Winning back lapsed customers on WhatsApp

The 90-day win-back is the campaign everyone wants to run and the one most likely to damage the account that runs it. Not because the idea is wrong, but because the list is cold, the consent is old, and WhatsApp measures both.

Published 23 August 2026 · By Adjoltz

Why this send is riskier than any other

Two mechanics compound against you.

Quality rating. WhatsApp computes a quality rating per phone number from user signals — blocks, reports, and how people respond — over the past 7 days, weighted by recency. A cold list produces exactly those signals. And quality rating gates your messaging tier: 250, 1,000, 10,000, 100,000, or unlimited business-initiated conversations per number in a rolling 24 hours. A bad win-back does not just underperform. It can take your sending capacity down with it, and that hits every other campaign you run.

The frequency cap. Error 131049 drops marketing messages per user after the API has returned success. Lapsed contacts are, almost by definition, people you have messaged before without result. On a UAE account we operate it accounted for 1,388 of 1,557 delivery failures.

Wave, measure, wave

1. Take the most recently lapsed cohort first — they are warmest and lowest risk.
2. Send a wave of a few hundred.
3. Stop and look at quality rating, block rate, and 131049 failures before the next wave.
4. If quality holds, go again with the next cohort. If it moves, stop entirely.

This is slower than a single blast, and it is the difference between a campaign and an incident. It also has a side benefit: steady volume across days is what the tier-up criteria reward, since raising your limit requires having initiated conversations with at least half your current limit in unique customers over the last 7 days.

Segment by what they bought, not by how long ago

"90 days since last order" is a database query, not a segment. It puts someone who bought a mattress next to someone who bought coffee beans, and the coffee buyer is three months overdue while the mattress buyer is nine years early.

Useful segmentation:

• Consumables past their replenishment cycle. The only genuinely time-based win-back that makes sense, and the highest-converting one.
• One-time buyers who never came back. A different message entirely — the question is why the first purchase did not stick.
• Former regulars who stopped. The most valuable and the most likely to have a specific reason. Ask what happened rather than offering a discount.
• Seasonal. Someone who only ever buys in November is not lapsed in July.

Check the consent, honestly

If the newest thing you can show for a contact is a tick box from two years ago on a form you no longer use, that is a weak position. Practical filter: keep contacts whose consent you can evidence with a date, a source and a wording; drop the rest. A smaller list that you can defend is worth more than a larger one that generates reports — see bulk WhatsApp and UAE law.

What to actually say

Discount-first is the reflex and it is usually wrong: it trains the customer to wait for the next discount, and it converts the people who were going to return anyway. Better openings, roughly in order:

1. Something genuinely new since they last bought. A product, a service, a location.
2. A question. "We haven't seen you since March — did something go wrong?" Low volume, high value, and the replies are free because they open a 24-hour window.
3. A replenishment nudge, if the product has a cycle.
4. A discount, last, and only to the cohort that ignored the first three.

The numbers we have

On a UAE account we run, 79.2% of delivered marketing messages were read and 11.7% of contacts replied, with 6 opt-outs across the period. That is the general marketing baseline. We have not run a large enough win-back-specific sample to publish a separate reactivation rate, and anyone quoting you one for your market should be asked what it is based on.

Frequently asked questions