The short answer
- Same time every week, or it doesn't happen. Friday afternoon or Monday morning - pick the one your calendar defends.
- Five questions, one page: what came in, what's owed, what's next, what's at risk, what gets better.
- Money in and money owed come from the ledgers, not from memory. If the ledger is thin, that's the finding.
- 'What's at risk' gets one honest sentence per risk. A risk written down is a plan; a risk felt is anxiety.
- The review ends with the first action for Monday, chosen now, while the week's context is still loaded.
The page, question by question
- Money in: what landed this week, from whom, against what invoice. Total it. Compare to last week only - trends need weeks, not days.
- Money owed: every invoice outstanding, its age, and its next action. Anything over 30 days gets a chase this week, no exceptions - that rule alone pays for the review.
- Work next: the projects active this coming week and the one deliverable that matters most in each. Two per project maximum; a list of ten is a list of zero.
- At risk: the one to three things that could go wrong - a client gone quiet, a deadline dependent on someone else, a pipeline with nothing in it. One sentence each, one action each.
- What gets better: the single process that annoyed you most this week, and the smallest fix. One improvement a week is fifty a year; that compounds.
Why one page is the point
- A one-page review gets done on the bad weeks - and the bad weeks are when it matters most.
- The same five questions produce comparable pages, so week 12 can be compared to week 2. Patterns only show in repetition.
- Writing by hand or in one plain file both work. The dashboard is not the product; the noticing is.
The free tool for this
Invoice total, with a discount and a tax rate - Invoice total, with a discount and a tax rate. It runs in the page, asks for your own numbers, and sends nothing anywhere.
If you would rather not build it yourself
Client OS writes the page for you: /weekly assembles money in, money owed with ages, active projects and the chase list from your live ledgers - you bring the judgment, it brings the numbers. ClientOS is $19. See it on the shop, and there is a plain page about it here: what it holds.
Questions people actually ask
Twenty minutes really enough?
For the review itself, yes - if the ledgers are current. Twenty minutes of review plus ten minutes of ledger catch-up is the honest budget, and it's still less than one lost invoice costs.
What about tracking marketing and pipeline?
Pipeline is 'what's at risk' when it's thin - you'll notice. Give marketing its own simple counter (posts out, leads in) rather than swelling the weekly page; the review is for steering, not for everything.
Read next
How to keep track of unpaid invoices without a system you hate, What to look for in a client tracker before you pay for one, How to write a late payment reminder that actually gets paid, The client onboarding checklist that prevents every later argument, What every invoice line item should say (and what never to write).