Attribution for Subscriptions: First Order Credit Is Not Renewal Credit
A channel that wins the first subscription order often gets credited for renewals it never touched. Separating acquisition attribution from renewal revenue, and what changes when you do.
A paid social campaign drives a shopper to a first subscription order at a discount. Four months later that subscription has renewed three times, and in the dashboard those renewals sit in the same revenue column, still carrying the first campaign's name. The campaign now looks like the best thing the business ever ran, and its budget case rests mostly on revenue it had no part in producing after the first click.
Subscriptions break the usual attribution habit because the first transaction and the value are separated in time. The channel genuinely caused the acquisition. It did not cause the third renewal in any meaningful sense. The product, the delivery, and the lack of a reason to cancel caused that, and crediting the ad for it distorts every decision built on the report.
Split the report at the first renewal
The simplest correction is structural. Report first order revenue and renewal revenue in separate columns, with separate attribution. First orders keep the acquisition model the team already uses. Renewals attribute to the subscription itself, or to the retention activity that actually intervened, an email, an offer, a product change, and not to a campaign from months earlier by default.
Run the split for one quarter before changing any budgets, because the first effect is presentational and it is large. Channels that recruit subscribers on heavy first order discounts often look dramatically worse once renewals stop flattering them, while channels that recruit at full price with a slower start look better. Both impressions are closer to the truth than the blended figure, which flattered whoever discounted hardest.
Cohort the subscribers, not just the orders
The next question is whether the subscribers a channel recruits stay. Group subscribers by acquisition channel and month, then track how many are still active at renewal two, four, and six. A channel whose subscribers cancel before the second renewal at twice the rate of another channel is not a cheaper acquisition channel. It is a more expensive one whose cost shows up later.
This is where first order discounting gets judged properly. A deep introductory offer recruits people who wanted the offer, some of whom never valued the product at full price. If their renewal rate is poor, the correct response may be a smaller introductory discount and a better second order experience, which is a conclusion a blended revenue report would never produce because the blended report was busy crediting the campaign for renewals that, in this cohort, never happened.
Keep a place for acquisition credit where it is due
None of this says acquisition channels deserve less credit. It says they deserve credit for the thing they did, winning a subscriber whose value is then determined by retention, and that the value assigned to that win should be an expected value based on the cohort's actual renewal behaviour, not a running total of whatever those subscribers happen to spend while the attribution window stays open.
Set that expected value per channel from observed cohort data and revisit it quarterly. Forecasts built on it will be wrong in detail and right in structure, which is a better place to make budget decisions from than a dashboard that cannot distinguish an order a campaign caused from an order that merely happened afterwards.
Related reading: Attribution Playbook, Answers, Glossary.
Frequently asked questions
Why should renewal revenue not go to the acquiring channel?
Because the renewal is caused mainly by the product, fulfilment, and the absence of a reason to cancel, not by an advertisement from months earlier. Crediting it to acquisition inflates that channel's return and biases budget toward heavy first order discounting.
What should an attribution report split for subscriptions?
First order revenue, attributed to acquisition, and renewal revenue, attributed to the subscription and any retention activity that intervened. Keeping them in one column is the error that drives the distortion.
How do you judge a channel that recruits subscribers who cancel early?
By cohort retention. Group subscribers by channel and start month and compare how many remain active at successive renewals. A cheap first order with poor retention is an expensive acquisition once the full pattern is visible.
How should acquisition channels be valued instead?
By expected subscriber value drawn from their own cohort's observed renewal behaviour, updated quarterly. That credits the channel for winning a subscriber without crediting it for every later transaction automatically.