How to win back lapsed customers in F&B — 2026 playbook

Win-back is the highest-leverage campaign in F&B and the one most often run badly. The typical version is a monthly blast to everyone quiet for 30 days, with 20% off, no segmentation and no holdout. It produces a redemption number that looks acceptable and an incremental effect close to zero, because most redeemers were coming back anyway.

Start with the trigger. Each customer has a normal gap between visits; the signal you want is deviation from their own gap, not a shared calendar. Fire at roughly 1.5× personal cadence for a soft touch and 2× for the offer. On the same customer list, cadence-based triggering typically converts two to three times better than a fixed 30-day rule, and it costs less because you send fewer, better-aimed messages.

Use two sends. The first is an acknowledgement with no offer attached — a short, human message that notices the absence and says something specific ('we have the single-origin you liked back on'). A meaningful share of lapsers return on this alone, at zero discount cost. The second, three to five days later, carries the actual offer for the people who did not respond. Leading with the discount trains customers to wait for it and gives away margin on people who did not need it.

Segment the offer by prior value. Three bands is enough. Light lapsers (one or two visits) get a small named item — they have not earned more and their expected value does not justify it. Established regulars get the standard offer: a named item or 10–15%, with a 14-day deadline. The top decile gets no discount at all; they get a personal gesture — a held table, something off-menu, a note from the owner. Discounting your best customers signals that the relationship was transactional all along.

Set expectations properly. Inside the recoverable window — roughly two personal cycles — recovery of 15–30% is realistic. Outside it, expect 5% or less, and treat those contacts as brand maintenance rather than a campaign. Chasing customers who have been gone eight months is where win-back budgets go to die.

Then prove it. Hold out 10% of every eligible segment permanently and never message them. The difference in 90-day revenue between the messaged and held-out groups is your real recovered value, and it will be lower than your redemption report suggests. That number is also the one that tells you whether to increase or decrease the offer — a decision that is otherwise guesswork.

The reason this is worth automating rather than running as a monthly ritual: the trigger is per-customer and continuous. A campaign calendar can only approximate it, and approximating a per-person deadline is precisely what makes the standard blast underperform.