A conversion funnel shows how many people move from each step to the next, from first contact to the valuable action, and where most of them are lost. Cohort retention follows the same group over time to see how many stay active. Together, they show whether the problem lies in bringing people to the value or in keeping them afterwards.
How to use
- Write the funnel steps in order, with the people who reached each one over the same period.
- See the conversion from the previous step and from the start, with the biggest drop highlighted.
- Optionally, fill in a cohort retention: period zero and the following ones, such as day 1, day 7 and day 30.
- Read the curve: if it flattens, a group keeps using the product; if it still falls, there is no stable plateau yet.
- Export the table as CSV or PNG and turn the biggest drop into an improvement hypothesis.
What each field asks for
| Field | What to enter |
|---|---|
| Step | E.g. started sign-up |
| People | How many reached this step. |
| Period | E.g. Day 7 or Week 2 |
| Active | How many people from the cohort were active in the period. |
Example
A fictional project management app gets 12,000 visits to its pricing page; 3,600 start sign-up, 2,700 confirm their email, 1,350 create their first project and 270 subscribe to a plan. Overall conversion is 2.25%, and the biggest drop is at subscription: only 20% of those who create a project subscribe. In the cohort of 1,350 people, 60% return on day 1, 36% on day 7 and 33.3% on day 30: the curve flattens, with a 7.4% loss between day 7 and day 30.
When to use
- To decide which funnel step to focus the next experiment on.
- To compare the funnel before and after a change to sign-up or onboarding.
- To see whether a new cohort retains better than the previous one.
- To bring a metrics review a number that shows where the product loses people.
Frequently asked questions
How do I build a conversion funnel?
Choose three to six steps in sequence, from first contact to the valuable action, and count the people who reached each one over the same period. Every step should count the same unit, such as people, without mixing people and sessions.
What is the biggest drop in the funnel?
It is the step with the lowest conversion from the previous one. It is usually the best place to investigate first, but compare the volume lost too: a smaller drop at a step with many people may be worth more.
What is a cohort?
It is a group of people who started in the same period, such as everyone who signed up in a given week. Following the cohort separates the effect of time in use from the effect of new people coming in.
How do I read the retention curve?
Every curve falls at first. What matters is whether it flattens: a stable plateau shows a group that has found recurring value. If the curve keeps falling period after period, the product is not yet holding on to any group.
What is the difference between retention and churn?
Retention is the share of the cohort that stays active. Churn is the share that leaves in a period. The calculator shows both: retention over the initial cohort and the loss over those remaining in the previous period.
Do the periods need equal intervals?
No, but the reading is fairer when they do. A loss between day 7 and day 30 covers more time than one between day 1 and day 7. Use regular days or weeks when you want to compare the loss from one period to another.