Three beat nine in every cohort we tracked: when a lobby’s search-and-filter panel surfaced three default sort options rather than nine, the share of players who made a deposit by their 45th session rose by 4.1 to 6.8 percentage points, and the effect held after controlling for bonus presence, game mix, and device. The finding comes from a pooled analysis of 2.3 million anonymized sessions across four U.S.-facing operators between January and September 2024, and it runs against the common assumption that more visible choice helps discovery. In this dataset, restraint at the point of search did more for the session-45 deposit metric than any promotional lever we tested.
What the session-45 deposit metric measures, and why it matters more than day-one conversion
Most operator reporting stops at first-deposit conversion or 30-day retention. Session-45 deposits capture something different: whether a player is still finding games worth paying for after roughly six weeks of intermittent use. It is a lagging indicator, which makes it harder to game and more useful as a diagnostic.
The metric is defined here as the proportion of newly registered players who place at least one deposit on or before their 45th recorded session. Median time to session 45 in the sample was 38 days, with an interquartile range of 24 to 61 days. Players who reach session 45 without depositing rarely convert later; in the same dataset, only 7.9% of that group deposited within the following 90 days.
That asymmetry is the reason the search-filter finding carries weight. A lever that moves session-45 deposits is moving players who are already engaged enough to keep returning; it is not inflating a vanity count at registration.
The default-count experiment: three versus nine
The comparison was not a clean A/B test across all four operators, which is a limitation worth stating plainly. Two operators ran a controlled split on 60% of traffic; the other two supplied observational data from a staggered rollout in which the filter panel was reduced from nine default options to three over a six-week window.
What changed in the interface
The nine-option default included provider, volatility band, RTP range, theme, max-win tier, feature type, release year, jurisdiction of license, and a "popular now" sort. The three-option default kept provider, theme, and "popular now," and moved the remaining six behind a "more filters" disclosure.
No game inventory changed. No bonus terms changed. The only variable was what appeared by default when a player opened the search panel.
The results
| Cohort | Defaults | Session-45 deposit rate | Delta vs. control |
|---|---|---|---|
| Controlled, Operator A | 9 → 3 | 18.2% vs. 22.8% | +4.6 pp |
| Controlled, Operator B | 9 → 3 | 16.9% vs. 23.7% | +6.8 pp |
| Observational, Operator C | 9 → 3 | 14.4% vs. 19.1% | +4.7 pp |
| Observational, Operator D | 9 → 3 | 15.8% vs. 20.0% | +4.2 pp |
Operator B’s larger effect coincided with a higher share of mobile sessions (71% versus 58% at Operator A), which is consistent with screen-constrained search benefiting more from fewer defaults. We treat that as a hypothesis, not a finding.
The aggregate effect, weighted by traffic, was +5.3 percentage points, or a 31% relative lift over the nine-default baseline.
Why fewer defaults appear to help: three mechanisms worth testing
The data cannot isolate a single cause, but three explanations fit the pattern and are testable.
Search abandonment falls. Players who opened the filter panel under the nine-default condition abandoned the panel without applying any filter in 34% of cases, against 21% under the three-default condition. A panel that presents too many axes invites the player to construct a query they never finish.
Applied filters are more likely to be decisive. Under three defaults, 62% of applied filters narrowed results to 40 or fewer titles. Under nine defaults, that figure was 39%. A filter that leaves 300 results has not helped the player choose; it has only reframed the lobby.
Session depth redistributes toward play. Median games launched per session rose from 4.1 to 5.3 in the three-default condition, while median time in the search panel fell by 19 seconds. The player is spending less time configuring and more time playing, which is the behavior the deposit metric ultimately reflects.
A fourth explanation — that fewer defaults simply reduce cognitive load for novices and do nothing for experienced players — was not supported. The effect was present, though smaller, among players with 20 or more prior sessions on the platform.
Caveats, and what would falsify the claim
The observational cohorts are the weak point. Operators C and D changed their filter defaults during periods that also included seasonal slot releases, and we could not fully separate inventory effects from interface effects. The controlled cohorts are cleaner but smaller: roughly 410,000 sessions combined.
There is also a novelty risk. The measurement window ran 45 sessions from each player’s first session, so players exposed to the three-default panel late in the rollout had less calendar time under the new interface. We corrected for this by indexing on session count rather than days, but the correction is imperfect.
The claim would be falsified if a preregistered replication found the effect reversing under a four- or five-default condition, which would suggest a threshold rather than a monotonic relationship. We did not test intermediate counts, and that is the most obvious gap in the design.
What the finding implies for lobby design, and an open question
If the effect replicates, the operational implication is uncomfortable for product teams that have spent years adding filter dimensions as a proxy for sophistication. A search panel is not a menu; it is a decision aid, and decision aids degrade when they present more axes than a player can resolve in one pass. The practical move is not to remove filters but to demote them — keep provider, theme, and a popularity sort in view, and let volatility band or RTP range live one tap deeper for the players who actually use them.
The open question is whether the same logic applies to the lobby itself, not just the search panel. Lobbies have drifted toward nine or more visible rails — new releases, jackpots, live dealer, crash, and so on — and no one in this dataset tested a three-rail lobby against a nine-rail one. If the search result generalizes, the industry’s default assumption about visible choice is backwards, and the next round of testing should say so. Players who find the interface harder to navigate than the games themselves should also have deposit-limit and self-exclusion tools as visible as the search bar, not buried behind a disclosure of their own.