KPI Tracking for Small Companies: Fewer Numbers, Better Decisions

A dashboard with forty metrics is a dashboard nobody reads. The discipline is choosing the few numbers that change what you do.

The test

For each metric, answer: what decision would I make differently if this number moved?

If there is no answer, it is a number you are collecting rather than a metric. Most dashboards fail this test on most of their contents.

Two supporting questions: is it timely enough to act on, and can it be gamed without improving the underlying thing?

A short list by function

Company level (three, at most five)

• ›Cash position and runway. The one that determines whether you exist.
• ›Revenue against plan.
• ›One customer health measure: retention or net revenue retention.

Sales

• ›Pipeline coverage against target.
• ›Win rate, trending.
• ›Sales cycle length. See increasing deal velocity.

Delivery

• ›On-time delivery rate.
• ›Utilization, paired with margin so it is not optimized alone.
• ›Project margin distribution, not the average. See project profitability.

People

• ›Regretted turnover.
• ›Time to hire, by stage.
• ›One engagement measure, quarterly.

Finance

• ›Days sales outstanding.
• ›Gross margin.
• ›Burn or profit, against forecast.

That is roughly fifteen numbers for an entire company, and most SMBs would be better run on these than on what they currently track.

Leading versus lagging

Lagging indicators tell you what happened: revenue, margin, turnover. They are accurate and too late to act on.

Leading indicators predict: pipeline coverage predicts revenue, missed one-on-ones predict turnover, budget consumption rate predicts project margin.

A good set has both, and the review conversation should spend more time on the leading ones, because those are the ones you can still influence.

Review cadence

Weekly: cash, pipeline movement, delivery exceptions. Short, operational.

Monthly: the full set against plan, with explanations for variances.

Quarterly: whether these are still the right metrics. Businesses change and metric sets calcify.

Where AI helps

Not in producing the numbers, which any tool does. In explaining movements: which segment drove the change, what coincided with it, which accounts are behind it. That analysis is what turns a review from reporting into deciding. See AI business intelligence.

FAQ

What if leadership wants more metrics?

Offer the short list as the standing review and the rest as available on request. Most requests for more metrics are requests for a specific answer, which is a question, not a dashboard.

How do we pick between two candidate metrics?

Choose the one closer to a decision and harder to game.

Should every team have its own dashboard?

Teams need their operational numbers. The company set should stay small and shared, or you get functional silos with their own versions of the truth.

Brainis BI OS builds dashboards across every module, with Cortex explaining what moved and why. Included on every plan. See BI OS.