A 400ms delay between tapping confirm and the bet landing lifted stakes by 29% in a test run across three sportsbook apps last spring. The effect held for both casual users and accounts with 500+ lifetime bets, which is the part that should bother anyone who designs these flows. Adding friction to a one-tap product usually costs you volume. Here it did the opposite.

What the test actually did

The setup: a standard snap-bet widget — pre-filled stake, single confirm button, no slip review. Variant A fired the bet immediately. Variant B inserted a 400ms spinner after the tap, then confirmed. Same odds, same markets, same promotional calendar. Sample was roughly 180,000 sessions over six weeks, split 50/50.

Average stake per snap bet went from €14.20 to €18.30. That's the 29%. Session count per user dropped 4.1%, so net handle still rose. The teams running it expected the delay to suppress impulse bets. It reshaped them instead.

Why a pause changes the number

The obvious reading is regret avoidance: users tap, see the spinner, and get a half-second to reconsider — but they don't cancel, they re-anchor. A stake field that says €10 feels final when the bet fires instantly. When there's a beat before it locks, some users read it as a confirmation of a decision rather than the decision itself, and confirmations invite a slightly bigger commitment.

There's a second mechanism, less flattering. The spinner creates a moment where the user is watching the product rather than the match. Attention on the interface, not the outcome, tends to correlate with larger stakes in every behavioural dataset I've seen from operators willing to share.

The segments where it broke

The 29% is an average, and averages hide the useful part.

• New accounts (first 7 days): +6%. Barely moved. They don't have a stake habit to re-anchor.
• Accounts with 500+ lifetime bets: +34%. The heaviest users responded most, which inverts the usual assumption that experienced bettors are friction-insensitive.
• Live in-play bets: +41%. Prematch snap bets were closer to +19%.
• Same-game parlays: flat. Parlay builders already have a review step, so the extra pause added nothing.

That in-play figure is the one to sit with. Live betting is where margin pressure and emotional decision-making peak anyway. A 400ms delay there isn't neutral design — it's a stake inflator aimed at the least considered bets on the board.

What operators did with it

At least two of the three apps kept the delay in production. One rolled it back after a compliance review flagged that the spinner text read "Placing your bet…" while the bet was, in fact, already placed — a wording problem, not a numbers problem, but it says something about how the feature was framed internally.

The regulatory angle is thinner than it looks. Most licensing regimes require clear confirmation of a bet's terms, not a delay. A 400ms pause satisfies nobody's definition of a cooling-off period, and it isn't sold as one. It's sold as perceived reliability — the spinner makes the app feel like it's doing work.

The question nobody's answering

If 400ms produces 29%, what does 2 seconds produce? Nobody in this test pushed it that far, and the rollback suggests someone got nervous about finding out. The honest implication is that the line between "smooth UX" and "engineered hesitation" is now under half a second wide, and the metric that would tell you which side you're on — stake per session among users flagged for gambling harm — wasn't in the dataset.

That's the number worth publishing next.