The Owner's Dashboard: Ten Numbers That Run the Company
By MercConsulting · Published 2026-08-03 · Updated 2026-08-30
The ten numbers every owner should see weekly: cash, AR aging, pipeline, close rate, speed to lead, margins, capacity and more, plus the 20-minute ritual and the automation that keeps a dashboard alive.
The small business KPIs that actually run a company fit on one page: cash on hand in weeks, receivables over 30 days, qualified pipeline value, close rate, speed to lead, gross margin by line, capacity utilization, revenue per employee, repeat or churn rate, and one customer-satisfaction signal. Ten numbers, pulled weekly, each with a target and an owner. For most companies under roughly $10 million in revenue, that is the whole dashboard, and it is enough to catch every problem while it is still cheap to fix.
Most owners live at one of two extremes: running the company off the bank balance and a feeling, or drowning in a 40-metric report nobody reads. Ten is deliberate. It is few enough to review in 20 minutes and complete enough that trouble anywhere in the machine, sales, delivery, cash, or customers, shows up in at least one number.
Here are the ten, how to adapt them to your business model, the leading-versus-lagging discipline that makes them useful, the weekly ritual, and the automation fix for the reason most dashboards die.
Why ten numbers and not forty
A dashboard is not a reporting exercise. It is an attention-allocation device, and its job is to answer one question every week: where does the owner's limited attention go right now? Forty metrics answer nothing because everything blinks at once. Ten forces ranking, and ranking is the actual work of running a company.
Every number on the page must pass three tests: you can get it weekly without heroics, a bad reading has a knowable response, and one person owns it. A number that fails any of the three is trivia, not a KPI.
Cash and collections: numbers 1 and 2
1. Cash on hand, in weeks
Not the raw bank balance; the balance divided by your average weekly operating outflow. The balance alone tells you nothing without the burn beside it. Most service businesses want 8 to 12 weeks; the direction of the trend matters more than the level. Three consecutive down weeks with flat revenue is the earliest honest warning most businesses ever get.
2. Receivables over 30 days past due
Total AR is a vanity number; aged AR is the truth. Track the dollar amount past 30 days and the three largest offenders by name. This number responds fast to systematic follow-up, which is exactly the kind of polite, persistent, rules-based work automation carries well.
The sales engine: numbers 3, 4, and 5
3. Qualified pipeline value
The dollar value of real opportunities, defined tightly enough that wishful thinking cannot inflate it, measured against a coverage target, commonly around three times the revenue goal for the period. This is your best leading indicator of revenue 60 to 90 days out. When it sags, next quarter already happened; you just have not billed it yet.
4. Close rate
Wins divided by decided opportunities over a trailing window. A falling close rate with steady pipeline points at pricing, targeting, or process, and each has a different fix, which is why the number earns its slot: it tells you where to look, not just that something is wrong.
5. Speed to lead
Minutes from a new inquiry to the first meaningful response. It is the most controllable number on the whole page, and one of the most consequential; contact rates collapse as response time stretches from minutes to hours. We cover the mechanics and the fix in the five-minute window.
Delivery economics: numbers 6, 7, and 8
6. Gross margin by line
By service or product line, never blended. A blended margin is where a money-losing line hides for years, subsidized by the winners. Owners who split this number for the first time almost always find one line to reprice or retire.
7. Utilization or capacity
For a service firm, billable share of available hours; for a shop or crew business, equipment or crew utilization. Read it in both directions: too low means overhead is eating you, but months of running near the ceiling means quality slips, people burn out, and there is no room to take the next good client.
8. Revenue per employee
Trailing twelve months of revenue divided by full-time equivalents. Benchmarks vary too much by industry to borrow; use your own trend. This is also the number that shows whether automation is genuinely working: it should climb as systems carry more of the routine load without headcount climbing alongside revenue.
Customer signal: numbers 9 and 10
9. Repeat rate or churn
The share of revenue from existing customers, or for recurring models, the share lost per period. Acquiring a customer costs several times what keeping one does, so a small slide here quietly raises the cost of every growth dollar. It is the number most likely to be ignored while it is still cheap to fix.
10. A satisfaction signal
Pick one and keep it consistent: a one-question would-you-recommend-us survey, review velocity and rating, or complaint rate. The specific instrument matters less than the consistency; what you are watching for is movement.
A restaurant swaps utilization for prime cost and speed to lead for table turns. A contractor tracks backlog weeks instead of pipeline coverage. A subscription business promotes churn to the top of the page. The structure holds: two cash numbers, three sales numbers, three delivery numbers, two customer numbers.
Leading versus lagging: the discipline that makes it useful
Lagging numbers tell you what already happened: revenue, margin, the cash balance. Leading numbers tell you what is about to happen: pipeline, speed to lead, utilization trend. Both belong on the page, but they get used differently, and confusing them is how owners end up managing the rearview mirror.
The rule: every lagging number you care about should have a leading number upstream of it on the same page. Revenue is downstream of pipeline and close rate. Cash is downstream of AR aging. Margin is downstream of utilization. When a lagging number goes bad, the correction always happens upstream; you cannot fix last month's revenue, but you can fix this week's pipeline. If a lagging number on your dashboard has no leading partner, that is the gap to close first.
The 20-minute weekly review
The dashboard only works attached to a ritual. This one takes 20 minutes because the assembly happened before the meeting, not during it.
Monday morning works for most. The dashboard arrives pre-assembled; the meeting is for reading it, not building it. Skip the meeting twice and the dashboard is already dying.
Each number is at target, off target, or trending wrong. Five minutes, no discussion yet. The point of the scan is the whole picture, because problems travel in pairs: sagging pipeline plus high utilization is a different disease than sagging pipeline plus idle capacity.
The one furthest from target, weighted by how much it matters. Resist the urge to address four things; that is how nothing moves.
Write it down. Next week's meeting opens by checking whether last week's action happened and what it did to the number. That closed loop is the entire difference between a dashboard and a decoration.
Why dashboards die, and the automation fix
Dashboards do not die of bad design. They die of manual assembly. Someone spends two or three hours every Monday exporting from the accounting system, the CRM, and three spreadsheets, pasting numbers into a template. Then one busy Monday it slips. Then it becomes monthly. Within a quarter the company is back to the bank balance and a feeling.
"For years I ran the company off the bank balance and a feeling. The first month with a real scoreboard, I found out the feeling was running about two months behind the facts."
The fix is structural: the numbers pull themselves. Every one of the ten lives in a system you already run, the accounting file, the CRM, the phone system, the project tool, and connecting those sources so the dashboard assembles itself is a one-time build, not an ongoing chore. An AI reporting layer can go further: drafting the Monday summary in plain English, flagging which number moved abnormally, and pairing each flag with the trend behind it, so the 20-minute meeting starts at step two.
Before wiring anything, map where each number actually lives and where the data quality is weak; process mapping is the honest first step. Then hold the build to the same standard as any other investment using the automation ROI checklist; a dashboard build that saves ten assembly hours a month typically pays for itself inside a year on those hours alone, before counting a single earlier-caught problem. Designing and wiring that pipeline is core AI and automation consulting work.
Act on one number at a time
The dashboard's power is sequencing, not surveillance. Pick the worst number, work it for four to six weeks until the trend confirms, then move to the next. A focused month on speed to lead, or AR over 30, or the margin on one line, produces visible movement. Trying to move six numbers at once reliably moves none, and teaches the team that the dashboard is weather, not work.
What a framework cannot do is pick your ten. The right set depends on your model, your margins, and where the business is actually leaking; that is a one-conversation diagnosis, not a template.
Frequently Asked Questions
What KPIs should a small business track?
Ten numbers cover most small businesses: cash on hand in weeks, receivables over 30 days, qualified pipeline, close rate, speed to lead, gross margin by line, capacity utilization, revenue per employee, repeat or churn rate, and one customer-satisfaction signal. Adapt the specifics to your model, keep the count near ten, and make sure every number has a weekly source, a target, and an owner.
How often should I review my business KPIs?
Weekly, in a fixed 20-minute review, for the operating numbers; monthly for a deeper financial review with your bookkeeper. Weekly is frequent enough to catch problems while they are cheap and infrequent enough that the numbers have actually moved. The cadence only survives if the dashboard assembles itself; manual assembly is the reason most KPI routines quietly die.
What is the difference between leading and lagging indicators?
Lagging indicators report what already happened: revenue, profit, the cash balance. Leading indicators predict what is coming: pipeline value, speed to lead, utilization trend. You correct a business through its leading numbers, because the lagging ones cannot be changed after the fact. A useful dashboard pairs them, so every result you care about has an upstream number you can still act on.
What is a good revenue per employee for a small business?
It varies too much by industry for a borrowed benchmark to mean much: labor-heavy service firms may run under $150,000 per employee while lean professional or software firms can run several times that. Use your own trailing-twelve-month figure as the baseline and manage the trend. It should rise as automation and better process carry more routine work without matching headcount growth.
How do I automate a KPI dashboard?
Connect the systems where the numbers already live, typically the accounting file, CRM, phone or lead system, and project tool, into one view that refreshes on its own, and add an AI reporting layer that drafts the weekly summary and flags abnormal movement. Map the data sources first, because automating a number nobody trusts just delivers a wrong answer faster. Most builds are one-time projects that pay back on saved assembly hours alone.
Build the scoreboard once
You have the framework: ten numbers, paired leading and lagging, a 20-minute ritual, and a dashboard that assembles itself. What the framework cannot tell you is which ten fit your business and where the data for them actually lives. In a free 30-minute strategy call we will rough out your ten on one page and tell you honestly what it would take to automate the assembly.
Book a free strategy call