Committed vs actual, not just actual
Most overruns hide in 'committed' spend — POs raised but not invoiced, contractors engaged but not yet billed, subscriptions started but not yet renewed. Track committed alongside actual and you see the overrun coming, not after it lands. This single change catches 70% of mid-project surprises in our experience. The committed line is the leading indicator; the actual line is the trailing indicator. Most spreadsheets only show the trailing indicator and the project manager finds out too late.
A 10-minute weekly burn review
Project sponsor plus project manager plus finance, 10 minutes a week, single agenda: 'are we on plan, on budget, on risk?' Three traffic lights, three numbers, written down. That's the meeting. Anything longer kills attendance, and the meeting is more valuable for its consistency than its depth. The weekly cadence catches issues at week two when they cost an hour to fix, rather than at week eight when they cost a fortnight.
Change requests with a budget delta
Scope changes without a budget delta are how projects overrun by stealth. Make every change request show its budget impact, however small. Five rejected £200 changes prevents the one approved £40,000 overrun, because the discipline of pricing the change is what surfaces the cumulative scope creep. Build the budget delta into the change-request template itself; do not rely on the project manager to remember to calculate it.
Burn-rate ratios, not just totals
A project that has spent 60% of its budget at the 40% mark of its timeline is in trouble, even if the total is still under cap. Burn-rate ratios (spend % vs time %) are the single most predictive metric for project completion at budget. Track them weekly, plot them against the original plan, and the divergence appears 3–6 weeks before the absolute overrun.
Contingency that means something
Most projects carry a 10–15% contingency line that gets drawn down silently by week six. The contingency is meant to absorb unforeseen events, not to backfill known underestimates. Require a named risk to draw down contingency, and the line lasts the whole project. The discipline is small but powerful — it converts contingency from a slush fund into a managed reserve.
Post-project review with numbers
Every project should produce a post-project review that compares forecast to actual on the major budget lines and identifies the two or three estimating errors that drove the variance. Done consistently across a portfolio, these reviews dramatically improve the estimating quality of the next round of projects. Done once and filed, they teach nobody anything.
Project discipline is three habits and a five-minute weekly conversation. The tools follow the habits, not the other way round — and the projects come in on budget as a side effect of the habits, not as a result of heroic effort at the end.