A Quarterly AEO Review Agenda
Run the agenda in this order: what shipped, what the numbers did, what the control set did, what we cannot explain, what we got wrong last quarter, and what changes next quarter. Putting shipped work before outcomes separates the two ledgers deliberately — the team controls the first and not the second. Putting the control set immediately after the numbers stops a platform-wide movement being presented as a result. And an explicit 'what we got wrong' item is what makes the review load-bearing: a programme that has never revised a claim is not being reviewed, it is being presented.
The failure mode of a quarterly review is not dishonesty. It is that the format rewards a narrative, and the person preparing it is the person whose work is being evaluated.
The agenda
| Item | Question | Why in this position |
|---|---|---|
| 1. Shipped | What did we do? | Controllable; belongs in its own ledger |
| 2. Numbers | What moved? | Observed, not claimed |
| 3. Control set | Did it move too? | Immediately after, so it cannot be skipped |
| 4. Unexplained | What don't we understand? | Makes uncertainty a standing item |
| 5. Corrections | What did we get wrong? | Forces revision of prior claims |
| 6. Next | What changes? | Decisions come last, after evidence |
Why shipped work goes first
Separating shipped work from outcomes protects both. The team gets credit for work that has not paid off yet, and nobody is tempted to attribute a good quarter to whatever happened to ship in it. Presented together, the two always merge into a story where the work caused the number.
The item people remove
"What we got wrong." It is uncomfortable, it makes the review longer, and it is the item that distinguishes a review from a presentation. A programme that has never withdrawn or revised a claim across several quarters is not one with an exceptional record — it is one where nobody checks.
Keep the same metrics, or say you changed them
Changing a metric definition between reviews makes the series meaningless and is nearly invisible to an audience. If a definition changes, show both the old and new figure for the overlapping period. This is the single most common way a review becomes quietly misleading without anyone intending it.
For agencies reporting to clients
The incentive runs the other way — a quarter with no legible result is a difficult meeting. We would still run this agenda, and we would rather say a quarter was flat than claim a movement we cannot separate from a model update. Vendors and agencies who report this way are believed when they do claim a result, which is the only durable version of this.
FAQ
How long should the review be? — Short. Six items, most of them a sentence. A long review is usually one where activity is padding for the absence of outcomes.
What if nothing moved? — Say so, and show the shipped ledger. A flat quarter with honest reporting and a clear next step is a better position than an ambiguous claim that has to be walked back later.
Who should present it? — Whoever did the work, with someone else asking the control-set question. The separation matters more than the seniority of either person.
Put this to work on your own website.
Fenn finds what your customers ask, drafts the articles and site fixes, and measures what ChatGPT, Claude, Gemini, Perplexity and Grok say about you — with every change waiting for your approval.