Earned Value for Professional Services, Without the Ceremony
Spend against budget tells you nothing on its own. Half the money and a third of the work is a different situation from half the money and half the work.
Earned value management compares three quantities: what you planned to have spent by now, what you have actually spent, and what the work completed is worth against the plan. The third is the one most firms do not have, and it is the one that makes the other two mean anything.
Without it, a project reporting fifty per cent of budget consumed is unreadable. It could be exactly on track, or it could have consumed half the money for a third of the work. Those are opposite situations and the spend figure alone cannot distinguish them, which is why budget-versus-actual reporting reassures people right up until it does not.
The three numbers, in plain terms
Earned value is the one that requires judgement and therefore the one that gets fudged. If a deliverable was budgeted at twenty thousand and is genuinely half complete, its earned value is ten thousand regardless of whether it has consumed five thousand or thirty. Deciding what half complete means is the discipline, and the discipline is the entire value of the method.
- Planned value: what the plan said should have been delivered by this date, expressed in money.
- Actual cost: what has genuinely been spent, including effort at cost and third-party expenses.
- Earned value: what has actually been completed, valued at what the plan said it was worth. Not what it cost to do it, and not what it was billed for.
Two ratios worth reporting
From those three numbers come two indices, and they are more useful than the raw figures because they are comparable across engagements of different sizes.
Cost performance is earned value divided by actual cost. Above one means the work completed is worth more than it cost. Below one means the opposite, and the shortfall is the amount already lost on the work done so far.
Schedule performance is earned value divided by planned value. Below one means less has been completed than the plan expected by this date. It is a schedule measure expressed in money, which is unintuitive at first and useful in practice, because it makes schedule slippage comparable with cost overrun in the same unit.
The forecast that follows from these is the number a partner actually wants: estimate at completion. Dividing the total budget by the cost performance index gives a projection that assumes current efficiency continues, which is usually the honest assumption in the middle of an engagement and is a great deal better than restating the original budget with confidence.
Where it goes wrong in consulting
The method was built for construction and manufacturing, where percentage complete can be observed. On a consulting engagement it is estimated by the person doing the work, which introduces the two failure modes that give earned value its reputation.
The first is the ninety per cent complete problem. Knowledge work reaches an apparent ninety per cent quickly and stays there, because the remaining work is the hard part. A method that trusts self-reported completion will report an engagement as nearly finished for a third of its duration.
The defence is to stop estimating and start counting. Value is earned when a deliverable reaches a defined state, not when somebody feels it is nearly there. Nothing, or half on start and half on acceptance, or a fixed fraction per milestone: any rule is better than a judgement, because a rule cannot be optimistic.
The second failure is applying it at all where it does not fit. On a time and materials engagement with no fixed scope, earned value is close to meaningless because there is no baseline to earn against. On a fixed-price engagement with defined deliverables it is exactly the right instrument. Using it everywhere because it is in the methodology is how it acquires a reputation for ceremony.
The lighter version most firms should run
Full earned value with weekly recalculation is disproportionate for most engagements. A reduced version captures most of the benefit at a fraction of the cost.
Budget each deliverable in money at the point the plan is baselined. Earn value only on acceptance, so nothing is partially earned and the ninety per cent problem cannot occur. Compare earned against actual monthly. Report the estimate at completion whenever it moves by more than a threshold you set in advance.
That gives a firm the early warning that matters, which is the moment when cost performance drops below one and stays there, without a planner maintaining a model nobody reads.