How to run a subscription commerce ninety day plan without guesswork

What to take away

Weeks 1 to 3: baseline the numbers

Pull twelve months of order data for active, paused and cancelled subscribers. Split it by acquisition source so you can see which channel brought the customers who stayed.

A team paying £400 a month for email and £900 a month for paid search should know which produces a subscriber still paying in month nine. If nobody knows, that is the first gap the plan closes.

Set three baseline figures: average revenue per subscriber, monthly churn rate and support contacts per hundred subscribers. Note the date each was taken. A plan built on undated numbers cannot be audited later.

Define each cohort before you count it. A customer who joined in January and cancelled in March belongs to January, not March, and mixing the two hides whether retention is improving or getting worse.

Use the Subscription commerce: strategy and planning guide for 2027 to align the baseline with your wider objectives before you commit budget to any test.

Weeks 4 to 9: run controlled tests

Change one thing at a time: a trial length, a renewal reminder, a payment retry rule. Each test needs a start date, an end date and a success measure agreed in advance.

Keep tests small enough to close inside the block. A six-month pricing test cannot inform a ninety day decision, so treat it as a separate project.

Write the decision rule before the test opens. For example, if churn falls by two points against the control group, keep the change; if it holds flat, revert and log the result.

Retail demand moves with the wider market, so frame your assumptions against the Retail Sales Index reference tables from the Office for National Statistics rather than a single month of your own trading.

Where compliance bites

Subscription advertising sits inside the CAP Code. A promotion promising a free trial without making the rolling cost clear can be acted on, and the ASA's published sanctions guidance sets out what that means for the brand.

Data adds a second layer. If your platform, payment provider or support desk sits outside the UK, moving subscriber records is a restricted transfer. The ICO's international transfers guidance explains the safeguards required.

Decide in week one who signs off copy and who signs off data flows.

Weeks 10 to 12: decide and document

Score every test against the rubric below. Anything under 12 goes back for another cycle rather than into the live programme.

Maximum score is 20. Record the score, the owner and the review date in one place. The Subscription commerce strategy framework shows how these decisions feed the annual plan rather than sitting in a folder.

Governance that survives quarter end

One named owner, one scorecard, one decision meeting every fortnight. Keep the meeting to forty minutes and publish notes the same day.

Keep one page per test: hypothesis, dates, owner, result and the decision taken. That page is what a new finance lead or an auditor will ask for first.

If a test fails, write down why. Failed tests are cheap evidence, and they stop the same idea returning next quarter under a new name.

Common questions

How long should each test run?

Long enough to cover at least one full billing cycle for every subscriber in the sample. For monthly plans that usually means four to six weeks, leaving room to read the result before week ten.

What if we have no reliable baseline data?

Spend weeks 1 to 3 building one, even if it delays testing. A plan built on estimates will produce decisions nobody can defend.

Who should own the ninety day plan?

One person with authority over pricing, retention and the platform roadmap. Shared ownership across marketing and finance usually means the decision date slips.

Do we need legal sign-off before week one?

Yes for advertising copy and any transfer of subscriber data outside the UK. Both can be reviewed in parallel with the baseline work rather than blocking it.