Stop 8 of 9 · Weekly
The client review
Turn twelve numbers into one story, attach a confidence line to the recommendation, and make the budget conversation part of the review.
What it is
The review is where a month of work becomes a renewal or a slow fade. Nour has twelve accounts of numbers and one hour per client. This stop is how those numbers become one story, why every recommendation carries a confidence line, and how the budget conversation stops being a separate awkward call.
Twelve numbers, one story
Owners do not read tables. They hold one sentence and one decision. The work is deciding which sentence, and it starts from the goal the client agreed, not from last month.
The shape that holds up: result against goal, the trade behind it, what we changed and what it did, what we are doing next, and the one thing we need from them. Everything else is an appendix.
[upload this month and last month exports for one client, plus the goal we agreed] Write the review for [client code] against their agreed goal. Structure: one sentence on result versus goal; the trade the numbers represent, in plain words, with both figures; what we changed this month and what each change did, in dollars; what we do next month; one thing we need from the client. Under 400 words. Mark any figure you cannot show rows for as unverified rather than dropping it, and list those at the end.
The confidence line
Every recommendation gets three sentences attached: how confident, on what evidence, and what would change the answer. This is the part most reviews leave out, and it is the part that makes a client trust the next recommendation after one goes wrong.
Take the budget recommendation above and add the confidence line. Say how confident you are and why, name the evidence with its numbers, state the level at which you would no longer recommend it, and give the specific trigger that would make us reverse within two weeks. Three sentences. No hedging words without a number attached.
The budget conversation
Ask for more budget when the account can prove it can spend it, and say when it cannot. Impression share lost to budget on campaigns already at target cost per lead is the evidence. A ceiling makes the ask credible, because it shows you are not simply asking for more every month.
Here is the C06 review and the budget recommendation. List the five questions this client is most likely to ask, hardest first, and give me the two sentence answer to each with the number it rests on. Then tell me the one question I cannot answer from this data and what I would need to pull before the call.
What Claude does not do here
The review recommends. Nothing in it changes the account. The budget moves after the client says yes and a person makes the change, and both the yes and the change go in the guardrail log with the date.
The measurable
Two numbers per review: how many of the recommendations from the previous review were acted on, and how many turned out the way the confidence line predicted. The second number is your calibration, and clients renew on calibration more than on results.