Dormant player segmentation: who to call, who to leave

Most guides to dormant player segmentation stop at recency, frequency and monetary value. RFM is a sound starting point, and Track360's reactivation playbook calls it the most practical scoring framework for casino reactivation (Track360). But RFM tells you how much a player is worth chasing, not how to approach them. Two players with identical scores, one who left after a withdrawal took a week and one who left after a run of heavy losses, need completely different calls. One needs an apology and a fix. The other shouldn't get a marketing call at all.

The dormant base is also large. Optimove analysed several iGaming brands with 5,341,332 players in total and found about 60% were dormant, using a one-year definition (Optimove). With that many people in the bucket, the order you segment in matters as much as the segments themselves.

Pass 1: who you may contact

Before any scoring, remove everyone you can't contact on the channel you're planning. Doing this first keeps excluded players out of every count downstream, so your segment sizes and forecasts are real.

• Self-exclusion registers. National and operator-level. In Great Britain, self-excluded players must be removed or flagged in marketing databases within two days of the notification (Gambling Commission, SR 3.5.3). Each market's register and rules are on the self-exclusion registers by country page.
• Consent per channel and product. In Great Britain, SR Code 5.1.12 has required opt-in per product and per channel, with phone call named, since 1 May 2025 (CMS). A dormant player who opted in to casino emails three years ago hasn't opted in to casino calls.
• Safer-gambling markers. Players with strong indicators of harm are out of marketing altogether. The Commission requires licensees to prevent marketing and new bonus take-up where there are strong indicators of harm (customer interaction guidance, requirement 10).
• Players who used safety tools. Anyone who set a deposit limit, took a time-out or asked to close their account just before going quiet. Their silence may be the tool working.
• Previous "stop calling" requests, from any campaign or channel.

Pass 2: dormancy against the player's own rhythm

Fixed calendar bands (30, 60, 90 days) are easy to run and wrong for many players. A player who deposited every Friday for a year and has missed three Fridays is a much stronger signal than a player who deposited four times a year and hasn't been seen for 60 days.

Measure the gap as a multiple of the player's own median time between deposits:

For players with too few deposits to have a rhythm, fall back to calendar bands. Many operators treat 30 days without a deposit as dormant and 180 as lost, but set the bands from your own deposit data.

Pass 3: what they were worth

Value decides how much contact a player is worth, and value is concentrated. In NatCen's study of 139,152 British online accounts, the top 20% of gaming accounts by stake generated just over 90% of operator revenue (NatCen, Patterns of Play). Optimove's dormant lottery study found something similar at the other end: its highest-potential dormant segment was only about 1% of dormant players, with the best reactivation rate of any group, 0.71%.

Use net gaming revenue, not gross, so bonus-heavy players don't look better than they are. Three value bands are enough for most programmes: high (your VIP and pre-VIP tiers), mid, and low. Dial Milo's free player LTV calculator helps you put a number on each band.

Two cautions. Treat a sudden spike in value just before a player went quiet as a possible harm signal, not a reason for priority. And remember the NatCen finding cuts both ways: the most valuable accounts were also the ones with the heaviest losses.

Pass 4: why they probably left

This is the pass most programmes skip, and it's the one that decides the call. You rarely know for sure why someone stopped, but the last few weeks of the account usually point somewhere.

The event-led row comes from Sportradar's segmentation of dormant sports users, which separates players waiting for a big enough event from those unlikely to return at all (Sportradar). For sportsbooks it's often the largest dormant group that can be won back cheaply.

Pass 5: whether you can reach them

A dormant segment is only as useful as the channel you can reach it on. Check:

• Email. Bounces, or no opens for months. Inactive players often have stale addresses.
• Phone. A number in valid international format, with phone consent. Mobiles change hands less often than email addresses go stale, but they do.
• Push. Only if the app is still installed and notifications are on, which for long-dormant players is rare.

The players whose email has gone quiet but whose phone is valid and consented are a segment in their own right. For them, a call may be the only channel that still works.

Putting dormant player segmentation together: the grid

Combine pass 2 and pass 3 into a grid, then let pass 4 override the action in any cell.

The dormancy segmentation template has this grid ready to copy or download, with the bands editable to fit your own data.

Why calls in the mid-value row? Because the cost changed. With a human call centre, the agent is paid for every dial, including the ones nobody answers. With an AI agent, unanswered dials have no minute charge, so mid-value players come into range for a call. Calls for high-value players still lean on the host. The VIP host guide covers how to split that work.

A worked example

Illustrative figures throughout. You have 10,000 dormant UK players.

Pass 1 removes 2,500: 900 on self-exclusion lists, 1,200 without phone consent, 400 with safer-gambling markers or recent use of limits. 7,500 remain.

Passes 2 and 3 put 900 in high value, 2,600 in mid value and 4,000 in low value.

Pass 4 moves 300 high- and mid-value players out of marketing altogether (heavy losses before going quiet) and flags 450 for service calls (withdrawal or verification problems).

The low-value 4,000 get email only. That leaves 3,200 phone contacts: 650 high value (host-led or AI check-in calls), 2,100 mid value with offers, and 450 service calls.

Say 30% answer, calls average two minutes, and you're calling UK mobiles from a UK number. The line starts at 8.5¢ a minute and is billed per started minute; the AI is 9.57¢ per talk minute. That's about 18¢ a minute all in.

960 answered calls × 2 minutes × roughly 18¢ ≈ $345 in call cost, plus a few dollars for voicemails. Unanswered attempts have no minute charge. Whether that's worth it depends on how many come back and what they're worth, which the reactivation ROI calculator works out from your own rates.

Turning segments into calls

1. 1

   Export each segment as its own list

   One CSV per segment, with phone, name and any columns you want the call to use: segment, favourite game, VIP tier, language, last-seen date. Dial Milo saves extra columns as player details and can use them in the script as placeholders such as {favourite_game}.

2. 2

   Write a script per contact type, not per segment

   A service script, a check-in script and an offer script cover most of the grid. The offer script takes an offer code per player, so the mid-value lapsing group and the mid-value dormant group can hear different offers from the same script.

3. 3

   Set calling rules

   Calling hours in each player's local time, two attempts a few hours apart, and a contact cooldown across campaigns so nobody gets a reactivation call and a VIP check-in in the same week.

4. 4

   Keep a holdout

   Hold back 10% of each segment from the call. Without it, you can't tell reactivation from players who would have come back anyway.

5. 5

   Read the outcomes and move people

   After the calls, each player carries an outcome and a next-step label. Players marked "stop contacting" go into suppression, "move to VIP support" goes to your host, and anyone flagged for safer gambling goes to your team and onto the do-not-call list automatically. Feed the outcomes back into your CRM so the segments update; the iGaming CRM guide covers the wiring.

Mistakes worth avoiding

• Scoring before suppressing. Your forecast includes players you can't call, and someone eventually calls them.
• One offer for the whole dormant base. The players most likely to come back for a bonus are the ones who least need it, and the ones who left for a reason won't come back for one.
• Treating long dormancy as an opportunity. Six months of silence is often a decision. Re-permission calls respect that; heavy offers don't.
• Ignoring the last session. The few weeks before a player went quiet say more about the right call than their lifetime value does.
• Not updating the segments. A player who said "call me after payday" belongs in a different segment next week.

For the full reactivation programme, including channel order and how to measure it, start with the player reactivation guide. Players who registered but never deposited are a different problem from dormant ones, and have their own guide: registered no-deposit players.

You can hear a reactivation call before committing to anything: the browser demo lets Milo call you, and a first campaign can run from a $10 top-up. Create an account and upload your first segment.