SkillMedev Win-Back Campaign
A win-back campaign is a structured last attempt to re-engage users who have stopped responding - not a discount blast. Done correctly it surfaces the real reasons users left and converts a meaningful minority back into active customers; done lazily it burns deliverability, trains the list to expect discounts, and buries the churn signal the business needs. The two costly mistakes this skill prevents: treating all dormant users as one segment, and sending a "final email" that turns out not to be final.
Operating procedure
Step 1: Gather inputs
Collect before writing anything. Label estimates as guesses.
- Product usage cadence: daily-use app, monthly subscription, or B2B tool - this sets the dormancy definition.
- The activity event that counts as "alive" (login, purchase, core action) and the re-activation event that counts as "won back."
- Segment sizes at each inactivity depth, and rough LTV per segment (sets the incentive ceiling).
- What has shipped since these users left (fuel for Beat 2).
- Available incentives: discount, free month, feature unlock, consultation - and what finance will approve.
Step 2: Define dormancy precisely
Dormancy means different things per product. Set a hard threshold: no login in 30 days for a daily-use app, no purchase in 90 days for a monthly subscription, no activity in 60 days for a B2B tool. The threshold determines segment size and message urgency. "Kind of inactive" is not a segment.
Step 3: Tier by inactivity depth
Depth of dormancy predicts winnability, so the offer must scale with it:
- Tier 1 - recently lapsed (past threshold up to ~2x threshold, e.g. 30-60 days for a daily-use app): highest win rate. No incentive yet; lead with value and what is new. Discounting this tier trains users to lapse for coupons.
- Tier 2 - cooled (roughly 2x-3x threshold, e.g. 60-90 days): value message plus a modest incentive.
- Tier 3 - cold (beyond ~3x threshold, e.g. 90+ days): the strongest approved incentive, then a genuine goodbye. Win rates here are low single digits; spend accordingly.
Run each tier through the same three-beat structure with tier-appropriate offers.
Step 4: Structure each tier as three beats
- Beat 1 (day 0): acknowledge the silence without guilt-tripping. Open by naming the gap plainly - the user has not been around lately. Do not pretend nothing happened; users who feel seen re-engage at higher rates than recipients of tone-deaf promotion. Remind them of the specific value they were getting (their data: projects created, money saved), not generic benefits.
- Beat 2 (day 7): show what is new since they left - a feature, an improvement, a piece of content. Give them a reason the product today is not the product they abandoned.
- Beat 3 (day 14): the direct ask with the tier's incentive and a real deadline. This is the final email - apply the honesty rule below.
Step 5: Apply the final-email honesty rule
If Beat 3 says "this is our last email," it must be true. After Beat 3, suppress non-responders from all marketing sends. A "final" email followed by more emails destroys the credibility that makes final emails work, invites spam complaints, and poisons every future campaign. Honesty here is also the mechanism: a true goodbye is the single highest-open email in most win-back sequences.
Step 6: Choose one strong incentive, not several
Offer one clear thing - a discount, a free month, a feature unlock, a consultation. Multiple small incentives read as desperation and dilute perceived value. Size it to segment LTV: an incentive worth more than roughly one month of expected margin on a won-back user is buying vanity re-activations.
Step 7: Measure re-activation, then close the loop
Success is users returning and completing a meaningful action, not opening an email. Track re-activation rate as the primary metric (open and click rates are diagnostics only). Include a one-question "what made you leave?" reply prompt or survey link in Beat 1 - the answers are churn-reduction's raw material. Suppressing non-responders is a win, not a loss: it protects deliverability and keeps the list healthy.
Worked artifact: three-email sequence, Tier 2 (B2B tool, 60-90 days inactive)
EMAIL 1 - day 0 - subject: "You've been missed at [FILL: product]"
Open: "You haven't logged in since [FILL: date] - no guilt, just noticed."
Body: one concrete value recap using THEIR data ("your team shipped
14 reports in your first two months").
Ask: one-question reply prompt: "What pulled you away?"
No incentive.
EMAIL 2 - day 7 - subject: "What's new since you left"
Body: 2-3 shipped improvements, each one sentence, most relevant first.
Ask: "Log back in and see" - deep link straight to the improved feature.
Incentive: soft - "your data and settings are exactly where you left them."
EMAIL 3 - day 14 - subject: "Last email from us - 1 free month if you're in"
Body: plain statement that this is the final email and why.
Offer: ONE incentive - 1 free month on re-activation, expires in 7 days.
Close: "If now's not the time, no hard feelings - we'll stop emailing."
After: non-responders suppressed from marketing sends. Actually.
Bad Beat 1 opener: "We have exciting news and a special offer just for you!" Good Beat 1 opener: "You haven't been around since March. Fair enough - here's what you were getting from us, and one question: what changed?"
Deliverable
Produce a campaign spec containing: the dormancy definition, the three tiers with segment sizes and per-tier offers, the full three-beat copy outline per tier (subjects, openers, asks, incentive, send-day), the suppression rule, and the measurement plan naming the re-activation event.
Do NOT
- Do NOT send one blast to all dormant users - a 35-day lapser and a 200-day ghost need different messages and different offers.
- Do NOT open with a discount to recently lapsed users; it trains lapsing and skips the cheaper value-first win.
- Do NOT stack multiple small incentives; one strong offer, proportional to segment LTV.
- Do NOT send a fourth email to non-openers; if three beats got nothing, more volume only damages deliverability.
- Do NOT call an email "final" and keep sending - the honesty rule is load-bearing.
- Do NOT report opens as success; only the re-activation event counts.
Quality bar
The spec ships only when: dormancy has a hard numeric threshold per the product's cadence; every tier has a distinct offer that scales with inactivity depth; Beat 1 names the absence plainly and asks why they left; Beat 3 contains exactly one incentive with a real deadline; the suppression rule is written into the automation, not left as intent; and the primary metric is re-activation rate.
Adjacent jobs
For recovering an abandoned checkout within hours or days, use abandoned-cart-sequence. For diagnosing why active users cancel and fixing the leak upstream, use churn-reduction (feed it the Beat 1 survey replies). For ongoing lifecycle flows rather than a bounded re-engagement push, use email-drip-builder.