If you’re drowning in session recordings and heatmaps, you’re not alone. Smartlook can capture everything your users do, but that’s only useful if you can actually find what matters. This guide is for product managers, UX folks, and anyone tired of guessing what needs fixing. We’ll cut through the noise and show you how to filter and segment user sessions in Smartlook so you can make targeted, real-world improvements—without wasting hours watching irrelevant recordings.
Why Filtering and Segmenting Sessions Actually Matters
Before you start, let’s get one thing clear: staring at random session replays hoping for insight is a waste of time. Good filtering and segmentation help you:
- Spot patterns that actually affect your KPIs
- Zero in on user groups (new users, churn risks, power users)
- Debug specific issues, not just generic “friction”
- Avoid “analysis paralysis” and focus on what you can fix
If you just want to feel productive, watch sessions at random. If you want answers, use filters and segments.
Step 1: Get Your Tracking Basics Right
Don’t skip this. Filtering only works if your Smartlook setup is capturing the right data. Here’s what you really need:
- Events: Button clicks, form submissions, errors—set these up as custom events. The default stuff (like page views) won’t tell you much by itself.
- User Properties (a.k.a. user identity and traits): Pass things like user IDs, plans (“free” vs. “pro”), or signup dates. This is gold for segmentation.
- Error Tracking: Make sure errors (JS errors, failed API calls) are tracked as events, so you can filter for “broken” sessions.
Pro tip: If your developer shrugs at “user properties,” show them Smartlook’s docs. It’s a few lines of code, not a month-long project.
Step 2: Use Smartlook’s Filters—But Don’t Overcomplicate
Once your data’s coming in, head to the session recordings dashboard. Here’s how to actually use filters without getting lost:
The Filters That Actually Matter
- Visited Page/Screen: Filter for users who visited a specific page (e.g., checkout). This is the fastest way to zoom in on “problem” flows.
- Event Occurred: Only show sessions where a specific event fired (like “payment_failed” or “signup_complete”).
- User Properties: Slice by cohort—new users, paying customers, users on mobile, etc.
- Duration: Long sessions aren’t always better, but super short ones (“bounce”) can flag onboarding issues.
- Device/Browser: Got a bug report “only on Safari”? Filter by browser/device and save hours.
Stuff to mostly ignore: - Filtering only by “country” or “referrer” is usually a dead end unless you’re chasing a known geo bug or campaign issue. - “Entry page” is less useful than you think; focus on outcome, not just where people started.
Example: Find Real Drop-off Points
Want to see why people abandon signup? Try:
- Filter sessions where “Visited Page” includes your signup page.
- Add “Event did NOT occur” > “signup_complete”.
- Optional: Filter for new users only.
Now you’re watching sessions that tried—but failed—to sign up. Way more useful than just looking at heatmaps.
Step 3: Build and Save Segments for Reuse
Manual filtering is fine once, but real value comes from saved segments. Here’s how:
- Create a Segment: After filtering, save your query as a segment (“Churn Risks”, “Mobile Power Users”, “Failed Checkouts”, etc.).
- Share with Your Team: Most product teams waste time recreating filters. If you find a good segment, share the link.
- Iterate: Segments aren’t forever. If a segment stops surfacing new insights, tweak or kill it. Don’t let your workspace turn into a graveyard of stale segments.
Honest take: “Out-of-the-box” segments are usually too generic. Build your own based on real business questions.
Step 4: Pair Segments with Funnels for Deeper Insight
Filtering tells you who did what. Funnels tell you how many did or didn’t complete a sequence. Combine them:
- Build Funnels: Set up a funnel for key flows (signup, checkout, onboarding).
- Filter by Segment: Want to see if mobile users or a specific cohort drop off more? Apply your saved segment to the funnel.
- Watch Targeted Sessions: Click into the funnel drop-off step, then watch those specific session recordings.
This lets you stop guessing why your funnel conversion sucks—you’ll see exactly where people get lost or frustrated.
Pro tip: Don’t build a million funnels. Start with the flows that matter to your core metrics, then add more as needed.
Step 5: Use Notes and Labels (But Don’t Become a Librarian)
While watching sessions, use Smartlook’s Notes or Labels features to mark important moments:
- “User got stuck on step 3”
- “Typed password twice—UI unclear”
- “Weird bug on mobile nav”
This helps you and your team jump straight to the good stuff later. But don’t go overboard. If you’re labeling every mouse move, you’ll never ship anything.
What to ignore: Fancy color-coded taxonomies. Just write what matters, in plain English.
Step 6: Avoid Common Pitfalls and Red Herrings
Filtering and segmentation are powerful, but it’s easy to fall into these traps:
- Over-segmenting: The more filters you layer on, the fewer sessions you get. If you’re down to 1–2 sessions, you’re probably too narrow.
- Chasing Edge Cases: Not every outlier matters. Focus on patterns that repeat, not one-off “weird” sessions.
- Assuming Correlation = Causation: Just because users from a certain segment fail more often doesn’t mean that’s the root cause. Watch the sessions, look for real blockers.
- Ignoring Sample Size: Watching 3 sessions and declaring victory is bad science. Try to view a decent range before making big decisions.
Step 7: Make It Actionable—Not Just “Interesting”
You’re not here to collect “insights”—you want product improvements. After using filters and segments, ask yourself:
- What did I actually learn that I can act on?
- Can I write a clear ticket or hypothesis based on this?
- Did this session pattern tie back to a business goal (retention, conversion, support volume)?
If the answer’s no, go back and refine your filters or segments. If yes, congrats—you’re doing it right.
Wrap-Up: Keep It Simple, Iterate Often
Filtering and segmenting in Smartlook isn’t about chasing every possible user journey—it’s about focusing on what actually moves the needle. Start simple, save your best segments, and use what you find to make concrete changes. You can always go deeper later, but don’t let “analysis” get in the way of action.
Just remember: It’s better to deeply understand a few key flows than to drown in endless session recordings. Keep it real, keep it useful, and keep improving.