What Klaviyo's attributed revenue actually measures

Every retention agency gets this question eventually, usually from a client who just opened their Shopify dashboard next to your report: "why doesn't this number match?" The honest answer isn't that one of the two is wrong. It's that Klaviyo's attributed revenue was never trying to answer the same question Shopify's total revenue answers.

The number is a model, not a ledger

Shopify's revenue report is a ledger: it counts orders that happened, full stop. Klaviyo's attributed revenue is a model: it counts orders that happened and that a specific email or SMS message can be credited with influencing, according to rules Klaviyo defines. A flow's "attributed revenue" is the sum of the value of every order tied to a conversion event that falls inside that flow's attribution window. Change the window, change the number — nothing about the store changed.

How the attribution window works

For accounts created after October 9, 2024, Klaviyo's default attribution window is 5 days for both opened and clicked email, and 5 days for clicked SMS (older accounts may still be on a shorter default). Push notifications default to 1 day. These are configurable — up to 60 days for email and up to 72 hours for SMS in the standard settings, with an extended option up to 90 days and 720 hours respectively — so two Klaviyo accounts can report different attributed revenue for an identical order history simply because they have different windows set. Before comparing two accounts, or two periods on the same account, check that the window didn't change in between (Klaviyo's own attribution documentation covers how to audit and change it).

Why Klaviyo and Shopify never agree

Even with the window held constant, a handful of structural differences guarantee a mismatch:

• Multi-touch overlap. If a customer clicked an email and later saw a retargeting ad before buying, Klaviyo may credit the flow while an ad platform credits itself — both can be "right" under their own model and the two numbers won't sum to the order value.
• Returns and cancellations. Klaviyo's attributed revenue is typically captured at the point of the conversion event and doesn't always net out later refunds the same way Shopify's reporting does.
• Currency and timezone. An order placed near midnight can land in a different reporting day depending on which platform's timezone setting is in effect.
• Test and duplicate orders. Shopify includes every order unless explicitly excluded; Klaviyo's numbers depend on which events actually fired against the customer's profile.

The three claims, ranked by how well they hold up

Take one real figure — say a flow attributed €12,480 over 90 days — and see how it survives three different phrasings:

1. "We generated €12,480 for you this quarter." A causal claim. It says the money wouldn't exist without the agency's work — a counterfactual nobody measured.
2. "This flow made €12,480." Ambiguous about source and period. A client checking Shopify next to it will find a different figure and assume an error.
3. "Klaviyo attributed €12,480 in conversions to this flow over the last 90 days." A specific, sourced, checkable statement. It survives being forwarded to whoever runs the client's own analytics.

Only the third version is something you can stand behind under a follow-up question.

What to put in a client-facing report

A defensible report names, at minimum: the conversion metric used (Placed Order, Ordered Product, or a custom metric — see why that choice matters), the exact date range, the attribution window in effect, and the platform the figure came from. It should also say, once, what the figure does not include — usually anything outside that window, and anything the model didn't attach to a message event at all.

Saying it neutrally

The phrasing that holds up in practice: "Klaviyo-attributed performance for this flow during [period]." It reads as more careful, not less impressive — and it's the sentence a client can't push back on, because it isn't claiming anything beyond what actually happened in the data. A number with a clear source is an argument. A number without one is just a claim waiting to be challenged.