A Subscription Business vs. Churn

What happened

A software product reached about $1 million in annual recurring revenue while losing close to 5% of customers every month, so marketing was filling a leaking bucket. Cohort analysis found that most of the leaving happened in the first thirty days, before customers ever reached the thing the product was good at. The team rebuilt onboarding around that first moment of value and introduced annual plans, and churn roughly halved. The same marketing spend then compounded instead of evaporating.

The case, start to finish

The topline was healthy and the cohorts were dying, and only one of those two numbers was on anybody's dashboard.

A leak that growth was hiding

An anonymized composite: a small business-to-business software company at roughly $1M in annual recurring revenue. It is built on the retention economics documented by Reichheld and Bain, whose central finding is that small improvements in retention produce outsized improvements in profit.

By the end of year one the product had real customers and a real revenue number, and roughly 5% of those customers were leaving every month. Monthly churn of 5% does not feel like an emergency. It feels like a rounding error. It is reported as a small percentage sitting next to a rising revenue figure, which is exactly why it survives so long unaddressed.

What it actually meant was that the company was refilling a leaking bucket at full price. New customers arrived at customer acquisition cost and roughly the same volume drained out the bottom. The topline still rose, because acquisition was outpacing the leak, and the topline was what leadership looked at.

Where the customers were actually leaving

The unlock was measurement, and specifically the right kind of it. A blended monthly churn rate is an average across every customer the business has ever had, which mixes a loyal three-year account with somebody who signed up on Tuesday. Grouping customers by the month they joined and following each group separately tells a different story. Here the story was sharp: most of the churn happened in the first 30 days.

That single fact reframes everything. Customers were not wearing out on the product over time. They were leaving before they had experienced what the product was for. The failure lived in the distance between signing up and the first moment the thing was obviously worth paying for. That is not a marketing problem and not a pricing problem. It was a product problem, sitting in week one.

The fix, and the one that felt easier

The easier response had already been tried, and it does not work. Offering a discount to somebody who is canceling delays the cancellation by about a month and teaches every customer that threatening to leave is a negotiating position. Worse, it shrinks the contribution margin of exactly the customers who were going to stay anyway. Price was never why people left.

The actual fix cost a quarter of roadmap. Onboarding was rebuilt around the measured first-value moment rather than around a feature tour. Annual plans were introduced, which makes nobody happier but does reduce the number of times a year a customer is invited to reconsider. And outreach was triggered by usage dropping rather than by a cancellation arriving, which is the difference between reaching somebody who is still deciding and somebody who has decided.

Churn roughly halved, from about 5% a month to about 2.5%. The consequence is larger than it sounds. At 5% the average customer lasts around 20 months. At 2.5%, around 40. The same acquisition spend, aimed at the same market, now buys twice the lifetime revenue.

Why retention compounds and acquisition does not

The asymmetry is the whole point. Acquisition buys one customer once. A point of retention improves every customer the business has already paid for and every customer it will ever pay for. In a business with high gross margins, where the cost of serving an existing customer barely moves, that improvement lands close to straight through to profit.

The transferable discipline is an order of operations, and it applies to a gym, a bookkeeping practice or a lawn-care route as readily as to software. Before spending more on winning customers, find out how long the ones you have already bought are staying, and where in their life with you they leave. If they are leaving early, more marketing simply buys more people who will leave early.

Timeline

• Year 1 Product hits ~$1M ARR but monthly churn sits near 5%, so the bucket leaks almost as fast as marketing fills it.
• Year 2 Q1 Cohort analysis shows most churn happens in the first 30 days, before customers reach the product's core value.
• Year 2 Onboarding is rebuilt around the first-value moment; annual plans introduced; churn roughly halves.
• Year 3 Same marketing spend now compounds: revenue growth accelerates because the bucket holds what it catches.

You're in the owner's chair

Your SaaS is at ~$1M ARR, but 5% of customers leave every month. The board wants faster growth. Where does next quarter’s effort go?

• Cohort-analyze churn, rebuild onboarding around first value
• Cut prices to keep people from leaving
• Pour the whole marketing budget into acquisition and outgrow the churn

Same acquisition spend, two futures

• Average customer lifetime at 5% monthly churn: 20 months
• Average customer lifetime at 2.5% monthly churn: 40 months

Halving churn doubles how long every already-purchased customer stays. It is the highest-leverage number in the whole model, and the one growth spend hides.

Business model

Recurring software subscriptions: pay to win a customer once, earn from them monthly. The entire model's economics hinge on how long the average customer stays.

Revenue model

Monthly and annual plans. At 5% monthly churn the average customer lasts ~20 months; at 2.5%, ~40. The same acquisition spend buys twice the lifetime revenue.

Cost structure

Mostly fixed (engineering, hosting scales gently), plus acquisition cost per customer. High gross margins mean retention improvements drop almost straight to lifetime profit.

Strategic challenge

Growth masked the leak: new-customer wins made topline look healthy while cohort retention curves showed each month's customers evaporating. Growth was refilling a leaking bucket at full acquisition price.

Key decision

Stop buying more water and fix the bucket: divert a quarter's roadmap from features to onboarding, aimed at the measured first-value moment; add annual plans to structurally lower churn opportunities.

What worked

Cohort measurement (finding that churn concentrated in month one), onboarding to first value, annual billing, and win-back emails at the moment usage dropped: each small retention gain compounded.

What failed

Discounting to retain cancellations delayed churn by a month and taught customers to threaten leaving. Retention improved only when the product reached value faster, not when it got cheaper.

Risk factors

Churn concentration in early cohorts; acquisition costs rising while lifetime value is uncertain; discount habits eroding price integrity; annual-plan refund exposure.

Lesson summary

In subscriptions, retention is the growth engine: churn compounds against you, retention compounds for you. Measure cohorts, find the first-value moment, and fix the bucket before buying more water.

Key data

• ~5% → ~2.5% Monthly churn, before → after
• ~20 → ~40 months Implied avg. customer lifetime

Sources & basis

The business in this story is a stand-in, not a company you can look up. This case is an illustrative composite: the operator, the people and most of the dollar figures represent a pattern rather than reporting one firm's history. What the list below cites is the other half, the documented industry data and public reporting the composite was assembled from, including any real company whose published figures the case draws on by name. The mechanism and the arithmetic are real even where the business is not.

1. Reichheld/Bain retention economics (HBR, 1990)
2. Composite pattern: see Subscription Economics and Customer Lifetime Value lessons