The short version: a sunset flow is the structured goodbye for subscribers who stopped engaging: one re-permission email offering a real choice, one goodbye, then suppression. It is the least glamorous flow in the account and one of the most valuable, because inbox providers score senders on engagement, and every send to someone who never opens teaches Gmail that your mail belongs in spam. Klaviyo's own hygiene guidance is blunt about the stakes: a damaged sender reputation can end in blocklisting, at which point your emails stop reaching even the engaged customers, and a smaller engaged list outperforms a larger one padded with the inactive. This post is the build itself; for where to draw the dormancy line, we wrote a dedicated guide to setting the threshold against your repeat purchase cycle, and this flow assumes that math is done.
Why brands refuse to build this flow, and why they are wrong
Suppressing subscribers feels like burning money you paid to acquire. The reframe: an unengaged profile is not an asset, it is a small ongoing tax on the deliverability of every email you send. The list-size number on the dashboard is vanity; placement and revenue per send are the business. Brands that sunset aggressively see open and click rates climb not because anyone got more interested, but because the mail started landing in inboxes again, and the engaged core was always carrying the revenue anyway. The sunset flow is how you stop paying for the ghosts.
Who enters: the threshold, in one paragraph
The full logic lives in the threshold guide; the operating summary is: define dormancy against your own repeat purchase window (roughly twice the median time between orders, with a floor around 120 days), never against a borrowed 90-day rule, and never on opens alone, because Apple's Mail Privacy Protection registers an open for every email delivered to Apple Mail whether it was read or not, inflating open rates 15 to 40% on Apple-heavy lists. Dormancy is built on clicks, site activity, and orders. Klaviyo's guidance frames the tenure side of it plainly: a week-old subscriber who has not engaged is just new, while three months of silence is a different story. Our own rule is more conservative: give every subscriber at least one full repeat cycle of normal sending before the sunset track can claim them. And the entries that never happen: recent purchasers, active subscribers, and anyone the winback flow has not yet had its commercial shot at, because the pipeline runs replenishment, then winback, then sunset, in that order.
The build: two emails, then the exit
Klaviyo's model is two parts: a re-permission email and a goodbye. We run it the same way. Sunset is not another sales sequence, and putting offers in it re-trains lapsers; the winback flow already made the commercial case.
Email 1: the re-permission
One honest question: do you still want these emails? One button to stay (a real click, which instantly re-engages the profile), a visible unsubscribe, and optionally a preferences link (fewer emails, only restock alerts, only sales) that saves the marginal cases. No guilt copy, no clingy subject lines. The brands that do this with dignity convert a surprising slice back to active, and everyone else exits cleanly instead of via the spam button, which is the exit you were preventing all along.
Email 2: 7 to 10 days later. The goodbye
For non-responders: you will stop hearing from us, here is where to find us if you want back in, one last click-to-stay. Then the automation adds non-responders to a suppression segment. Suppression, not deletion: the history stays, the profile can re-enter by purchasing or resubscribing, and your billing tier usually drops because most ESPs price on emailable profiles. Sunset flows quietly pay for themselves on the Klaviyo invoice alone.
The mechanics that make it safe
- Exit on any engagement, checked at every step. A click anywhere in the flow removes the profile from the sunset track and resets their engagement clock.
- Suppress, do not delete, and never re-import. The classic deliverability self-injury is exporting suppressed profiles and re-uploading them in a moment of list-size anxiety before BFCM. The engagement problems come back with them, at the exact moment sender reputation matters most.
- Campaign segments enforce the same line: your default campaign audience should already exclude the unengaged window, so the sunset flow is formalizing reality, not creating it.
- Seasonal and long-cycle brands stretch the clock, per the worked examples in the threshold guide: a full season of silence for a fishing brand, 18 to 24 months for leather and furniture. The flow build is identical; only the entry gate moves.
What to measure
- Re-permission recovery rate: the share of sunset entrants who click to stay, typically small but pure profit
- Deliverability trend after suppression: open and click rates on the remaining list, and inbox placement if you run seed tests, over the following 60 days
- Spam complaint rate, which should fall toward zero on the cleaned list: Klaviyo's guidance keeps it well under 0.1%, and Gmail and Yahoo reject bulk senders outright at 0.3%
- ESP cost per engaged profile, the quiet ROI line: suppression usually drops the Klaviyo tier while revenue holds, because the suppressed were not buying
Frequently asked questions
When should a subscriber enter the sunset flow?
After the dormancy threshold your repeat purchase cycle sets (roughly twice the median reorder gap, 120-day floor), never inside a new subscriber's first repeat cycle, and only after the winback flow has run. Recent purchasers and active subscribers never enter.
Should the sunset flow include a discount?
No. The winback flow already made the commercial case. Sunset is a permission question, and an offer inside it teaches people that ignoring you pays.
Suppress or delete unengaged profiles?
Suppress. History is preserved, re-entry by purchase or resubscribe stays possible, and the profiles stop counting against deliverability and, on most plans, billing. Delete only what compliance requires.
Does shrinking the list really help revenue?
Revenue per send rises and placement improves, which lifts every future campaign to the engaged core that was producing the revenue anyway. The list-size metric falls; the business metrics do not.
How is sunset different from winback?
Winback is a commercial sequence trying to recover a customer. Sunset is a permission sequence ending the relationship cleanly for people winback could not move. They run in that order, never merged.
Sources
- Klaviyo. Email list hygiene best practices and sunset flow structure.
- Mailflow Authority. Apple Mail Privacy Protection and open rate inflation.
- Klaviyo. Email marketing benchmarks.