Win-Back Campaigns with Braze

What Win-Back Campaigns Actually Require

Re-engaging churned users is not a messaging problem. It is a data and timing problem. Most win-back campaigns fail because they send the right message at the wrong moment, or they treat all lapsed users the same way. Braze gives you the infrastructure to do this correctly — but only if you configure it with precision.

This guide walks through a complete implementation of win-back campaigns in Braze, from user segmentation to multi-channel orchestration to measurement. Assume you already have Braze integrated and events flowing into the platform.

---

Defining "Lapsed" Before You Build Anything

Your win-back campaign is only as good as your lapse definition. Braze lets you create segments based on event recency, but you need to decide what "churned" means for your product before opening Canvas.

Common lapse thresholds by product type:

• Mobile app (daily active): 7–14 days without a session
• SaaS product (weekly active): 30–45 days without a login or core action
• E-commerce: 60–90 days without a purchase
• Subscription product: Any user who cancelled but has not yet lost access

In Braze, define these using Segment Extensions rather than standard segments. Segment Extensions let you query event data beyond the standard 30-day lookback window — up to 2 years — which matters when your lapse threshold is 60+ days. Create a Segment Extension using a Last Did rule: `Has not performed [core_event] in the last [X] days`.

Combine this with profile attributes like subscription tier, acquisition channel, or last purchase value to build sub-cohorts. A user who spent $400 and churned is not the same re-engagement opportunity as a user who completed one session and left.

---

Building the Win-Back Flow in Canvas

Canvas is Braze's multi-step journey builder. For win-back campaigns, use the Canvas Flow architecture (not the legacy Canvas). Canvas Flow gives you full control over step types, branching logic, and exit criteria.

Step 1: Entry Setup

Use a Segment Entry trigger — not an event-based trigger. Win-back campaigns target users who have already stopped acting, so there is no real-time event to listen for. Schedule the Canvas to evaluate entry eligibility once per day.

Set your entry audience to the Segment Extension you built above. Enable entry controls to allow re-entry after 30 days so users who lapse again can re-enter the campaign.

Step 2: Channel Decision Split

Not every lapsed user has the same reachable channels. Before your first message, add an Action Paths step — but use it as a pre-send channel check.

Branch logic:

1. Push enabled AND email subscribed → Full multi-channel sequence
2. Email only → Email-only sequence
3. Push only → Push + in-app sequence
4. Neither → Exit Canvas (do not waste the flow on unreachable users)

Step 3: Message Sequence

Structure the sequence across three phases.

Phase 1 — Re-introduction (Day 0):

Send a high-value reminder. This is not a discount. Surface something specific to the user's past behavior using Liquid personalization. Braze's Liquid templating lets you pull in `custom_attributes` like `last_product_viewed` or `last_category_browsed` directly into copy.

Example Liquid snippet:

```

{% if {{custom_attributes.${last_category}}} != blank %}

You left off in {{custom_attributes.${last_category}}}.

{% else %}

Here is what is new since your last visit.

{% endif %}

```

Phase 2 — Incentive (Day 5, if no conversion):

Use an Action Paths step after Day 0 to check whether the user completed your conversion event. If they did, exit them from the Canvas. If they did not, advance them to an incentive message — a discount, free trial extension, or exclusive content depending on your model.

Phase 3 — Final Attempt (Day 12, if still no conversion):

One more message. Keep it short. Acknowledge the gap directly. After this step, if the user still has not converted, suppress them from re-entry for 60–90 days. Continuing to message truly uninterested users damages your sender reputation and inflates suppression lists.

Step 4: Exit Criteria

Set global exit criteria at the Canvas level: any user who performs your core conversion event exits immediately. This prevents a user who re-engaged on Day 2 from receiving a discount offer on Day 5.

---

Personalization With Connected Content

If your Braze setup includes a recommendation engine or a real-time API, use Connected Content to pull dynamic data into your messages at send time. This is particularly effective for e-commerce win-backs where you want to surface products similar to what the user last purchased.

Connected Content makes a GET request to your endpoint and injects the response into the message template. Set a caching duration of 4–12 hours for recommendation data to avoid excessive API calls at high send volumes.

---

Measurement and Iteration

Track these metrics inside Canvas Analytics:

• Unique recipients per step — identifies where users are dropping off
• Conversion rate by variant — if you are running an A/B test on incentive timing
• Revenue per recipient — requires configuring a purchase event as your conversion event
• Unsubscribe rate per step — a rate above 0.5% on any single message is a signal to revise

Use Canvas Variants to test the incentive timing. Run a 50/50 split: one group receives the incentive on Day 3, the other on Day 7. Let it run for at least one full lapse cycle before making a decision.

---

Limitations of Braze for Win-Back Campaigns

Braze is strong on orchestration and personalization but has real gaps worth knowing before you build.

• Predictive churn is not native. Braze has a Predictive Churn feature (available on higher-tier plans), but it uses a generic model trained on session data. If your product has a complex behavioral signal for churn, you will get better results building a churn model externally and pushing the output back to Braze as a custom attribute.
• Segment Extensions have a refresh lag. Extensions do not update in real time — they refresh on a schedule (every 24 hours by default). If you need same-day lapse detection, use event-triggered Canvases instead and maintain lapse status as a profile attribute updated by your backend.
• Cross-device identity requires work. If a user churns on mobile but converts on desktop, Braze needs a resolved user identity via your SDK integration. Anonymous users who never logged in will not match back to the lapsed profile.
• SMS throughput costs. Win-back via SMS works well, but Braze charges per SMS segment. At scale, a three-touch SMS win-back sequence can become expensive quickly. Model this before adding SMS to the channel mix.

---

Frequently Asked Questions

How long should a win-back Canvas run before I declare a user permanently churned?

Thirty days of active Canvas engagement with no conversion is a reasonable cutoff for most products. After that, move the user to a long-term suppression segment and plan a once-per-quarter re-engagement attempt. Continuing to message beyond 30 days with no response generates unsubscribes and spam complaints at a rate that hurts your overall deliverability.

Can I use Braze's AI features to improve win-back targeting?

Yes, with caveats. Intelligent Timing optimizes send time based on individual engagement history, which is worth enabling on Phase 1 and Phase 2 messages. Intelligent Channel selects the highest-performing channel per user. Both features require sufficient historical engagement data — they work poorly on users who were always low-engagement before churning. Evaluate whether your lapsed cohort has enough signal before enabling them.

What is the right entry volume for testing a new win-back Canvas?

Start with 10–15% of your eligible lapsed cohort as a test group before rolling out fully. This gives you enough volume to reach statistical significance on conversion metrics while limiting the risk of messaging your entire lapsed base with an unproven sequence. For a cohort of 50,000 lapsed users, 5,000–7,500 is a workable test group.

Should win-back campaigns be separate from lifecycle Canvases or nested within them?

Keep them separate. Win-back campaigns have fundamentally different entry criteria, message tone, and success metrics than onboarding or retention Canvases. Nesting them creates dependency problems — a change to your retention Canvas can inadvertently break exit logic for your win-back flow. Maintain clean separation and use global frequency capping in Braze to prevent users enrolled in other Canvases from receiving win-back messages simultaneously.