Weekly Business Review for Founders: What Amazon Scans in Under an Hour

Amazon's Weekly Business Review put the top 50 leaders of the retail business in one room with 200 to 300 metrics, every week, and the meeting was not optional. Bill Carr, who spent 15 years as an Amazon vice president and co-wrote Working Backwards, describes the format in a post of his own: a 6-12 view, current 12 months against the trailing 12, plus the current and trailing 6 weeks, with every metric laid out the same way.

The sameness was the trick. Carr's other post on the practice gives the reason. Each new chart type forces the reader to reorient, and a hundred reorientations is cognitive overload.

With one format, the team could scroll hundreds of metrics in under an hour and still catch what had moved. He quotes Bezos in the same piece: "Humans are unbelievably data efficient."

A founder reading this usually takes one of two lessons, and one of them is expensive.

The expensive one is to build the 200-metric version. You have one product, a few real numbers, and no finance team. The ritual carries the value here, not the volume.

The useful one is the distinction the meeting is built on: controllable inputs against outputs. Adam Fishman wrote a practical playbook for running these reviews outside Amazon, and lists the problems they solve, including teams being "focused on the wrong things (lagging outputs rather than inputs)". In the Amazon review, Carr's account and Commoncog's write-up agree, output metrics were reported but not debated. The discussion stayed on the inputs that move them.

That constraint survives any scale.

At founder scale it looks like this: 20 minutes on Friday, one page, three inputs numbered against last week. Conversations with people who could buy. What shipped. The decision you are carrying, with what you expected and when you will know.

Amazon's own inputs took years to settle. Selection started as the number of new detail pages. It became detail page views once nobody got credit for a page no customer opened, then the share of those views where the item was in stock.

It ended as Fast Track In Stock: the share of detail page views where the product was in stock and ready for two-day shipping. Colin Bryar, who ran the process as Bezos's technical assistant, tells the story as a warning. Teams chase whatever number you give them, so a lazy input quietly fills warehouses with things nobody wants.

The lesson is speed of correction. A wrong input caught in the weekly review costs weeks. Caught in a quarterly plan, it costs the quarter.

David Allen's weekly review shares the calendar slot and almost nothing else. His 11 steps are about getting clear: collect the loose paper, get the inbox to zero, empty your head, review projects and waiting-fors. Tiago Forte calls that review an operating system, the master program the week runs on, and his own checklist takes him from total chaos to total clarity in 15 to 30 minutes.

Run both on the same Friday. Clear the loose ends first, then read the three numbers, then write the decision line.

The page is the artifact. A review you cannot re-read is a meeting you will have again next week.