The numbers worth looking at weekly.

Total revenue hides the thing that matters most. A month where marketplace orders replace direct ones can show flat sales and a materially worse profit, because the two channels have completely different costs. Every number on this list is more useful split by channel than in aggregate, and the split itself is one of the numbers.

Each of these answers a question you will actually act on. If a metric does not change what you do next week, it does not belong on a weekly report.

Of the eight, repeat customer rate is the most predictive and the least commonly measured, because it needs orders attached to customers rather than to tickets. It is the closest thing the trade has to a leading indicator: it moves before revenue does, in both directions. If it is falling while sales hold, you are replacing regulars with first-timers and the bill for that arrives a quarter later.

A number without a comparison is trivia. Compare against the same day last month and the same week last year, not against the previous week, because trade is weekly-seasonal and Monday against Sunday tells you nothing. Where you have it, a rolling four-week average smooths out the weather and the odd bank holiday.

Social followers, page views, app downloads and total revenue are the four that get reported most and change decisions least. They are not meaningless, but they are downstream of everything on the list above, and watching them weekly mostly generates a feeling rather than an action. If you look at one number a week, make it gross profit by channel.