The short version: a winback flow re-engages customers who stopped buying, and for a supplement brand the trigger math is different from every other category because the product has a consumption cycle. Define lapse relative to that cycle, run three emails (soft re-engagement, the reason to return, then one capped offer with a deadline), and hold the results against real benchmarks: 2 to 5% of recipients converting per email is healthy, 5 to 10% is top-quartile, and full programs reactivate 12 to 20% of lapsed customers, with the best reaching 20 to 35%. The economics are why this flow always earns its build: reactivating a lapsed customer runs roughly 5 to 10 times cheaper than acquiring a new one.
Define lapsed off the consumption cycle, not a calendar default
The stock Klaviyo winback triggers at a flat 90 or 120 days since last order, which is wrong in both directions for consumables. A customer on a 30-day bottle who has not ordered in 60 days is two full cycles lapsed and cooling fast; a customer who bought a 90-day supply is not lapsed at day 80, they are mid-bottle, and a winback email tells them your systems do not know them. The build:
- Lapse threshold per product: 1.5 to 2 times the product's consumption cycle, using the same days-of-supply math as the replenishment flow, quantity-adjusted at the order level.
- The replenishment flow runs first. Winback catches the people replenishment failed to convert, which means winback entry is really replenishment-exit-plus-a-cycle. If you have no replenishment flow, build it before this one; it converts better, earlier, at full margin.
- Exclude the deliberate leavers: anyone who unsubscribed, anyone in a cancel-save flow's graceful-exit track within the cool-down window, and anyone whose lapse is really a failed payment (that is the dunning flow's job).
Why supplement winback has a different job than retail winback
An apparel customer who lapsed simply shopped elsewhere or stopped needing clothes; the winback job is remembering. A supplement customer who lapsed usually stopped believing: they did not feel results, or never built the habit, and quietly concluded the product was not for them. The winback job is re-selling efficacy, which changes the content completely. A bare we-miss-you with a coupon answers none of that, and it is why generic winback sequences underperform in this category while objection-led ones hit the benchmark bands.
The build: three emails, then stop
The published structure matches what we run: soft re-engagement, then a moderate incentive, then a capped final offer with a real deadline, with diminishing returns and rising list-risk beyond three emails.
Email 1: at 1.5 to 2 cycles lapsed. The re-opening, no offer
Acknowledge the gap without guilt, then give the reason to return that is not a discount: what changed since they left (new formulation, new flavor, new third-party testing), the strongest recent review from a customer with their goal (the goal property from the welcome flow still routes here), and one honest line about consistency: most people who stop at week three stopped right before it works. That line, substantiated properly and inside FTC bounds, is the single highest-leverage sentence in supplement winback.
Email 2: 5 to 7 days later. The objection pass
Address why people actually quit, in their language: did not feel anything (results timelines, restated honestly), too expensive (cost-per-day framing, subscription price), too many pills (the simplified routine or the gummy variant). If your post-purchase surveys tag exit reasons, branch on them; if not, cover the top two objections in one email. A moderate value-add belongs here if anywhere: free shipping or a sample of the newer product, not yet the discount.
Email 3: 5 to 7 days later. One capped offer, one deadline
The discount, sized once and capped, with a real expiry, and the subscription option beside it (the lapsed customer restarting on subscribe-and-save is the best possible outcome, because the subscription's rhythm is the habit infrastructure they were missing). After this email, the flow ends. Non-responders route to the sunset track, because continuing to blast the unresponsive costs deliverability for everyone else.
Offer discipline: the anti-training rules
- The offer appears once, in the final email. Same logic as the checkout flow: escalating ladders teach the list that lapsing produces coupons, and supplement brands live on repeat buyers who learn patterns fast.
- Cap it below acquisition-offer levels. A winback discount richer than the welcome offer teaches the worst possible lesson: leave and come back for the best price.
- One winback offer per customer per year, tracked as a profile property, same guard as the cancel-save flow's.
What to measure
- Per-email conversion against the 2 to 5% healthy band (5 to 10% is top-quartile); below 2% means targeting, offer, or deliverability, in that order of likelihood
- Program reactivation rate against 12 to 20%, measured over the full sequence plus 30 days
- Second-order rate of reactivated customers: a winback that produces one discounted order and a second lapse was a coupon, not a reactivation. Route every reactivated customer straight into post-purchase and replenishment
- Subscription share of reactivations, the number that decides whether the flow rebuilt the habit or just rented it
Frequently asked questions
When should a supplement winback flow trigger?
At 1.5 to 2 times the product's consumption cycle since last order, per product and quantity-adjusted. Flat 90-day triggers fire too early for stock-up buyers and too late for 30-day-cycle customers.
What is a good winback reactivation rate?
2 to 5% of recipients per email is healthy and 5 to 10% is top-quartile; at the program level, 12 to 20% of lapsed customers is average and 20 to 35% is strong.
Should winback lead with a discount?
No. Lead with what changed and the efficacy case; hold one capped offer for the final email. Discounts answer a price objection, and most supplement lapse is a belief objection.
How is winback different from replenishment?
Replenishment is proactive and fires before run-out; winback is the safety net one to two cycles after it failed. Replenishment converts better and cheaper, which is why it runs first and winback only catches its misses.
What happens to customers who ignore the winback flow?
They enter the sunset flow. Emailing the permanently unresponsive erodes deliverability for the whole list, and the winback sequence was their last full-price touchpoint.