Re-engagement Campaigns: When to Fire Them, When Not To
Re-engagement campaigns are the easiest part of mobile growth to get expensively wrong. Teams fire push and email at every lapsed user, the unsubscribe rate climbs, and the campaigns that should recover real revenue burn out their channels before they get a chance. The right approach is more selective: identify the lapsed cohorts that are worth recovering, ignore the ones that are not, and time messages to value moments instead of calendar days.
This article is the re-engagement decision framework Semnexus uses with mobile apps. It covers the four cohorts to target, the three to ignore, the message timing that works, and the channels that earn their place.
What re-engagement actually is
Re-engagement is the deliberate process of bringing a lapsed user back into the app at a meaningful enough depth that they continue using it. Two clarifications:
- It is not the same as retention. Retention keeps a user active; re-engagement re-activates a user who stopped.
- It is not the same as winback. Winback targets cancelled or churned paying users specifically; re-engagement is broader.
The cost-benefit on re-engagement is usually attractive: re-activating a lapsed user is 3 to 7 times cheaper than acquiring a new one. But the cost-benefit assumes the right users are targeted.
The four cohorts worth re-engaging
Cohort 1: Recent lapse with prior depth
Users who used the app reliably for 14+ days, then went quiet within the last 30 days. They had clear value and recently lost the habit. Highest recovery rate of any cohort.
Recovery rate: 25 to 45% within 90 days of campaign start.
Cohort 2: Trial-end without conversion
Users who completed a free trial within the last 60 days without converting to paid. They demonstrated intent but did not pay. A well-timed offer or contextual nudge can convert a meaningful share.
Recovery rate (to paid): 8 to 18% within 30 days.
Cohort 3: Seasonal pattern lapsed users
Users whose engagement is tied to a season (fitness apps in January, finance apps at tax time, education apps in fall) and who are due back. Less about the lapse, more about timing the return.
Recovery rate: 15 to 30% within the seasonal window.
Cohort 4: Cancelled subscribers within 90 days
Subscribers who cancelled in the last 90 days for non-product reasons (price, life change, temporary disinterest). Winback-style outreach with a discount or a product update can recover a share.
Recovery rate: 5 to 12% within 60 days of cancellation.
The three cohorts to ignore
Targeting these cohorts produces zero or negative ROI:
Cohort 5: Users who never activated
Users who installed and used the app once or never at all. They did not experience the value. Re-engagement messaging cannot fix a missed activation; reacquisition through paid is more honest.
Cohort 6: Long-lapsed users (180+ days inactive)
Users inactive for more than 6 months have moved on. Recovery rates drop below 5%, and the unsubscribe damage to channel health is real. Skip.
Cohort 7: Cancelled subscribers for product reasons
Subscribers who churned because the product did not work for them. No discount or message recovers them; the right answer is product change, not re-engagement.
Message timing that recovers value
When you message matters more than what you message. Three timing principles:
Tie messages to value moments, not lapse durations
The wrong trigger: "User has been inactive for 14 days." The right trigger: "User completed a tracked workout, then stopped, then their typical workout day is tomorrow."
Calendar-day triggers produce generic messages. Value-moment triggers produce timely ones.
Avoid the first 7 days post-lapse
The strongest cohort (Cohort 1) is in a temporary lull, not a permanent lapse. Messaging too early reads as nagging and converts at lower rates than messaging at 10 to 21 days.
Cap frequency hard
Three messages per re-engagement window is the working maximum. After three unsuccessful messages, move the user to a quiet list for 60 days before trying again. Apps that send 7+ messages within a single re-engagement campaign accelerate unsubscribes.
Channel choices
The channels that earn their place, in order of cost-effectiveness:
1. Push notifications
Lowest cost; highest immediate response when the user has push enabled. Use for Cohort 1 and Cohort 3.
2. Email
Higher cost per message; reaches users who turned off push. Use for all cohorts where the user gave email.
3. In-app messaging
Only relevant if the user opens the app. Use for users who briefly opened and re-lapsed.
4. Paid re-engagement campaigns
Highest cost per message; reaches users who turned off everything. Use selectively for Cohort 2 (trial non-converters) and Cohort 4 (cancelled subscribers).
5. SMS
Highest cost and highest unsubscribe risk. Use only for high-value cohorts (Cohort 4 specifically) with explicit consent.
The 5-step re-engagement workflow
The Semnexus operational pattern:
- Define cohorts based on real behavioral data. Don't assume; query.
- Build a small library of message templates per cohort. 2–4 templates per channel per cohort is enough; more becomes noise.
- Tie sends to value-moment triggers, not calendar days. Requires product event instrumentation.
- Cap frequency to 3 messages per window. Stop after the third unsuccessful message.
- Measure recovery rate per cohort per channel. Drop cohort-channel combinations that fall below the recovery threshold.
Five mistakes that destroy re-engagement programs
The most common failures in audits:
- One generic re-engagement email blast to all lapsed users. Targets all seven cohorts at once. Average response rate is poor; unsubscribe rate is high.
- No frequency cap. Three messages becomes ten messages. Users leave the channel before they have a chance to come back.
- Discount-led messaging on Cohort 1. Users who lapsed for non-price reasons do not need a discount. The discount trains them to expect one.
- Re-engagement of users who never activated. Wastes channel health on users who cannot be converted by messaging.
- No measurement on recovery quality. Counting opens or clicks is not enough. Recovered users should reach a meaningful in-app event, not just open the app once.
How to measure if re-engagement is working
The minimum scorecard:
- Recovered user count. How many users returned to a value-defining event within the window.
- Recovery rate by cohort. Compare to the benchmarks above.
- Channel health. Unsubscribe rate per channel after the campaign vs before.
- Recovered-user LTV. Recovered users should retain comparably to organic returners. If they churn faster, the campaign is bringing back users who were not coming back willingly.
Frequently asked questions
When should I start running re-engagement campaigns? Around month 4 to 6 of an app's life, after the first lapsed cohorts have meaningfully accumulated. Earlier than that, the lapsed-user pool is too small to support a measurable program.
Are re-engagement push notifications subject to the same iOS policies as marketing push? Yes. They count toward the user's push permission decision and should respect quiet hours.
Should I personalize re-engagement messages with the user's name? Light personalization helps; heavy personalization is rarely worth the cost. The strongest personalization is contextual (what the user did, not who they are).
What about email deliverability for re-engagement? Lapsed users have lower engagement than active users, which drags down sender reputation. Use separate sending IPs or sub-domains for re-engagement campaigns to protect transactional and active-user email deliverability.
Where does paid retargeting fit? Paid re-engagement campaigns on Meta, TikTok, and Google work as a fourth channel option for users you cannot reach via push or email. Costs are higher; reach is broader.
If your re-engagement program is unfocused or you are unsure whether your current campaigns are healthy, the Semnexus mobile app marketing team runs lifecycle and re-engagement audits as part of every engagement. The app development team handles the product-side instrumentation work when value-moment triggers need to be wired up.
