The five numbers you should be able to answer in five minutes

Operational clarity is testable, and the test takes five minutes.

Ask for five numbers. If you get them from one place, with the same definition everybody else would use, you have clarity. If getting them requires a request, a reconciliation and a caveat, you have reporting — which is a different thing, produced by a different process, and trusted differently.

Here are the five, with what each one actually measures and how each one fails.

1. Pipeline to capacity

What it is. Weighted pipeline over the next quarter, set against the delivery capacity you can actually field in that period — by skill, not by headcount.

How it fails. Pipeline is held in the CRM at deal level and capacity is held in a resourcing sheet at person level, and the two use different time buckets. The number can only be produced by someone who knows how to map one onto the other, which means it exists in a person rather than in the business.

What movement tells you. Rising pipeline against flat capacity is a hiring or subcontracting decision with a lead time attached, and the lead time is the whole point of measuring it early. How that view is composed is covered on our resourcing page.

2. Project health

What it is. Burn against budget, open change requests, and risks, per engagement, as a current state rather than a fortnightly report.

How it fails. Budget lives at contract level and burn accumulates at task level, so the comparison is made by hand. Change requests sit in email until someone converts them. The number is right on the day it is compiled and drifts immediately afterwards.

What movement tells you. A project going amber is only useful if it goes amber while there is still scope to act. Health reported monthly is health reported after the decision window has closed.

3. Utilisation and throughput

What it is. Billable hours as a proportion of available hours, alongside the work actually completed. Both, deliberately — utilisation on its own rewards being busy.

How it fails. Time is entered late, so the current fortnight is always understated and last month is always being revised. Firms compensate by only trusting the number once it is old, which removes its operational value entirely.

What movement tells you. High utilisation with flat throughput is a rework signal. It is the single most useful pairing on this list and the one most often reported as a single figure.

4. WIP to invoice velocity

What it is. How long work takes to travel from delivered to invoiced, and what is currently sitting in between.

How it fails. WIP is assembled at month end rather than held continuously, so the question “what is in WIP today” has no answer. When the balance is challenged, the response is a set of conversations rather than a set of records — which is the difference between a WIP register and a WIP argument.

What movement tells you. This is the metric with the most direct cash consequence, and the one where a small structural improvement compounds every cycle rather than once.

5. Forecast versus actuals

What it is. What you said would happen — revenue, margin, delivery dates — against what did, at engagement level rather than in aggregate.

How it fails. Forecast is produced in a spreadsheet by a different process from the one that produces actuals, so variance is always partly a measurement artefact. Nobody can separate the estimating error from the accounting difference, and so nobody learns anything from the gap.

What movement tells you. Consistent one-directional variance is a pricing or estimating problem, not a delivery problem, and it is the cheapest thing on this list to fix once you can see it cleanly.

What the scorecard cannot do

Two limits, and the first one is the important one.

A scorecard sitting on an untrustworthy record produces confident wrong answers faster than a spreadsheet does. Every number above depends on the underlying entries being governed — time approved against the right engagement, WIP traceable to the work and the clause, forecast and actual drawn from the same object. If that is not true, measuring more frequently makes the disagreement more frequent, not smaller. This is downstream of a governed record, not a substitute for one.

And five is a deliberate under-count. A firm with three service lines and two billing models will need more than five, and some will need a different five. The discipline is not in the specific list. It is in refusing to add a sixth without deleting one, because a scorecard that grows every quarter is a report pack with better branding.

Where to start

Run the test before you change anything. Ask for the five, time it honestly, and note which ones required a person rather than a system. That list is your actual backlog, in priority order, and it is usually shorter than expected.

The reason DAY ONE can answer them from one place is that pipeline, delivery, time, cost, WIP and invoicing sit on the same governed record rather than being joined afterwards — which is also what makes the analysis on top of them worth trusting. How that reporting layer is grounded is set out on our analytics page.