A panicked client email, verified line by line — and three claims corrected
An e-commerce software client sent an alarming email: traffic had dropped hard, roughly 1.5x worse than usual, and they'd lost a large sum in ad revenue that week. AOS refused to take the narrative at face value. It checked each claim against owned data, produced a verdict table that corrected all three claims as false, and identified the real problem — ROAS erosion driven by average order value falling ~17.6%.
01 Ground
When a client emails in a panic, the path of least resistance is to mirror their alarm and start firefighting the problem they named. That's how teams end up spending a week chasing a traffic collapse that didn't happen while the actual issue compounds quietly underneath.
This client's email contained three specific, emotionally charged claims: traffic had dropped sharply, the drop was about 1.5x the usual volatility, and a large amount of ad revenue had been lost in a single week. Each was stated as fact. None had been checked against the owned data — they were the client's interpretation of a number that felt wrong.
The job was not to reassure and it was not to agree. It was to verify each claim independently against the data the business actually owns, and — if the claims were wrong — to say so plainly and point at the real cause, even though that means contradicting an anxious client who is certain they already know what's happening.
- Claim 'traffic dropped hard' — corrected: not supported by owned traffic data GA4 2026-06-21 88% confident
- Claim 'lost large ad revenue this week' — corrected: real issue is ROAS erosion, not a revenue cliff Google Ads / GA4 2026-06-21 87% confident
- Actual root cause: average order value down ~17.6% GA4 / Google Ads 2026-06-21 86% confident
02 Decide
AOS structured the response as a verdict table: each of the client's three claims on its own row, checked against owned data, marked true or false with the supporting figure. All three came back false as stated — the traffic collapse and the one-week revenue loss were not what the data showed.
Crucially, refusing the narrative was only half the work; the other half was pointing at the truth. The real problem was ROAS erosion, and AOS traced it to its driver: average order value down ~17.6%. That reframes the entire conversation — the client thought they had a traffic and ad-spend problem, when in fact the same volume of customers were simply spending less per order. The fix lives in basket value and merchandising, not in panic-buying more traffic.
03 Deploy
AOS took the client's email as a set of hypotheses to falsify rather than facts to act on. It pulled the owned traffic and revenue data, tested each of the three claims against it, and assembled a verdict table plus a root-cause finding into saved files and a child run.
The parent message is deliberately thin — the substance is in the artifacts. That's a deployment caveat worth knowing: the value lands in the saved deliverables and the child run, and the router took ~27 minutes, so this is a correctness instrument, not a speed one.