Zappush predicts when each customer is due to reorder or about to churn. It then fires replenishment, win-back, and VIP campaigns across ads and email, with the discount sized to the actual risk. The scores run on complete server-side purchase history, not a decaying pixel that only ever saw the checkout.
From $59/month · 14-day free trial · Cancel anytime
Elena Brooks
Bought once, went quiet
Most small stores do retention in two modes: a monthly newsletter blasted to everyone, and a panic discount when revenue dips. The signals that would let you act earlier are each customer's purchase cycle, fading engagement, and rising or falling spend. They sit in transaction data no one has time to model. The pixel only sees the checkout, not the profile behind it, so it cannot tell you a 45-day-quiet VIP is drifting. By the time a customer looks lapsed on a static 90-day segment, the model that could have caught the behavioral drift weeks earlier was never built. So you over-discount loyal buyers who were going to reorder anyway, under-invest in the at-risk buyers who quietly leave, and let churn happen silently instead of triggering a win-back while the customer is still recoverable. Acquiring a new customer costs roughly 5-25x more than keeping one, so every silently churned repeat buyer forces you to spend acquisition dollars just to stand still.
Zappush computes recency, frequency, and monetary value on each persistent profile. From there it derives average inter-purchase interval, expected next-order date, predicted LTV, and a live churn-risk probability, in the spirit of BG/NBD + Gamma-Gamma models. All of it runs on complete server-side history, not a partial pixel view.
Customers due to replenish, VIPs gone quiet past their normal cycle, and first-time buyers entering the re-order window each become a self-updating audience. You define each one as rules over these scores (e.g. past 1.5x their average interval AND top-20% LTV). Members enter and leave automatically as their behavior and risk score change.
Zappush hashes each audience with SHA-256 and syncs it to Meta and Google via the Conversions API for retargeting or suppression, and to your email/SMS tool for the flow. The discount is sized to the risk: deep for a true win-back, none for a Champion who was buying anyway.
A live probability that each customer is drifting away. It rises as their gap since last order stretches past their own purchase cycle, so you catch behavioral drift before a customer is technically lapsed.
Predicted next-order date and average inter-purchase interval per customer. Reminders for consumables and refills land exactly when someone is running low, not on a fixed 30-day guess for everyone.
Previously valuable customers who have slipped past their normal cycle become a recoverable segment, synced to ads and email while there is still a relationship to save.
Champions do not need 20% off. They were buying anyway. Zappush reserves the deep discounts for genuine win-backs and protects contribution margin on loyal buyers who would have reordered at full price.
Rank by predicted LTV to give your top buyers VIP treatment, and suppress recent purchasers from acquisition retargeting so budget stops chasing customers who already bought.
REPLENISHMENT
A fixed 30-day refill reminder is wrong for most of your list. Zappush derives each customer's average inter-purchase interval and expected next-order date from complete server-side history, then turns those scores into a self-updating audience. Prompts land when a specific customer is actually running low.
WIN-BACK
Static 90-day lapsed segments catch people long after the relationship cooled. Churn risk rises once a buyer passes 1.5x their own purchase interval, so a quiet VIP surfaces weeks earlier than a fixed threshold would show. The audience syncs hashed to Meta, Google, and your email tool while there is still something to recover.
MARGIN CONTROL
Blanket discount blasts pay Champions to do what they were going to do anyway. Ranking by predicted LTV and live churn risk separates loyal buyers from genuine win-backs, so the offer is sized to the risk: nothing for Champions, deep for customers you are about to lose. Contribution margin survives the retention program.
Start free in minutes, or see it live on a call.
From $59/month · 14-day free trial · Cancel anytime
