Subscription commerce attribution methods explained for recurring revenue

What to take away

Subscription commerce attribution methods are the rules that decide which touchpoint gets credit for a signup, a renewal or a cancellation.

Why recurring billing changes the maths

In a one-off shop, attribution stops at the order.

A subscription keeps producing events after the sale: renewals, upgrades, pauses and cancellations.

Credit for revenue therefore depends on where you draw the line.

Draw it at signup and paid search looks heroic; draw it at month six and the same channel looks expensive.

Churn timing matters more than churn volume. A cohort that leaves in month two costs you different money from one that leaves in month eleven.

The subscription commerce measurement and reporting guide sets out how these choices feed a board pack.

The main attribution methods compared

Last-touch attribution gives all credit to the final click before signup.

Set-up effort is low: one last-click rule in your analytics or subscription platform, plus a monthly check for duplicate conversion events.

It ignores the podcast, the review site and the friend who recommended you.

Last-touch attribution: cheap to run, easy to misread

Last-touch works when your funnel is short and your paid channels carry clear labels.

It misleads when a long consideration period sits between the advert and the signup.

A shopper who reads three reviews before subscribing is credited to whichever link she clicked last.

First-touch attribution: credit the earliest known touch

First-touch attribution does the opposite, crediting the earliest known interaction.

It gives the whole signup to the first advert, article or podcast that introduced the brand, and ignores every later prompt.

Set-up effort is low too, but most platforms record first touch only if you configure a persistent identifier; otherwise you export raw events and rebuild the field.

Neither model touches renewals, because both stop at the first payment.

Renewal-weighted attribution: credit the second payment

Renewal-weighted models split credit between acquisition and retention.

The first payment goes to the acquisition channel; later renewals go to whoever kept the customer.

This matches how subscription profit actually arrives, in small instalments over many months.

Worked example: one cohort, three scores

An illustrative cohort of 100 subscribers pays £10 a month and stays 12 months, so total revenue is £12,000. 60 subscribers had paid search as both first and last click. 25 first clicked a podcast and last clicked paid search. 15 arrived direct.

Renewal-weighted splits the £1,000 first payments by last touch and gives the £11,000 of renewals to the retention channel.

The same cohort produces three different stories. The method you pick decides which story reaches the board.

Self-reported attribution fills the gaps

Add a required question at checkout asking how the subscriber heard about you.

You trade precision for coverage, because answers rely on memory.

It still catches podcasts and word of mouth that pixels never see.

Ask at signup rather than by email weeks later, when recall fades. Keep the answer list short and include a free-text option you review each month.

Compare the reported channel mix against your subscription commerce key metrics in England: signup-to-renewal rate, average revenue per subscriber and cancellation timing by channel, before you trust either source.

A decision table for choosing a method

Match the method to the shape of your funnel and your renewal rate.

Marketing rules sit underneath every method

Consent decides which identifiers you may store, join and act on.

The ICO step-by-step guide to direct marketing for small businesses sets out the records a small operator needs.

The Data Protection Act 2018 underpins the lawful basis you rely on.

Attribution reports built on data you cannot lawfully hold will not survive an audit.

What physical subscription boxes add

For boxes containing toys, cosmetics or electricals, the Office for Product Safety and Standards safety regime matters, but the attribution distortion starts with returns and recalls.

Returns and recalls then distort retention figures you believed were clean.

A recall removes customers from the base without a cancellation event, so your active subscriber count and your attribution base drift apart.

Most platforms book a refund as a cancellation, which drags renewal-weighted results down.

How to check a method is telling the truth

Pick one month and rebuild the numbers by hand from raw events.

Then compare against the platform dashboard, channel by channel.

Start with the channel that carries the most spend, because errors there cost the most.

Where a consent gap removes an identifier, expect undercounts. Show the gap in the same report so nobody reads the shortfall as falling demand.

Large discrepancies usually point to a broken naming rule or a consent gap.

Common questions

Can marketing and finance use different methods?

They can, provided the differences are written down. Finance needs renewal-weighted revenue while marketing optimises acquisition, so reconcile both in one monthly view.

How often should the rules be reviewed?

Review when you add a channel or change the billing cycle, and otherwise once or twice a year. Constant changes make period-on-period comparisons meaningless.