Last updated:

|Edit this page

For every flow in your product, more people will start it than complete it successfully. Funnels enable you to visualize your flows and understand where the friction points are so that you can improve them.

You can learn the following from funnels:

How to create a funnel

  1. Click Product Analytics on the left sidebar
  2. Click the + New Insights tab
  3. Select the Funnel option in the dropdown

Adding steps

Select the steps to include in your funnel, ideally starting with the first event or action a user will trigger in the flow. Next, add intermediate events users need to go through to be successful and finish with the event you consider to be the success for this flow. It's best to start with the simplest flow and avoid using optional steps to ensure you don't filter out or skew results.

Funnel steps

Filtering steps

To refine your steps, you can filter out events using event, person, or group properties, autocapture elements, or HogQL. To do this, click the filter icon next to the step, choose your property, and fill in the details. You can also set a global filter in the Filters section in the bottom.

Global filters

Note: Using person properties in funnels requires capturing person profiles.

Excluding steps

You can also exclude people who completed certain events between two specific steps. To do this, add the step in the Exclusion steps section.

These people are completely excluded from the entire funnel.

Conversion rate calcuation

There are two options for showing conversion rates in a funnel:

  1. Overall conversion – for each step, this shows the conversion relative to the first step.
  2. Relative to the previous step – for each step, this shows the conversion relative to the previous step.

Overall conversion is helpful for understanding the entire funnel, whereas relative conversion shows you which steps have biggest opportunity for improvement.

Overall vs relative conversion


To understand how different types of user interact with your funnel, it's helpful to breakdown results. You can breakdown steps by event and person properties. To do this, click on the + Add breakdown button and select the property you want to break down by.

The example below shows a breakdown by operating system:

Breakdown funnel steps by Operating System
Funnel steps broken down by property

Attribution types

When breaking down funnels, it's possible that the same properties don't exist on every event. For example, if you want to break down by browser on a funnel that contains both of frontend and backend events.

In this case, you can choose from which step the properties should be selected from by modifying the attribution type. There are four modes to choose from:

1. First touchpoint

In this case, the first property value seen in any of the steps is chosen.

Consider this example sequence of events:

  1. $pageview with browser = Chrome
  2. user signed up with browser = Firefox

The first touchpoint means that Chrome will be the breakdown value.

In the first touchpoint mode, one person will always only have one value.

2. Last touchpoint

In this case, the last property value seen from all steps is chosen.

Consider the same example as above:

  1. $pageview with browser = Chrome
  2. user signed up with browser = Firefox

The last touchpoint means that Firefox will be the breakdown value.

In the last touchpoint mode, one person will always only have one value.

3. All steps

In this case, the property value must be seen in all steps to be considered in the funnel.

4. Specific step

In this case, only the property value seen at the selected step is chosen.

Consider the same example as above:

  1. $pageview with browser = Chrome
  2. user signed up with browser = Firefox

For the example above, when step two is chosen the value will be Firefox.

With the specific step mode, one person can have more than one property value. For example, if the user captured the same event with different browsers:

  1. $pageview with browser = Chrome
  2. $pageview with browser = Firefox

Then the person would have both Chrome and Firefox as their breakdown values.

Another common use for Specific step is to display only a single URL for pageview steps in the funnel columns. The default, First touchpoint uses the first property value seen from all steps, which will include the subdirectory URLs in the breakdown if they appeared in later steps.

For example, if you've

  • set your first query step as Pageview where Current URL = equals,
  • and /bar has subdirectories (e.g. etc.),
  • and Breakdown by is set to Current URL

but you don't want pages under the subdirectories of to appear in the funnel chart.

In this case, set the Attribution type to Specific step / Step 1 to see only in the pageview column of the chart.

For an unordered funnel, the specific step is meaningless, as the first step can be the last step, and vice versa. So, it's referred to as "any step" in the UI. It has the same semantics as a specific step in an ordered funnel, but it looks at property values from all steps.

Graph type

The graph type dropdown let's you choose between conversion steps, time to convert, and historical trends.

Conversion over time using Trends

Each type serves a unique purpose:

  • Conversion steps – shows you where users are dropping off in your funnel and number of people converting between steps.

  • Time to convert – shows you steps with the highest friction. Steps with a long time to convert are likely much harder than steps with a short time.

  • Historical trends – shows you how your conversion rate has changed over time. Useful for understanding how changes, fixes, and new features have affected your funnel success rate.

Tips for analzying funnels

Understand where people are getting stuck during your flow

The most common use case for funnels is understanding where people are getting stuck or dropping off in your flow.

Identifying drop-offs in a funnel

There a few ways to identify problem steps:

  • Absolute numbers show you where you're losing the most people.
  • Relative conversion rate show you which steps have the greatest impact on your overall conversion rate.

Usually the steps with the lowest relative conversion rate are where the largest opportunities are.

Identify unsuccessful users

Once you have a funnel drop-off you'd like to explore further, the first step is to find out why users are struggling. There are many ways to get this information, from talking to users directly to replaying their sessions.

Click on the chart or the linked column values below the funnel to view the individual people who COMPLETED or DROPPED that step of your funnel.

View users in a funnel

You can also save this list of users as a cohort for further analysis.

Explore user paths

It can be useful to explore the paths people take between steps in your funnel.

Click on the '...' next to any step in your funnel and, depending on which step you're selecting, you'll be presented with options to "show user paths":

  • leading to step
  • between previous step and this step
  • after step
  • after drop off
  • before drop off
View user paths in a funnel

Selecting any option creates a new insight showing the paths users took. This is useful for getting a complete picture of the real funnel created by user interactions, rather than the imagined perfect funnel engineers and designers have in their head. Spoiler: the real funnel and the perfect funnel are rarely the same!

Understand seasonality in your conversion rates

It's unlikely that your conversion rate will remain stable every day or every week. This is normal, but it's important to understand the external factors that can cause these fluctuations so you don't jump to conclusions too quickly when analyzing a big change or drop-off.

To look at the seasonality of your conversion rates, set the graph type to historical trends and set the time period to when you may expect seasonal variations.

A common seasonality to watch out for is weekends, particularly if your product is B2B. Another one is a yearly sale such as "Black Friday", which increases the number of people visiting your site and increases the expectation of discounts, making it difficult to compare against the rest of the year.

In this view, you can adjust the date ranges to observe key seasonal trends. You can then look at your funnel to determine if it's just a seasonal trend or whether something else has affected the experiece of your product (like a product change).

Correlation analysis

Where is this feature available?
Free / Open-source

Correlation Analysis automatically highlights significant factors that affect the conversion rate of users within the funnel.

Correlation analysis works well when you don't already have a hypothesis for what is affecting conversion through a funnel. This report will automatically highlighting significant events or properties that either negatively or positively impact conversion rate.

Further reading

Want to know more about what's possible with Funnels in PostHog? Try these tutorials:

Want more? Check our full list of PostHog tutorials.


Was this page useful?

Next article


Dashboards are the easiest way to track all your most important product and performance metrics. Unlike notebooks , which are ideal of adhoc analysis of specific issues, dashboards are designed for tracking common metrics over time. You can create a new dashboard from scratch, but we also offer numerous dashboard templates for tracking things like website metrics , product health metrics , and metrics for large language models . Creating a new dashboard Click on Dashboards in the left…

Read next article