Every user interaction is a funnel. Most PMs only look at the top (acquisition) and the bottom (conversion or retention). What happens in the middle is where the product either earns or loses users — and most teams have almost no visibility into it.
Funnel analysis is one of those tools that seems simple and is actually very sophisticated. The idea: map out the steps users take to accomplish a goal, track how many users complete each step, and identify where they fall off. Fix the biggest drop-offs. Repeat.
In practice, funnel analysis fails product teams in predictable ways. They build the wrong funnels. They analyze drop-offs without understanding why they're happening. They optimize individual steps without understanding how changes to one step affect the rest of the funnel. And they miss the funnels that matter most — the ones that happen between their defined steps.
Why Most Funnel Analyses Miss the Point
The most common funnel analysis mistake: defining the funnel as the steps you can measure rather than the steps users actually take. If your analytics tracks sign-up, email verification, profile setup, and first action — you'll build a funnel with those four steps and optimize for completion of each. But if users have to do three things between "email verification" and "profile setup" that you're not tracking, you're optimizing within a funnel that doesn't represent the real user journey.
The second mistake: treating all drop-offs as problems to be fixed. Some drop-offs are healthy. If 60% of users who reach your paywall don't subscribe, that's not necessarily a paywall problem — it might be a qualification problem. You're attracting users who were never going to convert, and the paywall is doing its job of separating them from the buyers. Trying to improve that conversion rate might mean converting users who churn quickly, which is worse than a lower conversion rate of long-term subscribers.
How to Build Funnel Analysis That's Actually Useful
Start from the user's goal, not your product's flow
Define the funnel by what the user is trying to accomplish, not by the screens they pass through in your product. "User wants to book a service" is a user goal. "User visits homepage → searches → views listing → contacts provider → books" is a funnel. The funnel should map the path to the user's goal, not your product's navigation.
Segment your funnels by user type
Aggregate funnel data is often misleading because different user segments have different funnels. New users have a different path than returning users. Mobile users have different conversion rates than desktop users. Enterprise users navigate differently than SMB users. Running funnel analysis on undifferentiated user populations produces averages that may not describe any actual user group accurately.
Pair quantitative drop-offs with qualitative research
Funnel data tells you where users are falling off. It doesn't tell you why. A 40% drop-off at step 3 could mean: the step is confusing, the step is too long, users get a phone call and never come back, users find what they need in step 2 and don't need step 3, or users are discovering that the product doesn't solve their problem and are self-selecting out. Each of these diagnoses leads to a completely different solution. You need qualitative research — user interviews, session recordings, or on-page surveys — to determine which one is actually happening.
Diagnosing your conversion funnel?
I've helped teams identify the real drop-off points in their user journeys — the ones that traditional funnel analysis misses. If you're seeing unexplained drop-offs or conversion stagnation, let's find out what's actually happening.
Work with Yogesh →