If you work on a SaaS growth team, you’ve probably heard that “funnels” are the secret to unlocking product insights and driving signups. The reality: funnel analysis is powerful, but only if you set it up right and focus on the stuff that actually moves the needle. This guide walks through building funnel analysis reports in Amplitude, with a focus on what works, what’s a waste of time, and how to avoid getting lost in vanity metrics.
Let’s get into it.
What is Funnel Analysis (and Why SaaS Teams Should Care)
Funnels map the steps users take toward a goal—like signing up, starting a trial, or upgrading. By tracking how many users drop off at each step, you can spot friction, test ideas, and (hopefully) fix what’s broken.
But here’s the truth: A funnel report is only as good as the events you track and the questions you ask. If you just measure everything, you’ll end up with noise, not insight.
For SaaS, funnels are best for: - Understanding onboarding drop-off (where do new users get stuck?) - Testing new flows (did changing the signup flow help?) - Measuring conversion to key milestones (trial → paid, activation events, etc.)
If you’re just curious about “engagement” or “overall usage,” a funnel isn’t the right tool. Use it for step-by-step flows with a clear goal.
Step 1: Decide What Funnel to Build (Don’t Skip This)
Before you click anything in Amplitude, figure out what you’re trying to measure. This sounds obvious, but lots of teams end up with useless funnels because they didn’t start with a simple question.
Ask yourself: - What user behavior do we want to improve? (e.g., more people completing onboarding) - What are the actual steps users take to get there? - Is there a single, clear “success event” at the end?
Pro tip: Don’t try to analyze everything at once. Pick one conversion path that matters for your current growth goal.
Example:
You want to track your SaaS onboarding flow:
1. User signs up
2. User verifies email
3. User completes onboarding checklist
4. User creates their first project
If you can’t list the steps in plain English, you’re not ready to build a funnel.
Step 2: Make Sure Your Events Are Set Up (and Named Clearly)
Amplitude can’t track what you don’t send it. Before building a funnel, double-check that: - The events you need are instrumented in your product. - Event names actually make sense to humans (“Signup Completed” is better than “evt_45321”). - You’re tracking the right properties if you want to segment (like plan type, signup source, etc.).
Honest take: Most SaaS teams have at least one missing or mislabeled event in their funnel. If you find gaps, fix those first. Otherwise, your report will just be a fancy chart of bad data.
What to ignore: Don’t get hung up on tracking every micro-interaction. Stick to the main events that map to user value.
Step 3: Build the Funnel in Amplitude
Here’s how to actually create a funnel report:
- Go to “Analysis” → “Funnels” in Amplitude.
- Add your funnel steps, in order.
- Start with your entry event (e.g., “User Signed Up”).
- Add each subsequent step (e.g., “Verified Email,” “Completed Onboarding Checklist,” etc.).
-
Make sure you’re using the exact event names from your implementation.
-
Choose how to handle repeated events (optional).
- Amplitude lets you decide if you care about the first time, any time, or the most recent time a user did something.
-
Tip: Stick with “first time” for onboarding funnels, unless you have a reason to do otherwise.
-
Set the time window.
- How long do users have to complete the funnel? For onboarding, 7 or 14 days is typical.
-
Shorter windows = more actionable data. Longer windows can muddy the picture.
-
Apply filters if needed.
- Want to look at just paid users? Or people from a specific signup source? You can filter by event or user properties.
That’s it—you’ll see the classic waterfall chart showing how many users make it through each step.
Step 4: Interpret the Results (and Avoid Common Traps)
This is where most teams go wrong. It’s easy to stare at a funnel and overthink it, but here’s what actually matters:
Look for big drop-offs.
- If you lose half your users between steps 2 and 3, that’s not subtle—dig into why.
- Tiny changes (2-3% drop) aren’t always worth chasing unless you’re operating at huge scale.
Check your sample size.
- Funnels with 50 people are almost always noise. Wait until you have hundreds (or thousands) for a reliable signal.
Ignore “perfect” conversion rates.
- If your funnel shows 99% conversion, you’re probably missing an event, or your definition is too narrow.
Ask “what would we do differently based on this?”
- If a funnel step is vague (“Engaged with Product”), you’ll never know what to fix. Make steps concrete.
Pro tip: Use the “Conversion Over Time” view in Amplitude to see if your changes actually improved the funnel, or if you’re just seeing random noise.
Step 5: Segment and Compare (But Don’t Overcomplicate)
Segmentation is where funnels go from basic to actually useful. In Amplitude, you can break down your funnel by user property, event property, or cohort.
Start simple: - Free vs. paid users - Signup source (organic, paid, referral) - Device or platform
What to watch out for: - Don’t throw ten segments at a funnel and hope for magic. Pick 1-2 that actually map to how you think about your users. - If a segment is tiny, ignore it. You can’t make decisions off a dozen users.
Example:
If paid users convert at twice the rate as free users, you might want to dig into why. Maybe your onboarding is tuned for folks already motivated to pay, but confusing for everyone else.
Step 6: Share and Act on the Insights
The point of a funnel report isn’t to make a pretty dashboard—it’s to help your team make decisions.
How to do this well: - Take screenshots or share direct links to the Amplitude report with a short summary (“We’re losing 60% of users at email verification. Next step: test in-product reminders.”) - Don’t just share the chart—add your takeaway and the next experiment you’ll run. - Revisit the funnel after you ship changes to see if things actually improved.
What not to do:
- Don’t turn every funnel report into a weekly status update. Only share when there’s a clear story or decision to be made.
What to Ignore (and What to Watch For)
Ignore: - Funnels that require users to jump through unnatural hoops (“User clicked three different help docs, then upgraded”). If you have to squint to explain the flow, don’t bother. - Micro-steps that don’t change behavior (like “Opened Settings” if it’s not critical to your flow). - Any funnel built “because the boss asked”—unless it’s tied to a real goal.
Watch for: - Steps that aren’t tracked (missing events = missing insight). - Funnels with super-high conversion rates—they’re usually hiding a data problem. - Overcomplicating with too many segments or steps. Simple is actionable.
Summary: Keep It Simple, Iterate Fast
Funnels are a great tool, but only if you keep them focused and grounded in real user behavior. Don’t get distracted by every possible flow or segment. Start with one key path, make sure your data is clean, and use the insights to run experiments. Then repeat.
If your funnel report doesn’t lead to an experiment or decision, it’s just busywork. Keep it simple, ship fast, and let real-world results guide what you do next.