Win-back campaigns built on your actual churn window
Most win-back flows trigger at ninety days because ninety is a round number. For half of catalogues that is far too late, and for the other half it is premature.
A win-back campaign is only as good as its trigger. Send it too early and you are discounting customers who were going to return anyway; too late and the relationship has already gone somewhere else. The Analyst agent derives the actual repurchase interval for your catalogue — usually different per category — and Retention builds the sequence around that rather than around a convention.
Find the real churn window first
Repurchase behaviour is category-specific and often bimodal. Averaging across the whole catalogue produces a number that describes nobody. Analyst segments it properly before any campaign is built.
Repurchase interval distribution per category and per first-purchase value
The point at which return probability drops sharply — the real churn edge
Which customer segments are genuinely recoverable versus permanently lapsed
Whether lapse correlates with a product problem rather than a marketing one
Say something other than "we miss you"
Generic re-engagement performs poorly because it gives no reason to act. The agent writes around what actually changed since the customer left — new products in their category, a restock of what they bought, a fixed issue they complained about — and reserves incentives for the segment where they demonstrably work.
Know when to stop
A subscriber who has ignored a full win-back sequence is costing deliverability. The agent enforces a sunset policy, moving unrecoverable contacts out of active sending rather than mailing them indefinitely.
How it works
Derive the window
Analyst computes repurchase intervals per category and identifies where return probability collapses.
Build segmented sequences
Retention writes different sequences for recoverable segments, with incentives only where they change outcomes.
Sunset the rest
Contacts who do not respond are moved out of active sending to protect deliverability.
What you get
Triggers based on your real repurchase data, not a default interval
Messaging built on what changed, not on generic sentiment
Incentives limited to segments where they measurably work
A sunset policy that protects list health
Handled by these agents
Works with
Frequently asked questions
When is a customer actually lapsed?
When their probability of returning drops materially below the baseline for their category — which for consumables can be weeks and for durables can be a year. That is why the window is derived rather than assumed.
Do discounts work for win-back?
Sometimes, and more often than in cart recovery — but they also risk training customers to lapse deliberately. The agent tests it per segment instead of applying it universally.
How many attempts before giving up?
Until incremental revenue per send goes negative for that segment, which is a measurable point rather than a matter of taste. After that, sunsetting protects deliverability for everyone else.
What if customers lapsed because of a product problem?
Then a marketing sequence will not fix it, and the agent should say so. Analyst checks whether lapse correlates with returns or support contacts before Retention writes anything.
Keep reading
Win back lapsed customers
A re-engagement programme built on real churn windows, honest offers, and a sunset policy.
Customer segmentation
Segments built from purchase behaviour, not from generic engagement windows.
Subscription retention
Reduce subscriber churn with intervention before the cancel, not a survey after it.
Post-purchase upsell
Raise repeat rate and order value in the window after purchase, without souring the experience.
Put the team to work on your store
Connect your storefront, analytics, and email stack, set a goal, and let the agents run the work end to end. Start on the free plan with your own model key.