Why customers stay, and what churn does to everything

A business adds 500 customers every month and loses five percent of the ones it has. It feels like growth. It is growth, for about two years.

Then it stops, permanently, at 10,000 customers, and no amount of extra effort on the acquisition side moves it. The leak and the tap have met.

Question six: why do they stay?

Every business has an answer, and the answers are not equally durable.

Because it keeps working and they never think about it. The strongest and quietest reason. It looks like nothing on a dashboard.

Because leaving is a project. Their data is in it, their team is trained on it, three other systems talk to it. This is a real reason and it has a shelf life, because resentment accumulates.

Because everyone else they deal with uses it too. Leaving means leaving the group, which is a much larger decision than leaving a product.

Because they forgot they were paying. Also real, also revenue, and the subject of an entire lesson two tracks from now.

Predict

A subscription business cuts monthly churn from 5 percent to 3 percent. Roughly what happens to the value of one customer?

Choose one answer

It rises by about 2 percent, matching the changeIt rises by about 40 percentIt rises by about two thirdsIt roughly doubles

Lock in answer

What a customer is worth, and when

Subscription lifetime value

≈ Hypothetical model, not a forecast

Turns a monthly churn rate into an expected lifetime and a lifetime value, discounted and not.

It assumes

• The same share of remaining customers leaves at the end of every month.
• Revenue per customer and gross margin both stay flat forever.
• The discounted figure charges a steady rate for every month of waiting.
• The business itself lasts as long as the customer lifetime shown.

It ignores

• Cohort effects, where churn falls as a cohort ages
• Upsell, downgrade, pauses and reactivation
• The cost of acquiring the customer in the first place
• Any risk that the business itself does not last the lifetime shown

One customer is worth 1,000 at the opening settings. Move churn alone until they are worth more than 2,000, then look at what happened to the share collected inside the sixty-month horizon.

Customers lost each month4%

Gross profit a month

40

Expected months as a customer

25

Lifetime value, undiscounted

1,000

Lifetime value, discounted

808

Collected by the horizon

914

Share of lifetime value collected by the horizon

0.914

• Collected, undiscounted
• Collected, discounted

◷ Not there yet — keep moving the controls.

Two things should have happened at once, and the second is the interesting one.

The lifetime value doubled. And the share of it you actually collect within five years fell — because at very low churn most of the value sits in months you have not reached yet. At two percent churn the customer is worth 2,000 and you will have seen about 70 percent of it after five years. At eight percent they are worth 500 and you have essentially all of it inside two.

Two churns, and the one people quote

Customer churn counts people. Revenue churn counts money. They come apart the moment customers are different sizes, and they come apart in opposite directions depending on who is leaving.

A business can lose eight percent of its customers and two percent of its revenue in the same month — the departures were all small accounts. The reverse also happens, and it is much worse news delivered in a much calmer voice.

There is also a third figure. Where existing customers grow — more seats, more usage, a higher tier — the revenue gained from the ones who stayed can exceed the revenue lost from the ones who left. Revenue then grows from the existing base with no new customers at all. It is rare, it is largely a property of what is being sold rather than of how well it is sold, and it is the single most valuable characteristic a subscription business can have.

Check

A business adds 800 new customers a month and has 4 percent monthly churn. It currently has 12,000 customers and the founders are planning for 40,000. What does the arithmetic say?

Choose one answer

It reaches 40,000 eventually, it just takes a whileIt settles at about 20,000 unless churn or intake changesIt depends on how long it keeps goingIt reaches 40,000 if it can keep growing intake at the same rate

Check answer

›Why cohort curves flatten, and why averages hide itOptional

Plot one month's intake and follow it. The first two or three months lose a large share — people who were never going to stay, who misunderstood what they bought, who were experimenting. After that the line bends and the remaining customers churn far more slowly.

A single average churn rate mashes those two populations together, and the average is dominated by whichever group is larger this month. A business growing fast has a lot of new customers, so its average churn looks terrible. The same business stops growing and its average churn improves, with nothing having got better.

Everything in this lesson uses a single constant rate, which is exactly this simplification. It is the right first approximation and the wrong thing to run a retention programme on.

Worth remembering

• Average customer lifetime is roughly one divided by the churn rate, so five percent monthly churn is about a twenty-month customer.
• A business adding a fixed number of customers a month settles at that number divided by its churn rate, and then stops growing.
• Because churn sits in the denominator, a small change in it produces a large change in the value of a customer.
• Very low churn pushes most of a customer's value into distant months, which is worth less to a business that has to fund the present.
• Customer churn and revenue churn are different numbers, and they can move in opposite directions in the same month.