Short answer: acquisition and retention offers have different jobs, but there is no portable wagering multiple or bonus-cost band. Terms must comply with the exact licence and promotion rules; economics must be modelled on eligible cohorts with bonus, tax, fee, abuse and net-revenue definitions declared. Low- or no-wager retention value may be useful where permitted, but only behind explicit controls and a pre-registered test.
Part of the Tier-1 Operations Canon. The series opener put "bonus restraint" first among the disciplines that separate tiers; this article unpacks the measurement architecture behind that restraint. It is not a cross-market schedule of acceptable terms.
Two wager regimes, two jobs
The most common structural mistake in bonus design is treating wagering as one dial. Tier-1 programmes split it:
| Acquisition layer | Retention layer | |
|---|---|---|
| Instruments | Deposit match, free-spin packs, descending ladder to the 3rd deposit | Weekly cashback, loss-back, rakeback, wheel prizes |
| Terms | Exact licence cap, eligible games, expiry and disclosure checked before launch | Low or no wagering only where permitted and economically modelled |
| Job | Acquire an eligible cohort without hiding the effective value | Test whether targeted value changes repeat behaviour incrementally |
| Failure mode | Opaque or unlawful terms, adverse selection and misleading headline value | Arbitrage, margin loss or subsidising behaviour that would occur anyway |
In Great Britain, the UK Gambling Commission caps wagering requirements on bonus funds at x10 from 19 January 2026. Other markets require their own current primary-source check.
The welcome pack is an experiment, not a gift
A welcome offer changes who converts as well as how many convert. Measure eligible impressions, opt-in, deposit, completion, abuse review, repeat deposit and net contribution by pre-defined cohort. A ladder across early deposits is a hypothesis to test against a simpler offer, not a universal prescription. FTD-to-second-deposit conversion is useful only when organic behaviour and offer exposure are separated.
A bigger welcome cannot be assumed to fix retention. If repeat deposit falls, inspect product, payments, support and CRM before changing terms; otherwise the programme can increase cost while subsidising the same users.
Low wagering works only behind explicit controls
Cashback and loss-back with low or no wagering can make the value clearer where rules permit it. They also increase abuse exposure. Before launch, document the wagering term, eligible games, bet and cashout limits where lawful, expiry, stacking, duplicate-account controls, source of funds, manual-review triggers and player-protection overrides. The terms must be clear to players and reviewed against the exact licence.
Game contribution and exclusions must come from the approved promotion design, not copied industry percentages. Record versioned rules, expected liability, realised cost and false-positive review outcomes; keep an explicitly monitored abuse cohort without presenting it as a reason to weaken responsible-gambling controls.
Bonus cost as a share of revenue: define the control states
The steering metric is not "bonuses issued". Bonus cost can be compared with GGR or net gaming revenue only after definitions for cash value, wagering conversion, taxes, fees, fraud, reversals and cohort period are fixed. Set thresholds from the deployment's approved budget and observed variance:
| Control state | Required evidence and action |
|---|---|
| Planned | Expected liability, eligible cohort, duration, stop rule and net-contribution model approved before launch |
| Warning | Observed cost or abuse rate leaves its pre-registered control band; inspect mix, caps, formula and tracking |
| Stop / rollback | Hard budget, legal, player-protection or data-quality condition is breached |
Publish the denominator, window and exclusions beside every ratio. A percentage without them is not reusable evidence.
The threshold discipline has a per-player corollary: exposure caps. Without a ceiling on bonuses per player, a handful of heavy accounts can quietly absorb most of the programme's budget and drag the ratio into the abuse band while the averages still look calm.
Generosity without segmentation is a margin leak
Every rule above is really one rule: a bonus is a targeted, testable intervention. Loss-back, calendar reloads and early-lifecycle offers need a holdout or credible comparison cohort before anyone claims incremental repeat deposit. Player-protection signals override commercial targeting. Restraint without segmentation is blunt; generosity without measurement can leak margin. Evidence is what turns either into strategy.
Where the discipline runs as software
Operationally, the canon is a checklist: every bonus passes a pre-launch anatomy review (trigger, caps, wagering, game weighting, expiry, stacking rules), the programme reports cost as a share of GGR against the thresholds, and per-player exposure is capped and monitored. On our platform that logic ships in the reporting stack — bonus economics land in the nightly brief as a share of GGR, and the win-back engine applies the same discipline to retention offers, pricing each one against expected value before it is sent rather than after it is regretted.
Frequently asked questions
Why do Tier-1 operators keep welcome wagering requirements high?
There is no universal high-wager standard. Terms must comply with the exact licence and be tested for eligible conversion, abuse, repeat deposit and net contribution. In Great Britain, the UK Gambling Commission caps wagering requirements on bonus funds at x10 from 19 January 2026.
When can low wagering be used on reloads or cashback?
Low- or no-wager value can be clearer where permitted, but its incremental effect must be tested against a holdout or credible comparison cohort. It requires explicit terms, liability limits, abuse controls and player-protection overrides.
What abuse controls make low-wager bonuses safe?
Document the wagering term, eligible games, lawful bet and cashout limits, expiry, stacking, duplicate-account controls, manual-review triggers, liability cap and player-protection overrides. Version the rules and audit realised cost and false positives.
What share of GGR should bonus cost take?
No portable percentage is evidence-safe. Define cash value, wagering conversion, taxes, fees, fraud, reversals, denominator and cohort period; then set warning and stop bands from the deployment's approved budget and observed variance.