Most restoration scorecards are built backwards. They open with last month's revenue, spend twenty minutes explaining it, and end without anybody learning something they could act on this week. Revenue is a receipt: it tells you what the company already did, long after the decisions that caused it.
A restoration company is really three businesses stapled together — mitigation, rebuild, and contents — and they run on different clocks. Mitigation bills fast and collects in weeks. Rebuild carries a backlog for months. Contents is its own margin question entirely. Roll them into one company number and the signals cancel each other out: a strong mitigation month hides a rebuild pipeline that quietly emptied. So the first rule is that every number below is tracked by department, not only at the company line.
leading beats lagging
A leading indicator moves before the money does. Jobs entered this week, estimates out the door, and production backlog by department are leading. Revenue, collected cash, and last quarter's margin are lagging.
The distinction matters because only one of those lists can still be changed. If jobs entered drops three weeks running, you have found next quarter's revenue problem while you still have this quarter to work on it — call the referral sources, check whether the phones are actually being answered, look at what the ad account did. Find the same problem in the revenue line instead and you find it a quarter late, when the only move left is explaining it.
A scorecard full of lagging numbers is a rear-view mirror. Useful for knowing where you have been, but a meeting spent staring at it produces commentary, not decisions.
a scorecard worth arguing about
Here is the short list. Eight to twelve rows is plenty — a forty-row scorecard gets skimmed, and a skimmed scorecard is a decoration.
- jobs entered this week, by source — the earliest signal you have, and the only one that says whether marketing and referral relationships are still working.
- estimates written and approved — the conversion step between a job and revenue; a widening gap between written and approved is a scoping or sales problem, not a lead problem.
- production backlog by department — how many weeks of work each department is holding. This is next month's revenue before it is revenue.
- gross margin by department — where the money actually is. Mitigation, rebuild, and contents carry different margins, and averaging them hides whichever one is bleeding.
- AR days — how long cash sits between invoice sent and money deposited. The number most likely to be quietly getting worse while nobody watches it.
- collection rate on closed jobs — the share of what you billed that you actually collect. The difference between a good month on paper and a good month in the bank.
- ad cost per booked job — not cost per lead. The only ad number that connects spend to work you can invoice.
- closeout time, work-complete to invoice-sent — pure self-inflicted delay, and usually the fastest number on this list to improve.
Those last two are where owners tend to be most surprised. Say your AR sits at 47 days and your average closeout takes nine days after the crew finishes. Nine of those 47 days are yours — recoverable without a single collections call, just by getting paperwork off a desk. The rest is a negotiation. That one is not.
the ownership rule
A number with no owner is a decoration. Every KPI on the scorecard gets exactly one name beside it — not a department, not two people, one person who is expected to know why the number moved and what they are doing about it.
One name, not two — be strict about this. Shared ownership means that when a number goes red, both owners quietly assume the other one is handling it. A KPI that genuinely belongs to two people is usually two KPIs.
Ownership only means something if there is a standing time to be asked. That is the weekly leadership meeting — a Level 10, if you run EOS — where the scorecard is reviewed first, every owner calls their number on track or off track in a single sentence, and anything off track becomes an issue solved later in the same hour. If you have not run one before, here is what an L10 meeting actually looks like. The scorecard and the meeting are one mechanism; either one alone decays within a quarter.
make it automatic or watch it die
Hand-filled scorecards die on a predictable schedule. Week one it gets filled on Monday. Week four it gets filled the night before the meeting. Week nine somebody pastes in last week's numbers because the report did not run. By week twelve the meeting opens with we will get to the scorecard next week.
That is not a discipline problem, it is a plumbing problem. The numbers already exist. Jobs, estimates, and backlog live in the CRM. Margin, AR, and collections live in the books. Spend lives in the ad account. They sit in three systems that do not talk to each other, so somebody retypes them every week and produces a scorecard nobody fully trusts — including the person who typed it.
Pull the numbers directly instead. When the CRM, QuickBooks, and the ad platforms feed the scorecard on a schedule, the numbers are on screen before the meeting starts and the conversation moves from is this right to what are we doing about it. That is what our scorecard was built to do, and it is how the weekly meeting runs in the restoration company these tools came out of. It is the same wiring our data tools work installs for a restoration company that already has the meeting but cannot trust the sheet.
Start smaller than feels right. Six numbers, pulled automatically, with a name on each, reviewed at the same hour every week, will beat a forty-row scorecard assembled at 11pm. Add a row only when somebody asks a question the current scorecard cannot answer.