The tool, in your browser
The weekly sales scoreboard
One card per week, filled in the same morning every week: inspections booked, inspections you actually ran, estimates out, jobs signed, contract dollars, and how far the schedule reaches. The page works out your own normal week — your average, your best, your slowest — and then does the only comparison that matters: the last four weeks against your own history. By the time the bank account tells you sales are down, the sales were down six weeks ago. This is the front of that chain.
Free, and it asks for nothing. No email, no account, no sign-up. What you type stays in this browser on this device.
Read this first
This scoreboard only knows about this browser. Nothing you type is sent to us and nothing syncs, so the phone and the office computer each keep their own copy — and this one is only worth anything if it’s the same copy every Monday. Clearing your browsing data wipes it, and a new phone won’t bring it along. If this is going to run all year, run it on the-weekly-sales-scoreboard.xlsx at the office instead.
The scoreboard fills in right here in your browser. If this line stays put, JavaScript is off — the spreadsheet below does the same job.
What you actually read
Inspections booked. Everything else is context.
Ad conversions drop, then calls drop, then inspections drop, then signed contracts drop, then the schedule stops reaching as far out as it used to, and finally payroll gets tight on a Friday. Same chain every time. Inspections booked sits near the front of it, it is already in your calendar, and it does not care what kind of work is coming — repairs and re-roofs average out over enough weeks.
So the line to read is the last four weeks against your own average. Under about 90% two readings running is the alarm, and the fix goes into the ad account and the phone that week — there is nothing you can do to the schedule that changes it. Your own average is the only benchmark here. Somebody else’s good week is their size, their market and their crew.
And when the alarm goes off
The other four lines tell you which problem it is.
- Ran well under booked — cancellations, no-shows, weather, a truck that ran out of daylight. You are booking work you never get to. That is a scheduling problem, not a lead problem — the ten-day board.
- Estimates per inspection under about 0.8 — prices are dying in the truck, not at the kitchen table. Running the same visit every time is what fixes that.
- Close rate off — know it, don’t build your week around it. Pull it out when cash gets strange, because a cash problem is usually a sales problem wearing a different coat.
- Schedule reaching less far out — the number a sub-crew shop watches hardest. A board thinned to ten days is the week your crews start taking somebody else’s jobs, and you’re short-handed exactly when your own work comes back.
How it gets used
- Same morning every week, with the calendar open. Four numbers, ten minutes. Fill it in before the week talks you into a better version of itself — the sheet recalculates happily off a number you rounded up, and lies to you every week after. Alongside the money meeting is the easiest place to put it.
- Newest at the bottom, and no gaps. The bottom four weeks are what the answer reads. If a week was dead, log the zeros — a skipped week quietly changes the number, and the page will tell you when it spots one missing between your dates.
- Booked is what was on the calendar; run is what you got on the roof for. They are different numbers and the gap between them is the point.
- Give it four weeks before you read the verdict. Under four the page says so and prints nothing else. It is a trend, not a week.
- If you track nothing at all today, don’t start here. Start by costing a job — the job’s report card. A shop that can’t say whether its jobs make money won’t be around long enough for a lead metric to matter.
- When it goes light, look upstream. This sheet doesn’t know where the leads came from; the lead-source count is the sheet that says which faucet to turn.
Want it in a spreadsheet?
The same sheet as a spreadsheet
the-weekly-sales-scoreboard.xlsx is this scoreboard as a spreadsheet — same columns, same rolling four-week average, same verdict — for the computer at the office. Three differences in the page’s favour: the file stops at 26 weeks and tells you to start a new copy, this doesn’t; when a week doesn’t line up the file’s whole answer block prints one sentence instead of every figure, where here that week says what is wrong and the rest still adds up; and the file cannot see a week you skipped entirely, where this page counts the Mondays missing between your dates and says so. Two the other way: the file is a real file, so it survives a cleared browser and prints for the Monday meeting — and it ships 26 example weeks that quietly slow down at the end, so you can watch the alarm go off before you trust it with your own. Both refuse the same things: a week with a number missing, numbers on a row with no week, more jobs signed than estimates sent, money on a week with no job, and the same week typed twice. The sheets ask for your email once; this page never does.
The play behind it — M9-14 The Number That Tells You Next Month — is where the four numbers come from, and the sheet notes spell out every column and formula.