How to Build a Weekly Business Dashboard

Six is not arbitrary. It is roughly the number a person can hold in mind and genuinely react to. Beyond that you are scanning rather than reading, and scanning produces no decisions.

1. Cash on hand

Reconciled, across operating accounts. The single most important number in a small business and the one most often known only approximately.

2. Cash committed, next fourteen days

Payroll, rent, loans, bills due. Cash on hand without this is half an answer — $60,000 is comfortable or alarming depending entirely on what leaves next week.

3. Revenue booked this week

Work won, not work invoiced. This is the leading indicator; invoiced revenue is a lagging one.

4. Gross margin on work completed

The number that reveals problems earliest. Volume can look healthy for months while margin erodes underneath it, and only this catches it.

5. One pipeline measure

Backlog in weeks, quotes outstanding, or pipeline value. Whichever genuinely predicts your next month.

6. One operational measure

Specific to you — utilization, on-time completion, callback rate, jobs per crew. The one that most often explains the others.

A bare number carries no information. "$47,300 in the bank" is not good or bad. "$47,300, down from $61,000 last week, against $52,000 the same week last year" is a situation.

Two comparisons are enough:

• Last week — direction and momentum.
• Same week last year — strips out seasonality, which otherwise makes every seasonal business panic annually on schedule.

Rolling four-week averages are worth adding for anything volatile. A single bad week in a lumpy business is noise, and reacting to noise is its own failure mode.

This is what separates a dashboard from a report. Each number gets a threshold, written in advance, that triggers a specific action.

• Cash below four weeks of operating costs → work the collections list this week.
• Gross margin down more than two points over three weeks → cost the last five jobs individually.
• Backlog below three weeks → quoting becomes the priority regardless of what else is happening.
• Overtime above a set share of payroll for two consecutive weeks → capacity conversation.

Set these when calm. A threshold decided during a bad week is a rationalization, and the whole purpose is to have made the decision before the emotion arrives.

Same time every week, fifteen minutes, four questions:

1. Did anything cross a threshold? If not, you are done. Most weeks, nothing has.
2. What moved most, and do I know why? An unexplained large movement is the finding.
3. What did I say I would do last week? One line, checked.
4. What is the one thing this week? One, not five.

Fifteen minutes is deliberate. An hour-long weekly review will not survive a busy month, and surviving busy months is the only thing that matters — because busy months are when the numbers move.

Friday afternoon works well for most owners: the week is complete, and the next one is not yet making demands. That is the same slot the weekly business review occupies.

Three things move fast enough to warrant a daily glance, and they are not the same six:

• Cash on hand — thirty seconds, and it prevents most surprises.
• Anything urgent from a customer — a complaint or dispute ages badly, as the complaint guide covers.
• Today's commitments — what you promised and to whom.

Everything else is weekly at most. Checking margin daily produces noise and, worse, produces reactions to noise. The measurement interval should match the rate at which the thing actually changes.

A useful test: if a number could not plausibly move between Monday and Tuesday, checking it on Tuesday tells you nothing you did not know on Monday. Backlog, margin and pipeline all fail that test. Cash passes it, which is why cash is the exception.

The design is easy and the plumbing is what determines whether it survives. Each of the six has a source, and knowing it in advance stops the weekly build turning into a scavenger hunt.

Cash on hand

The reconciled balance across operating accounts. Not the bank app figure, which includes transactions that have not cleared. This is why the dashboard sits downstream of a monthly close — an unreconciled starting point makes every derived number approximate.

Cash committed

Payroll from your provider's schedule, bills due from the AP aging, plus the fixed items that recur on known dates. The quarterly and annual ones are the trap, exactly as they are in the 13-week forecast.

Revenue booked

From accepted quotes or signed work, not invoices. Most businesses have to establish this deliberately, because their system records invoicing rather than winning.

Gross margin on completed work

The hardest of the six, and the one that requires job costing to be real. If labor is not burdened, this number will be wrong in a consistent direction and everyone will eventually stop trusting it.

Pipeline and operational measures

Whichever you chose. The rule is that both must be pullable in under two minutes, because anything slower is what turns a ten-minute build into a forty-minute one, and forty minutes weekly is the point at which the habit reliably dies.

Too many numbers

The commonest cause by a distance. Twenty-two metrics means no signal, and no signal means no reason to look.

Numbers nobody trusts

If margin is calculated on unburdened labor, everyone quietly knows it is wrong and stops using it. Accurate inputs come first — which is why this sits downstream of true job margin rather than upstream.

No threshold, so no action

Covered above, and it is the difference between a dashboard and wallpaper.

Manual assembly

If producing it takes forty minutes, it will not survive. The production has to be near-free or the habit dies regardless of how good the design is.

Built for someone else

A dashboard assembled to look rigorous to an accountant or investor will not be read by the owner. Build it for the person making decisions, which in a small business is one person.

Reacting to every movement

The opposite failure, and less discussed. An owner who changes something every week in response to a number is not managing, they are oscillating — and the business feels it as constant direction changes that never have time to work.

This is what thresholds are actually for. Most weeks nothing has crossed one, and the correct response to that is to close the dashboard and go back to work. A review where the answer is "nothing needs attention" is a successful review, not a wasted one.

Never revising it

The six numbers that mattered when you had four employees may not be the six that matter at twelve. Revisit the selection annually and be willing to retire a metric — including one you introduced and were fond of. A dashboard measuring last year's constraints is quietly misleading, because it still looks authoritative while pointing at the wrong things.

Sharing it with nobody

A dashboard read only by the owner keeps the whole picture in one head. Showing the two or three numbers a team can actually influence — backlog, on-time completion, callback rate — changes behavior in a way that instructions do not, because people can see the effect of their own work. Keep the cash and margin lines private if you prefer; share the operational ones. A number nobody outside your own head can see cannot change how anyone works, which for several of the six is most of the available value.

Design is a one-off. The killer is assembly: forty minutes every week to gather numbers is thirty-five hours a year, and it is the reason most dashboards are abandoned rather than any flaw in the numbers chosen.

If it cannot be produced in under ten minutes, reduce what is on it or automate it. Those are the only two options that work.

Choosing the six numbers and setting the thresholds is judgment about your own business and it should be yours. Reading it and deciding is the point and cannot be delegated at all.

Assembling it — pulling cash, calculating committed outflows, working out margin on completed work, comparing to last week and last year — is pure production, and it is the part that determines whether the habit survives past week six.