The claim
“I exported last month's Business Report on the 2nd. Today I pulled it again, and the numbers are different. Which one is right?”
Both were right, at the moment they were pulled.
The usual reading
A closed month is closed. Once a week or a month is over, its sessions, orders, and units are final. If today's pull disagrees with the file you saved, something went wrong — your export, your spreadsheet, or your pipeline.
Premises, laid out
Every conclusion below rests on these. Reject one and you can see which conclusions fall with it.
- ❌Past periods stop changing. They don't. Orders and units keep being revised downward for weeks as cancellations and refunds land.
- ❌Traffic is fixed even if orders move. I used to believe this; a short-interval test seemed to confirm it. Over longer intervals, sessions are revised too — upward.
- ⚠️Small revisions don't matter. At account level they're a few percent. On a single product-week, they can be large enough to change a conclusion.
How we take it apart
Free: conclusion + one screenConclusion: a Business Report number is a reading at a point in time, not a fixed fact. Every number you cite needs the time it was pulled. Comparisons across periods should use data pulled at the same time, and conclusions should lean on ratios, which survive proportional revisions better than absolutes.
The full step-by-step decomposition, including how AI joins the reports, is in the course.
Evidence
Free: one exampleExperience
The same weekly product-level data, pulled again four to five weeks later: sessions went up in every product-week I checked (+5.1% to +31.0%), while units went down or stayed flat (0% to −3.9%). At account level, sessions rose 16.8% and 17.6% in the two weeks I checked; units fell 2.85% and 2.95%. Aggregating did not wash the revisions out.
What would prove this wrong
Free: one sentenceDisproof condition. If repeated pulls of a month-old period returned identical sessions and units over several weeks, I would treat the earlier drift as a one-off and relax the rule.
How you run it
Course onlyReproduce this on your own account
- Step-by-step operations, including the AI prompts used to join reports
- Query templates and the exact fields to pull
- A checklist to rule out the common reading traps
- The full disproof conditions and the failed cases
Ledger status
Status: reading trap, confirmed on my accounts. Business Report figures are revised for weeks: units tend to drift down, sessions tend to drift up. Store the pull time with every number, re-pull before comparing, and prefer ratios to absolutes.
Type, in detail: Experience.
Have data that says otherwise? Share a case. Reversals are published.