A 2024 field study of 41,000 cash-out attempts across three licensed U.S. operators found that relocating the cash-out control from a buried account menu to a two-tap path on the live bet slip reduced "session-70" cancellations — bets abandoned at the 70th minute or later, when in-play prices move fastest — by 10.2%. The effect was not uniform: it concentrated in soccer and tennis moneylines, where the gap between displayed and executable price is widest, and it largely vanished in markets with fixed settlement windows.
That framing matters because the cash-out button is usually discussed as a UX nicety. It is better understood as a pricing surface. Every second between a user's intent to exit and the operator's execution is a second the book's in-play model reprices against the position. Placement, in other words, is not cosmetic; it is a lever on realized customer outcomes, and the operators who move it are quietly changing the distribution of who captures the margin on an early exit.
What "Session-70" Measures and Why It Is the Right Dependent Variable
Most cash-out analytics focus on take-up rate: what share of eligible bets get cashed out. That metric is contaminated by offer design. An operator that surfaces aggressive partial cash-out prompts will inflate take-up without improving anything for the customer.
Session-70 cancellations are cleaner. The term, as used in the study, refers to a cash-out request initiated after the 70th minute of a soccer match (or the equivalent late-stage point in tennis) that the user then abandons — closes the slip, backs out of the confirmation, or lets the request time out. The 70-minute threshold is not arbitrary. It is roughly where live in-play margins tighten and where a single goal can swing a moneyline by 40 cents or more. A cancellation at that stage is a revealed preference: the user wanted out and the interface, or the price it surfaced, talked them back in.
The 10.2% reduction decomposes unevenly. Roughly 6.4 points came from fewer abandoned confirmations — the user tapped cash out and completed it. The remaining 3.8 points came from users who never reached the confirmation screen at all in the old design because the control sat two menus deep. In the control cohort, 18.7% of session-70 attempts died at the navigation stage. Under the two-tap layout, that figure fell to 6.1%.
The Confirmation-Screen Problem
A meaningful share of cancellations were not navigational. They were the user seeing the actual cash-out figure — often 8–14% below the displayed live value — and balking. Two-tap placement does not fix that. It shortens the path to the number, which cuts abandonment from friction but does nothing for abandonment from sticker shock. Operators that paired the layout change with a pre-quote ("estimated $X, final price on confirm") saw the 10.2% figure hold; those that did not saw it erode to about 6% within six weeks as users learned the quote was soft.
Where the Effect Concentrates
The study's market-level breakdown is the most useful part for anyone deciding whether to replicate it.
| Market type | Change in session-70 cancels | Notes |
|---|---|---|
| Soccer moneyline | −14.8% | Largest effect; late goals dominate |
| Tennis match winner | −11.3% | Retirement risk drives urgency |
| Soccer spread/total | −7.2% | Tighter live margins |
| Basketball (late-game) | −2.9% | Settlement windows and foul dynamics |
| Pre-match only | −0.4% | Not statistically significant |
The pattern is consistent with a friction-and-price story. Markets where the executable price moves fastest and where the user has the strongest reason to want out immediately — a tennis player clutching a hamstring, a soccer side conceding a red card — show the biggest placement effect. Markets with predictable late-game structure show almost none. Basketball's late-game fouling and timeout patterns make the final two minutes more legible to users, so the interface matters less than the model.
The Mobile Share
Ninety-one percent of session-70 attempts in the sample came from mobile. That is not surprising, but it reframes the finding. On a 6-inch screen, "two taps" versus "four taps plus a scroll" is not a minor difference; it is the difference between a control that is reachable mid-action and one that is not. Desktop users showed a 3.1% reduction — real, but an order of magnitude smaller. Any operator reading this as a general UX principle rather than a mobile-specific one will over-invest.
The Counterargument: Faster Exits Are Not Automatically Better
There is a defensible position that the 10.2% figure describes a transfer, not a gain. If a user cashes out at the 72nd minute instead of holding to settlement, the operator captures the cash-out margin — typically 3–6% of the position — and the user forgoes the upside of the bet they originally placed. A smoother button may simply convert more users into that margin.
The study cannot fully adjudicate this. It measured cancellations, not realized user P&L. But two details push against the pure-transfer reading. First, the reduction was concentrated among users who had already initiated a cash-out in the prior 90 days — habitual exiters, not first-timers coaxed into a bad habit. Second, the operators in the sample reported a 4.1% drop in session-70 re-bets in the treatment cohort, meaning fewer users responded to a failed cash-out by doubling down on a live market. That is a harm-reduction signal, however modest.
The honest reading is that two-tap placement makes an existing behavior easier. Whether that is good depends on what the behavior was. For a user trying to lock in a profit on a 4-leg parlay with one leg live, it is plainly good. For a user panic-selling a position that has another 20 minutes to resolve, it is less clear.
Regulatory Overhang
Three states now require operators to display cash-out values as "estimates" with a stated variance band, and two more have introduced bills doing the same. A two-tap path that surfaces a hard number and then reprices on confirm is exactly the pattern those rules target. Operators weighing the layout change should assume the confirmation screen will get more disclosure, not less, and that the 10.2% figure will compress as a result.
What the Number Is Actually Worth
Treat 10.2% as a point estimate from a single 41,000-attempt sample, not a law. The confidence interval in the published version runs from 7.9% to 12.6%, and the sample skews toward operators with modern in-play engines. A book running a 4-second price refresh will see a different number than one refreshing every 400 milliseconds.
The more durable takeaway is directional: cash-out abandonment is dominated by interface friction and quote uncertainty, in that order, and friction is the cheaper of the two to fix. The operators that captured the full effect paired the layout change with an honest pre-quote. The ones that moved the button and nothing else got a third of the benefit and a slow decay.
The open question is whether the effect survives familiarity. Users learn interfaces quickly, and a control that feels fast in month one may feel ordinary by month six — at which point the remaining cancellations are almost entirely price-driven, and the lever shifts from design to the in-play model itself. The next study worth reading will measure the same metric eighteen months out, on operators that have held the two-tap layout the whole time. If the reduction holds near 10%, placement is a structural fix. If it drifts toward 4%, it was a novelty, and the real work was always on the quote.