Revenue Recognition in Services, explained

A services firm can send every invoice on time, collect every payment on time, and still not know how much revenue it has actually earned. Invoicing, payment and earning are three different events, and in a services business they can sit months apart. Revenue recognition is the discipline that answers the question underneath all three: how much of the promised work has genuinely been delivered, and therefore how much of the agreed price has genuinely been earned? One note before anything else: this entry explains the concept, its intuition and its failure modes. It is not accounting advice; the rules that govern recognition live in accounting standards and in conversations with your auditors, and they vary by contract type and jurisdiction.

Earned, invoiced, paid: three different clocks

Under the recognition view, revenue is recorded as the work is delivered, not when the invoice is raised and not when the cash lands. The three can run far apart in either direction. Invoice ahead of the work, as retainers and milestone schedules often do, and the firm holds deferred revenue: cash received for work still owed. Deliver ahead of the invoicing and it holds unbilled revenue: work performed and earned but not yet asked for. Neither position is wrong. What matters is knowing, engagement by engagement, which clock each number is running on, because a comfortable cash position can coexist with a large stock of work still owed, and a thin one with a large stock of work already earned.

The percentage-of-completion intuition

For a fixed-price engagement delivered over months, the intuition is natural: if the engagement is some fraction complete, then roughly that fraction of the price has been earned. Take a clearly illustrative case with invented round numbers. An engagement is priced at 120 and expected to take a year of even effort; after four months of on-plan delivery, roughly a third of the work is done, so roughly 40 has been earned, whatever the invoicing schedule says. The intuition is simple. Everything difficult hides inside the phrase “a third of the work is done”, because completion is not observed, it is estimated: usually from cost incurred against cost expected, or milestones passed against milestones planned. Both are judgements about the future dressed as measurements of the present. An engagement that has spent a third of its budget has not necessarily done a third of the work. It has spent a third of the money.

How recognition misleads

Because completion is an estimate, recognised revenue inherits every bias in the estimate, and the pressures all lean the same way. Three failure modes recur.

• Effort read as progress. When completion is measured by cost burnt against cost planned, a struggling engagement looks more complete than it is: burning money fast reads as delivering fast. The engagements most likely to be overstated are precisely the ones already in trouble.
• Optimism banked as income. A generous completion estimate books revenue that delivery has not yet backed. Each optimistic quarter borrows from future ones, and the borrowing stays invisible until the estimate finally meets the work.
• The restatement path. Aggressive recognition converts a delivery problem into a reporting problem. If the work is genuinely behind, revenue recognised against imagined progress must eventually be unwound: margin collapses late, prior figures are corrected, and a delivery issue that could have been managed quietly becomes a credibility issue that cannot.

Revenue recognition is where the delivery truth and the financial truth of a services firm are forced to agree, and the estimate of progress is the hinge between them. A firm with honest progress estimates has a revenue line that means something. A firm with optimistic ones has a number that holds right up until it does not.

One concrete example

Clearly illustrative, with invented round numbers and no customer implied. A fixed-price engagement is sold at 100 over ten months, with an expected delivery cost of 70. By month five the team has spent 40 of the 70, so a cost-based estimate says the work is more than half complete, and more than half the revenue is recognised. But the early months were consumed by technical problems that produced effort without output, and the scope genuinely delivered is nearer two fifths. The difference between the recognised fraction and the real one is not revenue; it is optimism recorded as income. If delivery recovers, nobody ever notices. If it does not, the gap surfaces all at once in a late quarter: recognised revenue reverses, margin collapses at closeout, and everyone asks how a profitable engagement lost money in its final months. It did not. It was behind all along, and the completion estimate was where that fact went unrecorded.

The decision-intelligence reading

The useful question about recognised revenue is the question worth asking of any conclusion: what evidence is it standing on? A completion estimate built from deliverables formally accepted by the client is close to measured. One built from a cost-burn ratio is modelled, with known biases. One built from a delivery lead’s assurance that the team will catch up is stated. Keeping those grades distinct is the whole point of the evidence hierarchy, and the decision-intelligence reading of revenue recognition follows directly: a revenue figure is only as strong as the evidence under its progress estimate. It is also why recognition belongs alongside delivery confidence: falling confidence that an engagement will complete as planned is early warning that revenue recognised against the plan is at risk. The decision this metric should trigger is never “restate later”. It is “interrogate the estimate now”, while rescoping, restaffing or renegotiating are still available, and before a delivery problem is converted, at the worst possible moment, into a financial one. The same discipline protects delivery margin, the other place where the sold story and the delivered story are forced to meet.