Project Budgeting & Cost Control

The Report That Tells You Nothing

Every project I have inherited mid-flight arrived with a cost report, and almost every one of those reports answered a single question: how much have we spent? It is a comforting number because it is easy to produce - you copy it from the accounting system. It is also nearly useless, because spend on its own tells you nothing about whether the project is in trouble.

Consider two projects, each with a \$500,000 budget, each three months in, each having spent \$250,000 - exactly half. One is on track to finish on budget. The other is heading for a \$150,000 overrun. The spend report is identical for both. The difference is invisible until you ask the question that actually matters: not “how much have we spent?” but “how much have we accomplished for what we spent?”

That second question is the whole of cost control, and answering it well is the difference between project managers who deliver on budget and those who are perpetually surprised.

Spending Is Not Progress

The trap is treating money out the door as a proxy for work done. It is not. You can burn half your budget and have almost nothing to show for it - rework, false starts, an under-specified interface that had to be rebuilt twice. Conversely, a frugal team can be quietly ahead.

The fix is a concept that sounds abstract but is brutally practical: earned value. You earn value not by spending money but by completing work, measured at the rate you budgeted for it. If a \$100,000 work package is genuinely 40% complete, you have earned \$100,000 times 40%, or \$40,000 - no matter whether you spent \$30,000 or \$60,000 getting there. Lay that earned figure next to what you planned to have done and what you actually spent, and the fog clears instantly.

Take our two identical-looking projects. Both spent \$250,000. But Project A has earned \$250,000 of value (on track), while Project B has earned only \$190,000 - it spent \$250,000 to accomplish \$190,000 of work. That gap is the early warning the spend report could never show.

The Three Numbers That Cut Through

You need exactly three measurements, all in the same currency, all as of the same date. What you planned to have done (Planned Value). What you actually accomplished, valued at budget rates (Earned Value). And what you actually spent (Actual Cost). From these three, the entire diagnostic toolkit falls out with arithmetic any spreadsheet can do.

Divide earned value by actual cost and you get the Cost Performance Index. For Project B: \$190,000 / \$250,000 = 0.76. The team is getting 76 cents of value for every dollar it spends. That single ratio is the most predictive number in project controls, because of an uncomfortable empirical truth: once a project is past roughly the 15-20% mark, its cost performance index tends to settle and rarely improves on its own. A CPI of 0.76 at the quarter mark is not a rough patch you will grow out of. It is, in all probability, the rate at which the rest of the project will run.

Which means you can forecast the ending today. Divide the total budget by that index - \$500,000 / 0.76 - and you get a projected final cost around \$658,000. The \$150,000 overrun that was invisible in the spend report is now a number you can put in front of a sponsor in month three, with the maths to defend it.

Honesty Is a Tooling Problem

Why do so many overruns surface late? Rarely because the project manager is dishonest. Usually because the tooling encouraged the wrong question. If your only instrument measures spend, you genuinely cannot see the overrun coming - and by the time the spend curve crosses the budget line, most of your corrective options have expired. You cannot recover a cost performance index of 0.76 to 1.0 across the back half of a project; the required performance to claw it back is higher than any team running at 0.76 can plausibly deliver.

Earned value changes the conversation from confession to forecast. Instead of “we have a problem” in month six, you say in month three: “at our current performance we will finish about \$150,000 over; here is the root cause, here is what I am doing about it, and here is the one decision I need from you.” Sponsors can work with that. They cannot work with a surprise.

What To Actually Do

You do not need enterprise software to start. You need a baseline that is time-phased rather than a single lump, an honest assessment of percent-complete that resists the temptation to claim optimistically, and the willingness to compute three numbers every reporting cycle. Set thresholds so you investigate the variances that matter and ignore the noise. Re-forecast every cycle, not only when things look bad - a stable, on-track forecast is itself valuable information.

Above all, change the question you lead your status report with. Stop opening with how much you have spent. Open with where the project will land if nothing changes, and what you propose to do about it. The spend number belongs in the report somewhere, but it should never be the headline. The headline is the forecast - and the forecast is the job.