Calculating churn rate and what it tells you

Churn rate is the share of your members who leave over a period of time. It sounds simple, and the basic arithmetic is. But the choices you make about who counts, and when, can make the same membership look healthy or alarming.

Here is how to calculate churn properly for a membership site, the decisions that change the number, and how to read what it is telling you about your business.

The basic formula

Monthly churn rate is the number of members who left during the month, divided by the number of members you had at the start of the month.

Churn rate = members lost during the month ÷ members at the start of the month

Say a language-learning site starts the month with 500 paying members and 40 of them cancel or lapse by the end. The churn rate is 40 ÷ 500 = 8%. The new members who joined during the month do not change this calculation; you count them separately.

Why divide by the starting number rather than the ending one? Because the members at the start are the group who had the chance to leave. Dividing by the ending count, which includes new joiners, makes churn look smaller whenever you are growing.

Decisions to make before you calculate

Most of the confusion around churn comes from small definitional choices. There is no single right answer to each, but you must pick one and stick to it, or your month-to-month comparisons mean nothing.

• When does a member count as lost? On the day they click cancel, or on the day their paid period ends? Most owners use the end of the paid period, because that is when revenue actually stops.
• What about failed payments? A member whose card declines has not chosen to leave, but if the payment is never recovered they are gone all the same. Count them, but record them separately as involuntary churn so you can see which problem you have.
• What about members who join and leave in the same month? They were not in your starting count, so the simple formula ignores them. If many new members leave within weeks, track that separately as early churn rather than letting it vanish.
• What about paused memberships? A paused member is not paying but has not left. Leave them out of both the starting count and the losses while paused, and note how many there are.
• What about people who leave and come back? Count the departure as churn in the month it happened, and the return as a reactivated member later.

Member churn and revenue churn

Counting heads is one view. Counting money is another, and the two can tell different stories.

Revenue churn is the monthly recurring revenue you lost from cancellations, divided by your recurring revenue at the start of the month. Suppose the language site has two levels: 400 members at $20 a month and 100 at $50. Starting monthly revenue is $8,000 + $5,000 = $13,000. If 30 of the $20 members and 10 of the $50 members leave, member churn is 40 ÷ 500 = 8%, but revenue churn is ($600 + $500) ÷ $13,000, which is about 8.5%.

Revenue churn higher than member churn means your more valuable members are leaving at a higher rate, which deserves attention. If it is lower, you are mostly losing members at your cheaper level. Downgrades also reduce revenue without anyone leaving, so it is worth tracking those alongside revenue churn.

Annual members need their own measure

Monthly churn gets awkward when some members pay yearly. An annual member cannot leave in eleven months out of twelve, so mixing them into a monthly calculation drags the rate down artificially and makes it jump around.

The cleaner approach is to measure annual members by renewal rate: of the members whose annual term ended this month, how many renewed? If 25 annual memberships came up for renewal and 20 renewed, the renewal rate is 80% and churn for that group is 20%. Report monthly churn for monthly members and renewal rate for annual members side by side, rather than blending them into one figure.

What your churn rate is telling you

A single month's churn is a noisy number, especially on a small site, where two or three extra cancellations can swing the rate noticeably. Look at the trend over several months, or use a three-month average, before drawing conclusions.

Once you have a steady trend, churn tells you several useful things:

• How long members typically stay. A rough rule is that one divided by monthly churn gives the average membership length. At 8% monthly churn, that is about a year. At 4%, about 25 months. This is why small improvements in churn matter so much.
• Where your growth ceiling is. If you lose 8% of members a month and gain 40 new members a month, you will level off where 8% of your members equals 40, which is 500 members. To grow beyond that, you need more joiners or lower churn.
• Whether a change worked. A new onboarding sequence, a price change or a new content series should show up in churn a month or two later.

Churn gets far more useful when you split it. Compare churn for members in their first three months with churn for longer-standing members; in our experience early churn is often much higher, which points to onboarding rather than content. Compare levels, pricing plans and the channels members came from. A single overall rate can hide one group leaving in droves while everyone else is content. When you want to know why people are leaving, your cancellation page is the best place to ask, and the answers turn churn from a number into a to-do list.

Mistakes that distort the number

• Changing a definition partway through, then comparing figures from before and after the change.
• Counting free or trial accounts as members, which inflates both the base and the losses.
• Ignoring failed payments until they are months old, so churn arrives in lumps.
• Comparing one month with one month, instead of looking at a trend.

Your churn routine

1. Write down your definitions for a lost member, failed payments, pauses and same-month joiners.
2. On the same day each month, record starting members, voluntary cancellations and involuntary losses.
3. Calculate monthly churn for monthly members and renewal rate for annual members.
4. Once a quarter, split churn by how long members had been with you and by level.
5. Pick the group with the highest churn and focus there first. For new members, that usually means looking hard at what happens in their first thirty days.