Funnel Analysis: How to Find Where Users Drop Off
Funnel analysis is a method for tracking how users move through a defined sequence of steps toward a goal — and, just as importantly, where they drop off along the way. Whether the goal is completing a purchase, finishing onboarding, or submitting a lead form, funnel analysis turns a vague sense that “people are leaving” into a precise, stage-by-stage map of where attention and intent leak away.
Aggregate conversion rates tell you that something is working or failing, but they never tell you where. Funnel analysis fixes that. By breaking a journey into ordered steps and measuring the proportion of users who advance from one to the next, you can isolate the single step that costs you the most users and focus your effort where it actually moves the number. This guide explains what funnel analysis is, how to build a funnel correctly, how to read the results, and the mistakes that quietly distort the conclusions teams draw from it. It pairs naturally with broader marketing analytics work and with cohort analysis, which adds a time dimension to the same data.
TL;DR — Funnel Analysis Essentials
- Funnel analysis measures the percentage of users who advance through each step of an ordered journey toward a goal
- The biggest single drop-off step is almost always the highest-leverage place to invest
- Step-to-step conversion (not just overall conversion) is what reveals where to act
- Time windows, step ordering, and entry definitions all change the numbers — define them deliberately
- Segmenting a funnel by source, device, or user type usually reveals more than the aggregate funnel
- Funnels describe where users leave; pairing them with qualitative data explains why
In This Guide
What Is Funnel Analysis
A funnel is an ordered series of steps a user is expected to complete to reach an outcome. Funnel analysis measures how many users enter the funnel, how many reach each subsequent step, and how many complete the final goal. The shape is called a funnel because each step typically has fewer users than the one before it — the population narrows as you move toward conversion.
The term covers two distinct measurements that are easy to conflate. Overall conversion is the percentage of users who start the funnel and reach the end. Step conversion is the percentage who advance from one specific step to the next. Overall conversion is a scoreboard; step conversion is a diagnosis. A funnel with a respectable 12% overall conversion might hide a single step where 70% of users vanish — and that step, not the headline number, is where the work belongs.
Overall conversion tells you how the funnel performs. Step conversion tells you what to fix. Always report both.
Why Funnel Analysis Matters
Most growth work suffers from a targeting problem: teams pour effort into steps that are already performing while ignoring the one quietly bleeding users. Funnel analysis solves this by ranking your steps by lost users, so the highest-leverage opportunity is unmistakable.
Consider the compounding math. If a five-step funnel converts 80% at each step, the overall conversion is 0.8⁵ ≈ 33%. Improve a single weak step from 80% to 90%, and overall conversion climbs to roughly 37% — a meaningful lift from one focused change. Funnel analysis is what tells you which step that should be.
It also reframes how you measure success. A campaign that drives more traffic but pushes those users into a step that converts poorly may produce fewer completions than a smaller, better-matched audience. Without a funnel view, that failure looks like success, because the top-of-funnel number went up.
| Step | Users | Step Conversion | Drop-off |
|---|---|---|---|
| Landing page view | 10,000 | — | — |
| Product page | 6,200 | 62% | 3,800 |
| Add to cart | 2,480 | 40% | 3,720 |
| Checkout started | 1,860 | 75% | 620 |
| Purchase complete | 1,450 | 78% | 410 |
In the table above, overall conversion is 14.5%. But the two steps with the worst step conversion — product page to add-to-cart (40%) and landing to product (62%) — account for the vast majority of lost users. That is where attention belongs, not at checkout, which already retains three of every four users who reach it.
Types of Funnels
Funnels appear throughout analytics under different names, but they share the same structure. The common varieties differ in what they measure and where in the lifecycle they sit.
Acquisition and conversion funnels
These track a user from first touch to a commercial outcome: visit → engage → sign up → purchase. They are the funnels most teams build first, because they map directly to revenue.
Onboarding and activation funnels
These measure whether new users reach the moment they experience a product’s core value — often called activation. Steps might be: account created → first project → first invite → first result. Drop-off here predicts long-term retention better than almost any other signal.
Engagement and feature funnels
These examine a specific workflow inside a product — for example, the steps to publish content or complete a key action. They reveal friction in features that are already in use, where small fixes can have outsized impact.
Name each funnel by the question it answers. “Where do trial users stall before activation?” produces a sharper funnel than a generic “signup funnel,” because it forces you to define the steps that matter.
The Anatomy of a Funnel
Every funnel is built from a small set of decisions, and each one changes the numbers. Getting them explicit is the difference between a funnel you can trust and one that quietly misleads.
- Entry definition — who counts as having entered? All visitors, or only those who reach a specific page or perform a specific action? This determines your denominator.
- Steps and order — the ordered events that make up the journey. Order matters: a strict funnel requires steps in sequence, while a relaxed one allows them in any order.
- Conversion window — how long a user has to complete the funnel. A one-hour window and a 30-day window can produce dramatically different conversion rates for the same data.
- Counting unit — are you counting users, sessions, or events? A user who tries checkout three times is one user but three sessions.
Changing the conversion window without telling anyone is one of the easiest ways to make a funnel look better or worse than it is. Document the window alongside every funnel report.
How to Build a Funnel Analysis
A reliable funnel comes from a deliberate process, not from dragging events into a tool and hoping. The following sequence works whether you use a dedicated product analytics platform or a spreadsheet built on event exports.
- Define the goal. State the single outcome the funnel measures — a purchase, an activation, a completed form. One funnel, one goal.
- List the required steps. Identify the events a user must logically pass through. Resist adding optional steps; they inflate drop-off that isn’t real friction.
- Set the entry point. Decide who enters the funnel and when. This is your denominator and shapes every percentage below it.
- Choose the conversion window. Match it to real behavior. Impulse purchases need a short window; considered B2B decisions need weeks.
- Confirm your event tracking is clean. A funnel is only as accurate as the events feeding it. Verify each step’s event fires correctly before trusting the funnel. See our guide to campaign tracking for keeping source data consistent.
- Compute step and overall conversion. Calculate both the step-to-step rates and the end-to-end rate.
- Rank steps by lost users. Sort by absolute drop-off, not just percentage, so you focus on steps that lose the most actual people.
Reading and Interpreting Funnel Results
The instinct when reading a funnel is to fixate on the lowest step conversion. That is a reasonable starting point, but two refinements make interpretation far more useful.
First, weigh percentage drop-off against absolute drop-off. A step that loses 50% of 200 users costs you 100 people; a step that loses 20% of 8,000 users costs you 1,600. The second is the bigger problem despite the better rate. Ranking by absolute lost users keeps you honest.
Second, distinguish expected narrowing from genuine friction. Some drop-off is natural — not everyone who views a pricing page intends to buy. The goal is not zero drop-off; it is identifying drop-off that exceeds what the step should reasonably produce. Comparing the same step across segments and across time is how you tell normal narrowing from a real problem.
A funnel tells you precisely where users leave. It never tells you why. Pair the funnel with session recordings, surveys, or support tickets to explain the drop-off you’ve localized.
Segmenting Your Funnel
The single most valuable thing you can do to a funnel is segment it. The aggregate funnel averages together populations that behave nothing alike, and the average hides the insight.
| Segment | Landing → Product | Product → Cart | Overall |
|---|---|---|---|
| Desktop | 68% | 46% | 19% |
| Mobile | 58% | 31% | 11% |
| Paid search | 71% | 38% | 16% |
| Organic | 61% | 49% | 20% |
The aggregate funnel might show a 40% product-to-cart rate, suggesting a single problem. Segmented, the data tells a sharper story: mobile users convert far worse at that step (31% vs 46%), pointing to a mobile experience issue rather than a universal one. The fix is now specific and testable.
Useful segmentation dimensions include traffic source, device, new versus returning users, geography, and entry page. Start with the dimension most likely to differ in behavior for your business, and look for steps where one segment underperforms the others by a wide margin.
Open vs. Closed Funnels
An important and often-missed distinction is whether your funnel is closed or open. In a closed funnel, only users who completed the previous step are eligible to be counted at the next — the population strictly narrows. In an open funnel, a user can enter at any step, so someone might appear at checkout without ever being counted at the product page.
Closed funnels measure a disciplined sequence and are right for journeys with a genuine required order, such as onboarding. Open funnels better reflect messy real-world behavior, where users arrive from bookmarks, ads, and search directly into the middle of a flow. Using the wrong mode is a frequent source of confusion: an open funnel can show a step “converting” above 100% relative to the prior step, which is impossible in a closed one and often mistaken for a data error.
Common Mistakes
A 30% step at the top of a high-traffic funnel usually costs more users than a 60% step near the bottom. Rank by absolute lost users.
Reporting a funnel without stating its time window invites apples-to-oranges comparisons. Two teams measuring “the same funnel” with different windows will never agree.
If a step’s event sometimes fails to fire, that step will show false drop-off. Validate event accuracy before drawing conclusions about user behavior.
The unsegmented funnel is a starting point, not an answer. The insight almost always lives in a segment that behaves differently from the average.
Funnel Analysis Tools
Funnel analysis is available in most modern analytics platforms, though the depth and flexibility vary considerably. The right choice depends on whether you need product-grade behavioral funnels or simpler web conversion paths.
| Tool Category | Best For | Funnel Strengths |
|---|---|---|
| Web analytics platforms | Site conversion paths | Page-based funnels, traffic-source segmentation |
| Product analytics platforms | In-app behavior | Event-based funnels, open/closed modes, deep segmentation |
| BI tools on a warehouse | Custom, blended data | Full control over definitions, joins across sources |
| Spreadsheets on event exports | Small datasets, prototyping | Maximum transparency, no tool lock-in |
Whichever you choose, the discipline matters more than the tool. A spreadsheet funnel with clean events and explicit definitions beats a sophisticated platform fed by ambiguous tracking. For a broader view of platform options, see our web analytics tools guide and the companion analytics dashboard guide for presenting results.
Frequently Asked Questions
What is the difference between a funnel and a conversion rate?
A conversion rate is a single number — the percentage of users who reach a goal. A funnel breaks that journey into ordered steps and shows the conversion between each one, so you can see exactly where users are lost rather than only whether they were.
How many steps should a funnel have?
Use the fewest steps that capture the real journey, typically three to six. Each extra step adds drop-off and noise. If a step is optional, leave it out — including it makes the funnel look leakier than it truly is.
What is a good funnel conversion rate?
There is no universal benchmark, because rates depend heavily on the funnel’s length, traffic quality, and goal. The more useful question is whether each step’s conversion is improving over time and how it compares across your own segments, rather than against an external number.
What is the difference between an open and a closed funnel?
In a closed funnel, users must complete each step in order to be counted at the next, so the population only narrows. In an open funnel, users can enter at any step. Closed funnels suit strict sequences; open funnels reflect real-world behavior where people arrive mid-journey.
How does funnel analysis relate to cohort analysis?
Funnel analysis shows where users drop off in a sequence; cohort analysis shows how behavior changes over time across groups. Combining them — running the same funnel for cohorts from different periods — reveals whether your funnel is getting better or worse, not just how it performs right now.
Why does my funnel show more users at a later step than an earlier one?
This usually means you are looking at an open funnel, where users can enter directly at a later step without passing through earlier ones. It can also indicate a tracking issue where an earlier event sometimes fails to fire. Check the funnel mode first, then validate the events.
Sources & Further Reading
- Marketing Analytics: The Complete Guide — where funnel analysis fits in the broader measurement stack
- Cohort Analysis — adding a time dimension to funnel data
- How to Build an Analytics Dashboard — presenting funnel results clearly
- Web Analytics Tools Guide — platforms that support funnel reporting