About this template
The deck opens with three wins and never mentions the targets you missed
Because nobody wrote them down anywhere findable. It takes a day to build and changes no decision.
The scorecard is the whole point
If a prior review sits in ~/qbr/reviews/, the first section reads last
quarter's targets and marks each one hit, missed, or unmeasurable.
That single section turns a QBR from a slideshow into a loop — and it's the part almost every small business skips, because the targets were in a deck nobody reopened.
Misses before wins
A review that opens with wins is a status report. One that opens with misses is a decision document.
Each miss gets the number, the gap, and the one cause the data supports — not
three speculative ones. If the data doesn't identify a cause it says cause unclear and names what would. An invented cause is worse than an admitted gap.
A bare number is a fact, not a finding
Every metric is reported against at least one of:
- Prior quarter — direction and size
- Same quarter last year — the only honest comparison for a seasonal business
- The target that was set — the most useful, and the most often skipped
No vanity metrics without a money link
Impressions, followers and page views appear only where your export connects them to revenue, pipeline or retention. Otherwise they're omitted — a QBR padded with reach numbers hides the quarter.
Every figure is sourced
No estimating, no "roughly", no filling a gap so a table looks complete. Missing
metrics show as [MISSING: what] with the decision each one blocks — which is
what actually gets them supplied next quarter.
A QBR is the document a business decides from. One invented number makes the whole thing untrustworthy.
Next quarter: three targets, maximum
Labelled proposals, not commitments. A review proposing eight targets proposes none.
What you supply
Your own Claude Code login, plus your P&L and sales exports in ~/qbr/exports/.
Keep past reviews in ~/qbr/reviews/ — that's how next quarter gets scored against
this one. The loop only closes if the targets stay findable.
No accounting API key. Exports are the input, which is also why nothing can be sent or committed.
Where this stops
It drafts. It does not email the review, present it, or commit the business to any target in it.
It won't read correlation as causation, won't cite an industry benchmark it doesn't have, won't forecast from two data points, and won't tell you the quarter was "good" — it reports against your targets and lets you judge.
Verified on build: the skill, the working tree, both guards, all three comparisons, the misses-first rule and the vanity-metric rule are present on a fresh fork. Running it on your own quarter is yours to do.