Who should own each number on a scorecard?
Exactly one named human owns each number, and that is a hard rule rather than a preference. A metric owned by a department is owned by nobody: when the number moves the wrong way, the weekly meeting produces a discussion instead of an action. The owner is the person who can actually move it, not the person who reports it — the dispatcher owns booked-call rate, not the bookkeeper who pulls it. This site enforces the rule structurally: every metric in our graph carries an edge to the seat that owns it, and a metric with no owner fails our build rather than shipping as a row with a blank column.
Sources: B1-01 · B1-06 · B2-01
Why did my weekly scorecard stop getting used after two months?
Scorecards die from four causes, and none of them is a lack of discipline. The numbers have no single owner, so nothing follows a miss. There is no written definition per metric, so two people report the same row differently and everyone quietly stops trusting it. Too many rows come from someone typing them in, and manual entry is the line that lapses first. And the scorecard is not reviewed in a meeting the owner attends, which is the only thing that makes filling it in consequential. Every framework steward publishes the artifact; almost nobody publishes the failure modes, which is why this reads as a personal failing rather than a design one.
Sources: B1-01 · B1-06 · INT-02
What's the difference between a dashboard and a scorecard?
A dashboard is continuous and self-serve; a scorecard is weekly, reviewed and owned; a report is periodic and narrative. The difference that matters operationally is the review: a dashboard is looked at whenever someone is curious, and a scorecard is walked line by line in a standing meeting where a miss has a name attached to it. Most newly-acquired companies buy a dashboard when what they need is a scorecard, because the dashboard is a purchase and the scorecard is a habit.
Sources: B1-01 · INT-02
How do I choose which KPIs to track in a business I just bought?
Start from the model, not from a list: what a trades business needs measured weekly and what a distributor needs measured weekly are different answers, not the same answer with different labels. Then apply three filters — can somebody in this company move it, does it come out of a system rather than out of someone's memory, and would a bad week show up here before it shows up in the bank account. The specific difficulty of an inherited business is that you do not yet know what normal looks like, so the first quarter's job is establishing your own baseline rather than comparing yourself to a benchmark you cannot verify.
Sources: B2-01 · A2-02 · INT-02
How many numbers should be on a weekly scorecard?
Roughly five to fifteen, and we present that as an editorial judgement rather than a finding. The framework stewards converge on "a handful" and none of them publishes a failure threshold, so there is no study behind any specific ceiling and we are not going to manufacture one. The mechanism is what to reason from: a scorecard is read line by line in a five-minute segment of a weekly meeting, and past some number of rows it stops being read and starts being scrolled. If yours has thirty rows, the question is not which fifteen to cut — it is which of them anybody acted on last quarter.
Sources: B1-01 · B2-01 · no benchmark value published
What is revenue per technician per day?
Revenue per technician per day is total revenue produced divided by technician-days worked in the period, and there is no verified benchmark for it — ours or anyone else's. Every figure in circulation for this metric traces back to field-service software marketing: a vendor blog states a number, a listicle cites the blog, a consultant cites the listicle, and it becomes "the industry average". The one industry financial survey that could settle it is 2020-vintage and paywalled. So use it as a trend against your own baseline and against your own dispatch board, and treat any specific target you are quoted as unverified until somebody shows you the survey.
Sources: B2-09 · no benchmark value published
What is OEE and is it worth measuring in a small shop?
Overall equipment effectiveness is availability × performance × quality, expressed as a single percentage for a machine or a line. One number attached to it is genuinely verifiable — the 85% composite — and it is routinely misquoted: its originator framed it as a minimum to strive for, not as an industry average, which is the difference between a target and a grade. The per-factor targets everyone reproduces alongside it exist only as an image on one site and are not published here. In a small shop OEE earns its keep only where one machine is the constraint; measured across a job shop it mostly produces a number nobody can act on.
Sources: B1-12 · B1-10 · no benchmark value published
What is utilisation rate and how do I calculate it?
Utilisation is billable hours divided by available hours for a given person or period — and before you compare yours to anything, check which definition the comparison uses. One benchmarking tradition computes it on hours and another computes it on dollars, so two firms both reporting "72% utilisation" may not be describing the same quantity at all. This is the one business model where free, current, formula-complete benchmark corpora genuinely exist, which makes getting the denominator right worth the trouble.
Sources: A2-10 · A2-11
How do I write a definition for each metric so everyone reports it the same way?
Write one line per metric covering four things: the exact formula, the system the data comes from, the period it covers, and the single person who owns it. That document — a metric definition dictionary — is what makes a scorecard survive past month three, because the alternative is two people computing "revenue" differently and nobody noticing for a quarter. It is also the antidote to the traps in this hub: utilisation, overhead rate and quick ratio each mean two incompatible things depending on whose tradition you inherited, and a written definition is where that gets settled once.
Sources: B1-01 · B2-01
What's the difference between a KPI, a metric, and a scorecard?
A metric is any number the business produces; a KPI is a metric someone is accountable for, with a target and a standing review; a scorecard is the weekly document a set of KPIs lives on. The promotion from metric to KPI is the entire distinction, and it is made by attaching an owner and a target — not by putting the number on a slide. Most newly-acquired companies have hundreds of metrics and no KPIs, because nothing is owned. The test: if the number moves the wrong way and no single named person has to answer for it in a meeting, it is a metric.
Sources: B2-01 · INT-01 · INT-02
What's the difference between a leading and a lagging indicator?
A lagging indicator reports the result you wanted; a leading indicator measures the behaviour that produces it and can still be influenced this week. Revenue, gross margin and retention are lags — they are true, and by the time they move the cause is a month old. Quotes issued, calls booked, jobs scheduled per crew and proposals out are leads. The 4DX formulation of this is the most transferable idea in the whole cadence literature for a company your size: a scorecard built only from lagging numbers is a report, and a report cannot change what anybody does on Tuesday.
Sources: B1-01 · B1-04
At what headcount does cascading KPIs start to make sense?
There is no sourced headcount threshold for cascading a scorecard, and the figure we would otherwise hand you — somewhere around thirty people — is our editorial judgement rather than a finding. What is defensible is the mechanism. Cascading pays only when a department has numbers its own manager can move without the owner in the room; below that, a departmental scorecard is a second document restating company numbers in smaller type. So the signal to watch is not headcount at all — it is whether a department head is already accountable for an outcome they cannot see on the company scorecard.
Sources: B1-01 · B1-06 · no benchmark value published
What's the difference between a job scorecard and a KPI scorecard?
A job scorecard defines a role before you hire for it; a KPI scorecard is the weekly document where company numbers get reviewed. They share a word and nothing else. The job scorecard states the outcomes the seat must produce, the competencies it needs and how success will be measured, and it is written once per role. The KPI scorecard is walked line by line in a standing meeting with an owner against each number. One is a hiring instrument, the other is a management rhythm. Confusing them produces the two familiar failures: a job description with no measures, and a weekly meeting that turns into a performance review.
Sources: B2-09 · INT-04
What is a first-time fix rate and why does it matter?
First-time fix rate is the share of jobs resolved on the first visit, and there is no verified benchmark for it — every figure in circulation traces to field-service-software marketing. It matters because the second visit is paid for twice by you and once by the customer's patience: the truck roll, the technician hour and the scheduling slot are all consumed again to deliver revenue that was already booked. Measure it against your own baseline at close, and split it by technician and by job type — the blended number hides the crew, the part or the diagnostic step actually causing it.
Sources: B2-09 · no benchmark value published
What is GMROI and how do I calculate it?
GMROI is gross margin dollars divided by average inventory cost, and it answers the question a distribution business actually turns on: how much margin each dollar tied up in stock returns over a period. We publish no benchmark for it, because not one primary distribution source is reachable — the trade associations return 403 and the named operating-performance reports 404 — so any industry figure you are quoted came from somewhere neither of you can open. Run it by product line rather than for the warehouse: the aggregate is a fast-moving line quietly funding a slow one, which is the thing you need to see.
Sources: B1-11 · B2-01 · no benchmark value published
