Meta claims a purchase because the buyer saw an ad six days ago. Google claims the same purchase because they clicked a brand search on the way to the checkout. Klaviyo claims it too, because they opened the abandoned checkout email that morning. None of the three is lying. Each is answering the question it was built to answer, which is whether it touched the sale, not whether it caused it.
The arithmetic that gives it away
Add the three attributed revenue figures together and set them next to the store's own total for the same dates. If the sum is bigger, and it usually is by some distance, then at least one purchase has been counted more than once and you have no way of knowing which. Any decision built on the sum is built on a number that does not exist.
There is exactly one revenue figure in the business that counts every order once. It is not in any of the marketing tools.
The denominator
The store's own total. It is the only one with a reason to be right, because it is the number the accountant and the tax authority both work from. So that is what email gets measured against: flow revenue as a share of store revenue, campaign revenue as a share of store revenue, same dates every week, no window games. The number goes down when the store has a big month on the back of something else, and that is correct, because that is what happened.
The split that matters more than the total
On Curated Chrome, 72.7 per cent of email revenue came from flows rather than campaigns. That ratio is worth more attention than the headline figure, because the two halves cost completely different amounts to produce.
A campaign is work every single time. Somebody writes it, builds it, checks the segment and sends it, and next week the meter resets to zero. A flow is work once. It is built, it is tested, and then it earns quietly for years while nobody touches it. A brand whose email revenue is mostly campaigns is renting that revenue by the week. A brand whose email revenue is mostly flows owns it.
The order we build them in, and why
- Abandoned checkout. The highest intent audience anybody will ever have. If only one thing exists, this is the one.
- Browse abandonment. More volume, lower intent, and it usually earns more in total than the checkout flow does precisely because more people reach it.
- Welcome. The only email somebody actually wants. It sets the price expectation and the tone for everything that follows, so it is worth writing properly rather than fast.
- Post purchase. Fires when goodwill is at its highest point and nobody is asking for anything. Cheapest second order in the business.
- Winback. Built last because it needs enough purchase history to know what lapsed actually means for this catalogue. Guessing at ninety days is how you annoy people who were going to come back anyway.
- Back in stock. Small, but it is the only flow where the customer asked to be interrupted.
What ends up in the report
- Flow revenue and campaign revenue, each as a share of store revenue for the same dates.
- The flow to campaign split, so you can see the owned half growing.
- Revenue per recipient rather than open rate, because opens stopped being a real measurement the day mail clients started prefetching images.
- List health: unsubscribes, spam complaints and the sending domain reputation, in the weeks where they are fine as well as the weeks where they are not.
The flows we build and what they cost is on the Klaviyo page.