If you’re trying to figure out if users actually stick around in your app, and you use Pendo, this guide’s for you. We’ll walk through how to set up retention cohorts—the right way—and cut through the noise on what’s worth paying attention to (and what’s just dashboard fluff).
You don’t need to be a data scientist, but a little patience and skepticism will help. Let’s get into it.
What’s a Retention Cohort, and Why Bother?
Before you start clicking around, it’s worth being clear: a retention cohort is just a group of users who started using your app around the same time. The point is to see if those people keep coming back—or if they bounce after the first login.
Why does this matter? Because it’s the best way to spot if your app’s actually “sticky,” or if new users try it once and disappear. Retention is a blunt but honest metric. If you’re in product, growth, or customer success, you need to know if improvements are paying off. Vanity metrics like daily active users are fine, but retention tells you if you’re building something people want to use again.
Step 1: Get the Right Data Into Pendo
First things first—Pendo only works with the data you give it. If you haven’t set up user tracking properly, your cohorts will be garbage-in, garbage-out.
Checklist: - Install the Pendo snippet in your app. Without this, nothing works. - Pass meaningful user IDs. Don’t use anonymous IDs unless you’re OK with muddy data. Tie actions to real, consistent user identifiers. - Feed in account info. If you care about team or company-level retention, make sure you pass account IDs too.
Pro tip: If your engineers say “it’s all set up,” double-check in Pendo’s Visitor and Account lists. If you see lots of “Anonymous” or weird IDs, fix that now. Don’t move on until this is clean.
Step 2: Define Your Cohort—Decide Who (and What) You’re Measuring
Don’t let the tool pick your cohorts for you. Decide: - Who counts as “new”? Is it sign-ups, first logins, first feature use? Be specific. - Which segment matters? All users? Only paying customers? Free trials? Pick a group that actually matters to your business.
How to do it in Pendo: 1. Go to People > Segments. 2. Click Create Segment. 3. Set rules for your cohort. For example: - “First visit date is between Jan 1 and Jan 7” - “Account type is ‘Pro’” - “Signed up via marketing campaign XYZ”
Don’t overcomplicate this: Start simple. If you try slicing 10 different ways at once, you’ll drown in data.
Step 3: Build a Retention Report in Pendo
Now for the main event: plotting your cohort’s retention.
- Navigate to Retention: In Pendo, go to Behavior > Retention.
- Choose your cohort: Pick the segment you defined earlier.
- Pick an activity to track: You need a clear definition of “retained.” Usually, it’s “any app visit,” but you can (and often should) track specific feature usage.
- Set your time window: Most teams use daily or weekly retention. If your product isn’t used daily, weekly is more honest.
- Run the report.
You’ll see a table (and maybe a chart) showing what percent of your cohort returns each period.
Pay attention to: - Day 1/Week 1 drop-off: This is where you’ll see if your onboarding is working. - Long-term tail: If you have users sticking around after 30 days, you’re doing something right. - Weird spikes or dips: Sometimes a product launch or bug will stick out. Don’t ignore these—investigate.
Step 4: Interpret the Results Honestly
Here’s where most teams go wrong: they see a “retention curve” and pat themselves on the back, but don’t dig deeper.
What to look for: - Steep drop-offs after signup? Your onboarding or first-use experience probably needs work. - Flatlining at zero after a week? You might be attracting the wrong users, or your product isn’t solving a real problem (yet). - Consistent, small group of loyal users? Find out what keeps them coming back—interview them if possible.
Don’t get distracted by: - Short-term bumps: A marketing push or new feature can inflate numbers for a week. Watch for lasting change. - Tiny sample sizes: If your cohort is 5 people, the data is basically noise. - Comparing apples to oranges: Don’t compare free users to paying customers. Segment clearly.
Step 5: Take Action (and Avoid Analysis Paralysis)
Data’s only useful if you do something with it. Here’s what to consider:
- If retention’s low: Don’t spin up another dashboard. Fix your onboarding, talk to churned users, or simplify your user flow.
- If retention improves after a change: Double-check it’s not a fluke. Wait a few weeks, and compare new cohorts.
- If you see unexpected trends: Dig deeper. Sometimes, a core group uses one feature heavily—maybe that’s where you should focus.
Pro tip: Set a calendar reminder to check your retention monthly. It’s easy to forget once the novelty wears off.
What Actually Works, and What’s Hype
What works: - Simple, consistent cohorts (like “all signups in March”). - Tracking real feature usage, not just logins. - Comparing before/after for product changes.
What doesn’t: - Chasing high retention for vanity’s sake. If users log in but don’t do anything meaningful, it’s a false positive. - Over-segmenting. You don’t need 30 cohort reports. Start broad, then get granular if you see something interesting. - Ignoring user feedback. Numbers are a starting point, not the whole story.
Ignore: - “Industry benchmarks” for retention unless they’re from companies with the same business model as you. SaaS, consumer apps, and B2B tools all have wildly different norms.
Common Pitfalls (and How to Dodge Them)
- Messy user IDs: If your data’s dirty, your retention numbers are fiction.
- Misaligned cohorts: Make sure your cohort definitions match your business goals.
- Overfitting: Don’t slice data until you find a story you like. Look for real trends, not just what feels good.
Wrapping Up: Keep It Simple, Iterate Often
Cohort analysis in Pendo isn’t rocket science, but it’s easy to overcomplicate or misread. Start with clear, simple cohorts. Check your retention regularly. When you see something odd, dig in—but don’t get lost in the weeds.
Remember: the best retention strategy is building something people want to come back to. The charts just help you spot if you’re on the right track. Keep it simple, act on what you learn, and tweak as you go.