Your Dashboard Is Full and You Still Cannot See Ahead

Most management packs we review are accurate, thorough, and blind. Revenue by month, gross margin, deals closed, headcount, utilisation against last year. Every number is correct. Every number describes a period that has already ended.

A business measured this way learns about its problems on the same schedule as its bank. By the time a revenue chart bends, the cause is two or three quarters old and the cheap window for fixing it has closed.

Measurement earns its cost when it buys time to change course.

The distinction that matters

A lagging indicator records an outcome. Revenue, profit, churn, turnover, project margin at completion. These belong in the pack, they are how you keep score, and they cannot be managed because the events that produced them are finished.

A leading indicator moves before the outcome moves. It gives you a period of warning while the response is still cheap.

The mechanical version of the question is this: which number, if it deteriorated for six weeks, would tell me that a lagging number will deteriorate next quarter? Answer it once per major outcome and you have a small set of measures worth reviewing.

Pairings that work in mid-market businesses

Revenue. Watch qualified pipeline against the coverage ratio your close rate requires, along with proposals issued and average sales cycle length. Pipeline thins before revenue does, and it thins in a way you can act on.

Cash. Watch weeks of runway on a rolling thirteen-week forecast, days sales outstanding, and the ageing of your largest three receivables. Cash crises are visible in advance in almost every case, and missed in advance in most.

Margin. Watch price realisation against list, rework hours, and scope changes granted without a fee. Margin erosion starts as a series of small concessions nobody records.

Delivery and client retention. Watch on-time completion, first-response times, support volume per client, and the number of accounts without a scheduled senior conversation in ninety days. Clients disengage before they leave.

Team. Watch missed one-to-ones, voluntary overtime concentrated in the same few people, internal application rates, and time to fill open roles. Turnover announces itself for months.

Concentration. Watch revenue share of the top client and top three clients, and revenue by acquisition channel. Concentration risk is a slow build with a sudden cost.

Pick two per outcome. A business with forty metrics has none, because nobody looks at forty numbers with attention.

Cadence, and one rule

Set the review rhythm before you need it.

Weekly, look at execution and the leading indicators. Short meeting, same measures, no narration of the numbers everyone can read.

Monthly, look at the financials against plan, and at the gap between what you committed to ninety days ago and what exists now.

Quarterly, look at the structure itself. Is the plan still pointed at the outcome, or has the market moved while the plan stayed still. This is the review most businesses skip, and the one that prevents the largest errors.

The rule that gives all of this force: treat a leading indicator falling for two consecutive periods as a present problem with an owner and a date, not as an item to monitor. Monitoring is what a business does while a problem matures.

Where measurement goes wrong

Vanity counts. Impressions, follower growth, meetings held, activity volume. These correlate with effort rather than outcome.

Averages hiding the distribution. An average project margin of nineteen percent can contain a client losing money every month. Segment before you average.

Metrics without owners. A number on a dashboard with no name against it is information. A number with an owner is management.

Instrumentation that cannot be trusted. If your team debates whether the figure is right, the review becomes an argument about data and never reaches the decision.

Good instrumentation is a small set of numbers, reviewed on a fixed cadence, each owned by a person, chosen because they move first.

If your reporting tells you what happened and never what is coming, the fix is a design problem, and it takes weeks rather than a new system.