Why Projects Go Over Budget, and the Six Controls That Stop It

An overrun almost never arrives as an event. There is no day on which a project becomes expensive.

What happens instead is that a decision is taken in month four, and its cost becomes visible in month fourteen, at which point it is attributed to month fourteen. The client sees the bill and asks what went wrong recently. Nothing went wrong recently. The bill is a delayed report on something that went wrong much earlier.

Here are the six mechanisms, and what actually works against each.

1. The incomplete drawing set

The most common cause, by a distance.

A contractor prices what is drawn. Anything not drawn is either omitted from the price or assumed at the cheapest reasonable interpretation. When the gap is closed later, and it will be, it is closed as a variation, priced without competition, on a site where the client's alternative is stopping work.

The control: do not tender an incomplete set. Schedules issued, details drawn, specifications written, TBCs resolved. This delays tender by weeks and it is the single highest-return delay available on any project. A contractor cannot change-order what has already been priced.

2. Uncoordinated drawings

Structure, architecture and services each correct, and mutually contradictory. A duct that cannot pass a beam. A drain that must cross a foundation. A ceiling too shallow for what is above it.

Each is discovered on site, resolved in an afternoon by whoever is present, and paid for twice, once in the rework, once in the compromise.

The control: a formal coordination review before construction, with a named owner and a signed-off clash list. On any project of scale, a federated model rather than a set of overlaid PDFs. The cost is a fortnight of a coordinator's time.

3. Client-side decision latency

A decision requested in week six and given in week twelve does not cost six weeks. It costs six weeks plus everything that could not proceed, plus the resequencing, plus the standing time, plus the second mobilisation of a trade that had to be released.

This is the overrun cause clients are least willing to look at, because the invoice does not say "delay caused by client."

The control: a decision register with dates and named owners, reviewed weekly, with the cost consequence of each pending decision stated in rupees next to it. Not as an accusation. Because a decision with a visible price attached gets made faster.

4. Scope creep by accretion

Never one large change. A better handle. A different tile in the lobby. One extra socket per room. An upgraded fitting in the bathrooms.

Each is individually trivial, individually justified, and individually approved without anyone summing them. Fifty such decisions across a project is a substantial number that no one ever decided to spend.

The control: a running variation log, with a cumulative total, reported at every client meeting. Not to prevent changes, changes are legitimate, but to ensure the sum is visible at the moment the next one is being approved.

5. Escalation and long-lead exposure

A rate quoted in January against a build running twenty months is a rate for a market that no longer exists. Materials in NCR have moved 8–12% in a year. A long-lead item ordered late is ordered at the price of the day it was ordered, not the day it was priced.

The control: an escalation clause both parties understand, procurement of major materials early against a frozen design, and a contingency that is genuinely held rather than spent in month three because the early works came in under budget. A contingency spent early is not a contingency.

6. Certification without independence

If the party doing the work is also the party measuring the work, cost control is a formality with a signature on it.

The control: independent measurement and certification. Bills checked against a bill of quantities by someone with no interest in the outcome. This is what quantity surveying and bill audit exist for, and it is the control most often removed to save fees, a fee saving that is almost always smaller than what it was protecting.

What all six have in common

Every one of them is a gap between a decision and its visible consequence. The overrun is not the decision. The overrun is the delay in seeing it.

Which means the whole discipline of cost control reduces to one thing: shortening the distance between a decision and its price. A drawing set that leaves nothing to be decided later. A coordination review that finds the clash before the concrete. A decision register that puts a number next to the delay. A variation log that shows the running total. A procurement plan that fixes the price early. A certifier who has no stake in the answer.

None of it is sophisticated. All of it is boring, and it is boring in a way that has to be maintained weekly by someone whose job it is.

That role has a name, and the projects that do not have someone in it are the projects that discover their real cost at the end.