity is not a win. Read cost and activity in the same report, against the control group.

More goals: reduce churn, VIP growth, reactivation. How the models fit together: predictive segments.

FAQ

How do I stop wasting bonuses on players who would play anyway?
Is Bonus Eligibility a bonus abuse detection tool?
What does Unknown mean in Bonus Eligibility?
What should a Suppress Bonus player get instead?
Does Bonus Eligibility block or grant bonuses by itself?
Can I use the four signals without the verdict?
What data does Bonus Eligibility need?