If you’ve ever sent a product message and watched it go ignored, you know the pain of shouting into the void. The fix? Target the right people at the right moment—with data, not guesswork. This guide is for product managers, marketers, and anyone using Heap to make sure messages land with users who’ll actually care (and maybe even click).
Why Behavioral Cohorts Matter (and Where Most Go Wrong)
A behavioral cohort is just a fancy way of saying “group of people who’ve done specific things in your product.” Think “users who tried feature X but never came back” or “folks who invited a teammate in the last 30 days.” If you’re blasting everyone with the same nudge, you’re wasting your chance to move the needle.
Tools like Heap promise you can slice and dice users by what they actually do. But here’s the dirty secret: most teams either get lost in the weeds or overthink it. The result? Overcomplicated segments that never get used, or “everyone who signed up” lists that aren’t actionable.
Let’s cut through the noise and set up behavioral cohorts in Heap that you’ll actually use for targeted messaging.
Step 1: Get Clear on the Message First
Don’t start with the data. Start with what you want to say, and who should hear it.
Ask yourself:
- What’s the goal of this message? (e.g., drive adoption, re-engage, upsell)
- Who shouldn’t get it? (Don’t annoy power users with “get started” tips)
- What action or inaction signals someone’s ready for this nudge?
If you can’t answer those, take five and figure it out. The best cohorts start with a real use case, not “let’s see what Heap can do.”
Example:
Suppose you want to prompt users to try your new reporting dashboard. You don’t want to bug people who already used it, or brand new signups who haven’t even onboarded yet.
Step 2: Map Out the Behaviors That Matter
Heap tracks a ton of stuff out of the box—pageviews, clicks, form submissions, and custom events. For your cohort, pick a clear, unambiguous behavior.
Some ideas:
- Adoption: Clicked “Export” button, used the dashboard, invited a teammate
- Engagement: Logged in 3+ times in a week, completed a workflow
- Churn risk: Haven’t logged in for X days, started but didn’t finish setup
Pro tips:
- Don’t get too clever. “Users who visited the pricing page 7 times but never clicked Sign Up” is usually overkill.
- Make sure you’re tracking the event you care about. If it’s not in Heap yet, set it up (or ask your dev).
What to skip:
Don’t bother with vague behaviors like “visited the homepage.” Focus on actions that tie directly to your message.
Step 3: Set Up Events in Heap (If Needed)
Before you build a cohort, make sure Heap is capturing the right events. Out-of-the-box tracking is great, but sometimes you’ll need more.
- Check your event list: Go to the Events tab in Heap. Is your key behavior already tracked?
- Define new events: If not, use Heap’s point-and-click event definition (for things like button clicks) or ask your dev to send a custom event (for stuff like “completed onboarding”).
- Name events clearly: Call it “Clicked Export Button,” not “Button123.” You’ll thank yourself later.
Pitfall to avoid:
If you’re not sure what an event means (“Is ‘Dashboard Viewed’ any time someone loads the page, or only after data loads?”), ask. Don’t assume.
Step 4: Build Your Behavioral Cohort
Here’s where Heap makes it easy (and where most people overcomplicate things):
- Go to the "Cohorts" section in Heap.
- Click "New Cohort."
- Set your criteria:
- Pick the event(s) that signal the behavior you care about.
- Add timeframes—e.g., “did this in the last 14 days” or “has never done this.”
- Stack criteria as needed (“has done X, but not Y”).
- Preview your cohort: Heap will show you how many users match. If the list looks way off, double-check your logic or event definitions.
- Name your cohort clearly: Use names like “Tried Reports - Not Adopted” or “Invited Teammates - Last 30 Days.” Avoid “Test Cohort 4.”
Examples:
- Onboarding nudge: Users who signed up but haven’t completed onboarding in 7 days.
- Feature promo: Users who have used the product 5+ times but never tried the new dashboard.
- Re-engagement: Users who were active last month but haven’t logged in this month.
What not to do:
Don’t make “kitchen sink” cohorts with 10+ criteria. Start simple—you can always refine later.
Step 5: Sync Your Cohort with Messaging Tools
Cohorts are only useful if you can act on them. Heap lets you sync cohorts to common messaging platforms (like Intercom, Braze, or HubSpot), either natively or via CSV export.
- Check your Heap plan: Some integrations are only available on higher tiers. CSV export is the fallback.
- Set up the sync: Follow Heap’s docs for your messaging tool. Usually, it’s a one-time setup.
- Automate updates: Most integrations let cohorts auto-update, so users enter/exit as their behavior changes. Don’t settle for static lists.
Pro tip:
Test with a small cohort before rolling out a big blast. Always QA—send yourself the message first.
Step 6: Craft Targeted Messages (and Don’t Overdo It)
Now that you have a dynamic, behavior-driven list, write your message for that group, not for “everyone.”
- Reference the behavior: “We noticed you haven’t tried the new dashboard yet…”
- Keep it short and actionable.
- Offer a clear next step (try, learn more, schedule a demo).
Don’t:
- Nag users who’ve already done the thing.
- Send too many nudges—burnout is real.
- Use fake urgency (“Act now!”) unless it’s actually urgent.
What works:
Personal and relevant always beats generic. If you can’t imagine a real person being happy to get this message, rewrite it.
Step 7: Measure and Iterate
Here’s where most teams drop the ball. If you don’t measure, you’ll never know if your segment or message worked.
- Track message opens/clicks in your messaging tool.
- Watch downstream behavior in Heap: Did more people adopt the feature? Did engagement improve?
- A/B test if possible: Try different messages or criteria to see what actually moves the needle.
Watch out for:
- Small cohorts (under a few hundred users) can be noisy—don’t over-index on one campaign.
- Don’t assume causation. If usage jumps, make sure it’s tied to the message, not just seasonality or a big launch.
What Works, What Doesn’t, and What to Ignore
Works: - Starting with a clear goal and mapping behaviors to that goal. - Keeping cohorts simple and focused. - Syncing cohorts automatically to messaging tools.
Doesn’t: - Overcomplicating cohort logic (too many filters = confusion). - Ignoring event quality—garbage in, garbage out. - “Set it and forget it”—cohorts need to be reviewed as your product evolves.
Ignore: - Shiny features you don’t need. If you’re not sure what “propensity scoring” or “AI segmentation” is, you probably don’t need it yet. - Vanity metrics (“Total users messaged!”) unless they tie to real product outcomes.
Keep It Simple, Ship, and Iterate
Building behavioral cohorts in Heap isn’t rocket science, but it does take a bit of discipline. Start with a real problem, define a clear behavior, and keep your cohort logic tight. Don’t get lost in the data. Ship your first targeted message, see what happens, and tweak from there.
Remember: The best segmentation is the one you actually use. Don’t wait for perfect—get something live, and improve as you go.