Menu

CRO

Funnel Analysis

A practical CRO method for identifying where users fail to progress through key journeys.

How to use funnel analysis to spot drop-off points, understand conversion performance, and prioritise optimisation work.

4 min read

What it is

analysis is a quantitative UX and CRO method used to track how users move through a defined journey and where they .

It breaks a journey into steps, such as landing, browsing, adding to basket, and completing a purchase, and measures how many users progress between each stage.

Unlike , analysis focuses on at scale using analytics data. It shows what is happening, but not why.

The goal is to identify points, , and opportunities to improve .

Funnel analysis is useful for showing where users are falling out of a journey, but it needs other methods to explain why.

When to use it

Use this method when you need to understand across a journey.

It is most useful when:

You are analysing conversion rates or drop-off points
You need to identify where users are failing to progress
You want to prioritise optimisation efforts
You are working on checkout, sign-up, or key flows
You have analytics data available

It is less useful when:

You need to understand user motivation or reasoning
Data is limited or unreliable
Journeys are not clearly defined
Funnel analysis is often used alongside usability testing and user interviews to explain why issues occur.

Key takeaway

Use funnel analysis when you need a clear view of where performance breaks down across a journey.

How to run it

Set up properly

Be clear on the journey, its steps, and how each is tracked before you look at a number. A is a you impose on , so its steps have to match something real.

Check what the tracking actually fires on. Steps defined by page view rather than by completed action will report progress that never happened.

Run the method

analysis is the simplest of the behavioural methods and the easiest to misread. It tells you where people stop, never why.

  1. Define the steps explicitly, and only include steps everybody must pass through. Optional stages in a linear produce that is not loss.
  2. Measure how many reach each step, using unique users rather than events. Repeat visits inflate the top of a more than anywhere else.
  3. Calculate the rate between consecutive steps, not from the start. A stage losing half its looks minor if you only ever measure against the entry point.
  4. Identify the largest drops in absolute numbers as well as percentage. A 60% drop on a step reaching two hundred people matters less than 10% on one reaching fifty thousand.
  5. Segment by device, and user type. An average frequently hides one segment performing well and another failing completely.

Focus on where, then go elsewhere for why. analysis earns its keep as a targeting tool for qualitative work, not as an explanation.

Capture and make sense of it

The value comes from knowing where to look. Look across the to identify:

  • The points that cost the most in absolute terms
  • Trends across segments, especially where one diverges sharply
  • Differences between devices and
  • Change over time, which separates a new problem from a permanent one

Use this to direct deeper investigation. A that produces a redesign on its own has been over-interpreted.

What to look for

Focus on:

Drop-off points: where users fail to continue
Conversion rates: how effectively users move between steps
Segmentation: differences across user types, devices, or channels
Step-to-step rates: measured between stages, not against entry
Anomalies: unexpected spikes or drops in behaviour

Where it goes wrong

Most issues come from:

analysis shows what is happening, not why it is happening.

Steps defined by page view rather than completed action, so progress is counted that never happened
Funnel definitions that change between reports, making the trend meaningless
Reading a drop-off percentage without knowing what happens at that step
Averaging across segments, which hides one audience failing completely
Treating the funnel as an explanation when it only ever locates the problem

What you get from it

Done properly, this method gives you:

The specific steps where people leave, in absolute numbers as well as rates
A view of conversion that separates a real change from a seasonal one
A shortlist of places worth investigating with qualitative work
Evidence to argue for research rather than an opinion about where to look

Key takeaway

It helps you focus on the areas that matter most.

Get in touch

If this sounds like something you need, we can help you identify where users are dropping off and how to improve conversion.

No guesswork. No assumptions. Just clear insight you can act on.

FAQ

Common questions

A few practical answers to the questions that usually come up around this method.

What is funnel analysis in UX?

Funnel analysis is a method used to track how users move through a journey and where they drop off.

When should you use funnel analysis?

Use it when analysing conversion rates, user journeys, or performance across key flows.

What tools are used for funnel analysis?

Common tools include Google Analytics, Mixpanel, Amplitude, and other analytics platforms.

What is the difference between funnel analysis and user research?

Funnel analysis shows what users do at scale, while user research explains why they behave that way.

Can funnel analysis improve conversion rates?

Yes. It helps identify where users drop off so you can focus optimisation efforts effectively.

Quick take

If you want to understand where users drop off and why conversions are not happening, start with funnel analysis.

LET'S WORK TOGETHER

Ready to improve your product?

UX, research and product leadership for teams tackling complex digital services.

Previous feedback

I had a fantastic experience working with Andy. One of his most impressive achievements during our time at NHS HEE was masterminding a deeply complex information architecture for a new platform that brought together a large number of legacy websites.

Will Parkhouse

Senior Content Designer