Session-Reminder Emails Cut Slot Play 26% at Day 9

A randomised field trial across four India-facing slot lobbies found that a single session-reminder email, dispatched at the 90-minute mark of continuous play, reduced average daily slot handle by 26% among treated accounts by day nine. The effect was measured against a matched control group of 4,180 accounts and persisted after controlling for deposit frequency, device type, and time-zone clustering. What makes the finding notable is not the size of the drop but its timing: the divergence between treatment and control widened sharply between day six and day nine, suggesting the intervention altered play rhythm rather than merely interrupting a single session.

The Study Design and Why Day Nine Matters

The trial ran from 14 August to 22 September 2023 across four operators licensed under the Meghalaya and Nagaland frameworks, chosen because both jurisdictions permit player-level data sharing for responsible-gambling research under existing consent clauses. Accounts were randomised into treatment and control at the operator level, not the individual level, to avoid contamination through shared devices and payment instruments — a common flaw in earlier reminder studies.

The treatment was deliberately minimal: one email, no push notification, no SMS, no in-app modal. Subject line variants were tested but the body was fixed — a neutral statement of elapsed session time, a link to the operator's self-set limit page, and a one-line note that play could continue without penalty. No bonus, no free spins, no incentive to return. This matters because a large share of prior "reminder" interventions bundled the message with a retention offer, which confounds any reduction in play with the pull of the offer itself.

The 26% figure is not the raw difference in handle. Raw handle fell 31% in the treatment arm and 7% in the control arm, giving a net difference of 24 percentage points. After adjusting for pre-trial baseline play, the modelled treatment effect settled at 26.3% (95% CI: 22.1–30.4). The day-nine marker was pre-registered as the primary endpoint because earlier pilot work showed that reminder effects decay within 72 hours if the message is a one-off. That decay did not occur here.

Why the Effect Grew Instead of Faded

The growth pattern is the most interesting result. By day three, the treatment effect was 11.4%. By day six it was 18.9%. By day nine it was 26.3%. Three explanations fit the data reasonably well.

The first is habit interruption. Slot play at these four operators clustered heavily in 60–120 minute blocks, often beginning between 21:00 and 23:30 IST. A reminder at 90 minutes lands inside that block, not after it, and may break the automatic transition from "one more spin" to "one more hour."

The second is selection into self-limits. Among treated accounts, 8.7% set a deposit limit within 14 days of receiving the email, against 2.1% in control. Deposit limits, once set, tend to persist, which would explain why the gap widens over time rather than closing.

The third is measurement artefact, and it deserves more attention than it usually gets. If treated players simply shifted play to a different operator, the reduction would be an illusion. The study attempted to control for this by cross-referencing hashed payment identifiers across the four participating operators, but a player using a fifth operator outside the panel would appear as a genuine reduction. The authors estimate this leakage at 3–6% of the treatment effect, which still leaves a real reduction of roughly 20–23%.

What the 26% Figure Does and Does Not Cover

A 26% reduction in daily handle is not a 26% reduction in harm. Handle is turnover, not loss. A player who cycles ₹10,000 through a 96.2% RTP slot loses roughly ₹380 in expectation per cycle, but variance means individual outcomes range from near-zero loss to full depletion. Reducing handle reduces expected loss proportionally but says nothing about whether the player's actual financial position improved.

The study did track two secondary outcomes: average session length and self-reported loss-chasing behaviour on a weekly survey (n=1,204 respondents, 29% response rate). Session length fell from a mean of 104 minutes to 78 minutes in the treatment arm. Self-reported loss-chasing — defined as playing specifically to recover a prior loss — fell from 34% of respondents to 26%. Both are directionally consistent with the handle data, but the survey response rate is low enough that these should be treated as suggestive.

There is also a composition question. The 26% reduction was not evenly distributed. Accounts in the top decile of pre-trial handle showed a 34% reduction. Accounts in the bottom decile showed a 9% reduction, which was not statistically significant at the 5% level. If the intervention works mainly on heavy players, that is a good outcome for harm reduction but a poor one for operators relying on that segment's volume — and it explains why some operators have been reluctant to adopt reminder emails despite the evidence.

The India-Specific Context

India's regulatory position on online slots remains fragmented. Games of skill operate under a different legal footing from games of chance in most states, and the 2023 amendments to the IT Rules introduced a self-regulatory framework that, among other things, requires "reasonable" responsible-gambling measures from permitted online gaming intermediaries. What counts as reasonable was left undefined. A 26% handle reduction from a single email is the kind of number that could plausibly inform that definition, though no regulator has yet cited this trial.

The practical constraint in India is delivery. Email open rates for transactional messages from gaming operators in the sample ran at 41% within 24 hours, well above the global average but far below SMS. A reminder that reaches only 41% of recipients still produced a measurable population-level effect, which suggests that the per-opened-email effect is considerably larger — perhaps in the range of 50–60% for those who actually read it. That per-user figure is more relevant for operators deciding whether to send, and more relevant for regulators deciding whether to mandate.

What Would Change the Conclusion

Three things would materially weaken the finding. If the payment-hash leakage is higher than estimated — say 15% rather than 3–6% — the true effect shrinks toward 15% and the day-nine persistence becomes harder to defend. If the effect is driven mainly by the 8.7% who set deposit limits, then the email is really a nudge toward a different intervention, and the question becomes whether the email is the cheapest way to achieve that nudge. And if a longer follow-up at day 30 or day 60 shows reversion, the "day nine" framing becomes a story about a transient effect rather than a durable one.

The study did not run past day 21. That is the gap. A 26% reduction at day nine is a solid result, but it tells us almost nothing about whether the same players are back at baseline by day 45, or whether they have simply moved to operators that do not send the email. The next useful trial would follow treated accounts for 90 days, match them against a national payment panel rather than four operators, and report the effect separately for players who did and did not set a limit. Until then, the honest reading is that session-reminder emails work, but we do not yet know for how long, or for whom, or at what cost to the operators who might be asked to send them.