Before you turn a Klaviyo revenue report into a new campaign brief, write down what the number includes. A report can be correct and still answer a different question from the one your creative team thinks it answers. Revenue associated with a campaign is not automatically evidence of what would have happened without it.
Use Migma for the next reviewed email once the analytics owner has accepted the finding. Keep the analysis and the creative decision connected by a short evidence note. That note should identify the metric, population, period, attribution definition, and the single change the team wants to test.
Ask a question that has a checkable answer
“Why is email revenue down?” leaves too much open. Which revenue, compared with which period, for which customers? A more useful starting question specifies the exact report and business context, then asks the analyst to explain the comparison.
Klaviyo’s October 13 data update describes using Composer to query data and inspect how a question was interpreted. That is a useful capability for discovery. It does not remove the need to verify definitions or make the answer a causal explanation.
Start with a plain-English metric card. Include the source event, selected dates, time zone, treatment of refunds, currency, audience restrictions, and attribution settings used by the actual report. If any item is unknown, ask the report owner rather than supplying a plausible default.
Put two different questions side by side
A business owner may ask how much the company sold. A lifecycle marketer may ask how much revenue the email platform associated with messages. Those questions can legitimately yield different figures. Before comparing numbers, inspect the current report’s definitions and filters.
For an original fictional example, suppose a store records 120 units of revenue during a period and a campaign report associates 35 units with a message. These values describe different selections. Neither number establishes that removing the email would have reduced sales by 35. The example uses abstract units to avoid suggesting a real campaign result.
Ask what happens to a purchase that follows several messages. Is it attributed to one message, counted as an assist, or visible in several overlapping reports? Record the answer before adding rows together. An attractive total can be misleading if the same order appears more than once.
Check the comparison before explaining the creative
Review whether both periods contain the same number of trading days, comparable offers, similar stock availability, and the same reporting configuration. A holiday, product shortage, or changed audience can affect the interpretation. Treat those as questions to investigate rather than invented explanations for a dip.
Use a simple decision record:
| Observation | What the reviewer must establish |
|---|---|
| Lower reported revenue | Same metric and attribution settings |
| Fewer purchases | Accepted purchase event and audience |
| More clicks but fewer purchases | Consistent link tracking and destination behavior |
| Higher order value | Comparable order population and currency |
This is a review framework, not a description of automatic Klaviyo or Migma diagnosis. It deliberately separates an observed change from an explanation that still needs evidence.
Translate one accepted finding into Migma
Migma prepares editable campaigns and series from a goal and approved context. Give it the accepted finding and the scope of the proposed creative change. Avoid dumping an ambiguous dashboard summary into the prompt and asking it to “fix performance.”
For example: “The reviewed report shows readers reach the product page but the current email does not explain the included accessories. Draft a version that makes those approved details easier to scan. Keep the offer, audience, and destination unchanged for this creative review.” This is a hypothetical brief; verify the underlying claim before using it.
Keep the original email available for comparison. If the new draft changes the offer, subject, layout, and audience simultaneously, the team will struggle to interpret what happened next. Choose changes according to the test design your measurement owner accepts.
Verify the final message independently
Run Email Preflight on the finished design and review links, writing, inbox previews, and warnings. Inspect the final populated message in the sending system as well. A correct report does not prove the new button reaches the intended product, and a passed preview does not verify a revenue definition.
Keep the release note concise: accepted evidence, creative hypothesis, controlled elements, changed elements, primary outcome, reporting window, and reviewer. Do not claim that an AI assistant automatically selected a winning variant unless the current product documentation establishes that precise behavior.
Close the loop with an honest result
After the campaign, compare the outcome using the definition agreed before release. If the audience or attribution configuration changed, explain that before presenting a creative conclusion. Report observed association separately from evidence that the email caused additional sales.
The aim is a better decision, not the largest available number. Start the next Migma brief with one verified finding and a specific reader problem. That gives both the designer and the analyst a clear reason for the change and a fair way to judge whether another revision is warranted.
Frequently Asked Questions
Is attributed revenue proof that email caused the sale?
No. A platform’s attribution rule associates outcomes with messages. Estimating additional sales caused by email requires a suitable causal measurement design, such as an appropriately designed randomized holdout.
What belongs in a revenue metric card?
Record the source event, population, period, time zone, currency, refund handling, filters, and attribution definition used in the actual report. Ask the analytics owner to resolve unknowns.
Can Composer’s answer go straight into the creative brief?
Review its interpretation and the underlying report first. Pass an accepted finding and a bounded creative hypothesis to Migma rather than treating an unreviewed answer as established fact.
Should overlapping campaign revenue rows be added together?
Only if the reporting owner confirms they are mutually exclusive under the chosen definition. Otherwise label overlapping assists separately and avoid presenting them as unique revenue.