EPC and contracting: order book, unbilled revenue, and percentage-of-completion as a lever

An engineering-and-construction contractor reports steady, healthy revenue growth and improving profit. Its receivables sit at 140 days - a number that, in almost any other business, would scream that customers are not paying. And its unbilled revenue is growing three times faster than its billed revenue, quietly, in a line most readers never open. One of these facts is completely normal and the other is a warning, and telling them apart is the whole skill of reading a contractor. illustrative

An EPC company - engineering, procurement, construction: a contractor paid to build an asset for its owner - builds large, long projects like a power plant, a metro line or a highway over several years, and two facts shape its accounts. First, the customer withholds a slice of every bill as retention money until the project is finished and proven defect-free, so a contractor is always owed a large sum that is not overdue but simply held back by contract - which makes 140 receivable days structural, not a collection problem. Second, because a project spans years, the contractor cannot wait until completion to book revenue; it recognises revenue over time, in proportion to how complete each project is judged to be. And "judged to be" is the crack in the floor: the percentage of completion is an estimate, and an estimate is a lever.

So this module reads a contractor through three numbers the ordinary reader skips - the order book, which is the visible future work and the true leading indicator; the retention money buried in receivables, which explains why the days are high without alarm; and unbilled revenue, which is where the percentage-of-completion lever leaves its first trace.

The previous module read a developer, whose revenue lands in one lump at completion. A contractor is the smoother twin - it recognises revenue over time as the project progresses, which avoids the developer's lumps but introduces an estimate the developer never has to make. That estimate is both what makes a contractor's accounts readable and the lever by which they can be gamed. Four ideas carry the reading.

The order book is the value of contracted work not yet executed - the leading indicator of future revenue, read against annual revenue as book-to-bill (orders in hand divided by one year's revenue, the years of work already booked). Retention money is the portion of each bill the customer withholds until completion, which sits in receivables and makes a contractor's receivable days structurally high without signalling a collection problem. Unbilled revenue is revenue recognised on the completion estimate but not yet billed to the customer. And the percentage-of-completion lever is the judgement behind it all: by estimating a project as more complete than it is, a contractor recognises revenue and profit ahead of the actual work, flattering the current year at the expense of a later reckoning.

The order book leads revenue. A contractor's revenue this year comes from executing work it won in prior years, so the order book is the leading indicator. An order book 2.9 times revenue means nearly three years of visible work - but the book's trend matters more than its level. Order inflow above revenue (book-to-bill rising) means the pipeline is filling faster than it empties, so future revenue is secured; inflow below revenue means the contractor is living off its backlog, and today's revenue growth will reverse as the book runs down.

Retention makes receivables look worse than they are. On a construction contract the customer typically withholds 5-10% of each bill as retention, releasing it only when the project completes and the defect-liability period - the warranty window after handover during which the builder must fix any faults - passes, often a year or more later. That retention sits in receivables the whole time, so receivable days of 130-150 are structural, not a sign customers are not paying. The check is to strip the retention out: receivables net of retention should collect on ordinary terms, and it is a rise in the non-retention receivables, or in retention that never gets released, that signals a real collection problem.

Unbilled revenue is where the lever shows. Because revenue is recognised on the percentage of completion, a contractor books revenue as work progresses and bills the customer separately as milestones are certified. The two rarely match exactly, and the gap is unbilled revenue - work recognised as revenue but not yet billed. A modest, stable balance is normal. But unbilled revenue growing much faster than billed is the tell that the contractor may be recognising revenue on an aggressive completion estimate the customer has not certified, or booking disputed claims still under arbitration as if they were certain. Estimating a project 70% complete rather than 60% recognises another slice of revenue and profit now, with no change in the actual work done - the same estimate lever seen in depreciation and provisioning, applied to the top line. And it reverses eventually, because a project can only be 100% complete once, so revenue pulled forward is revenue borrowed from a later year. A contractor whose margins look strong while unbilled revenue balloons is a contractor to distrust: the profit may be an estimate, and the cash - held back further by retention - is the slow truth that catches up.

The one inversion worth holding is that a contractor's 140 receivable days are the design of the business, while the same 140 days in FMCG are a warning. Retention has no equivalent in fast-moving consumer goods, where the trade pays in two to three weeks, so the identical figure reads opposite ways. Import the FMCG rulebook and you reject a healthy contractor for high receivables; import the contractor's rulebook and you excuse genuine channel stuffing - loading distributors with more stock than they can sell, to inflate reported sales - in FMCG. The number is the same; the meaning is set entirely by whether retention explains it.

This is the sharpest reason in the sector to read a ratio against the norms of its own industry: the receivable-days line, so revealing almost everywhere, means nothing for a contractor until it is decomposed into retention and ordinary trade credit.

Read the composite EPC contractor from the order book inward. Its order book grew from ₹14,000 crore to ₹26,000 crore over five years, against revenue of ₹8,900 crore in the final year - a book-to-bill of about 2.9 times, so nearly three years of work is visible. Order inflow in that year was ₹13,400 crore, comfortably above revenue, so the pipeline is filling faster than it empties. On the leading indicator this is a healthy, growing contractor with secured future revenue - a far more useful read than the revenue line alone. illustrative

Now the receivables. They sit around 140 days throughout, which an FMCG reader would find alarming. But retention money - ₹1,360 crore in the final year, out of ₹3,420 crore of receivables - explains much of it: nearly 40% of the receivables is cash the customers are contractually holding back until projects complete, not money they are refusing to pay. Strip the retention out and the ordinary trade receivables collect on normal terms. The 140 days is structural, and reading it as a collection problem would be a category error.

Then the line that matters most: unbilled revenue. It grew steadily for three years in line with billing, then in the fourth year jumped 35% while billed revenue grew only 11%. That gap is the flag. In that year the contractor recognised revenue well ahead of what it billed - which means either a genuine timing effect, work done late in the year and billable next year, or the percentage-of-completion lever being pulled, revenue booked on an optimistic completion estimate. It is not proof of manipulation, but it is exactly where you would look, and you would want the next year's billing to catch up and the cash to follow. A contractor whose unbilled revenue keeps outrunning its billing, year after year, is recognising profit its cash never confirms. The revenue and profit are real only to the extent the billing and the cash confirm them.

When unbilled revenue grows fast, a sharp analyst asks whether it is timing or the lever: "Unbilled revenue grew 35% against 11% billed. How much is genuine year-end timing that bills next quarter, how much is claims under arbitration, and what's your expected conversion to billing and cash over the next two quarters?" The question is testing whether the recognised revenue is real.

A good answer splits the growth and dates the conversion:

"Fair to probe. Of the increase, about two-thirds is timing - three large projects hit milestones in March that we billed in April, so it's already converting this quarter. The remaining third is a claim on a delayed government project, which we've recognised conservatively at 60% of the claimed value based on counsel's advice, and it's in arbitration with a hearing scheduled. We'd expect unbilled to normalise back toward one-times-billed growth by year-end. Retention is steady at about 40% of receivables and releasing on schedule."

It separates timing from claims, states how conservatively the claim is booked, gives a conversion expectation, and confirms retention is releasing. It lets you judge whether the revenue is real. The evasive answer does none of that:

"Unbilled revenue reflects the normal course of our project execution and our revenue recognition is fully compliant with accounting standards and reviewed by our auditors. We're confident in our order book and our execution capabilities, and we see strong growth momentum across our verticals. There's nothing unusual in our working capital."

Fluent and empty - it never splits timing from claims, never mentions the arbitration exposure, and gives no conversion timeline. "Compliant with accounting standards" confirms the recognition is permitted, not that it is conservative, and "nothing unusual" is the exact reassurance a contractor pulling the completion lever would also offer.

The follow-up nobody asks is "How much of unbilled revenue is claims under arbitration, and what has your unbilled-to-billing conversion actually been over the last four quarters?" - which forces the disputed portion and the conversion track record into the open. If "normal course of execution, fully compliant" is allowed to stand, an investor credits revenue recognised on an aggressive estimate and disputed claims as if it were cash. The silence is the tell: either the claims are a large, uncertain chunk of the recognised revenue, or the conversion history would show unbilled that never turns into cash.

The order book tells you the volume of future work, not its profitability. A contractor can win a fat book by bidding aggressively - taking projects at thin or loss-making margins to keep its people and equipment busy - so a growing order book can be a warning as much as a comfort if it was bought with price. Book-to-bill measures how much work is secured; it says nothing about whether that work will execute at a decent margin, which depends on bidding discipline, input-cost assumptions, and the risk of cost overruns on fixed-price contracts. A fat order book at bad prices is future revenue that destroys value.

Nor does unbilled revenue, on its own, prove the completion estimate honest or dishonest. A widening unbilled-to-billed gap is a flag that can be genuine timing or aggressive recognition, and the line alone cannot distinguish them - only the following periods resolve it. If the billing catches up and the cash arrives, it was timing; if the unbilled balance keeps growing and never converts, it was the lever. And the accounts cannot fully reveal the execution and counterparty risks that decide a contractor's fate: land-acquisition delays, a customer who stops paying, a design change that triggers a dispute, a cost overrun that turns a profitable contract into a loss - much of it surfacing as claims under arbitration, contingent liabilities and vague "expected to be resolved" notes rather than clean numbers. A contractor booking disputed claims as receivable is recognising cash it may never see, and whether those claims are real is a legal judgement the statements cannot make for you.

Which is where readers get fooled, in three predictable ways. They read a contractor's receivable days with an ordinary rulebook and reject a healthy business for the retention that is structural in construction. They admire revenue growth without reading the order book, missing that a contractor can grow revenue for a year or two while its pipeline empties, so the revenue falls when the backlog runs out. And they never open the unbilled-revenue line - the single clearest trace of the completion lever - and so trust an estimate dressed as a result, right up until the unbilled balance has to convert and does not. The whole reading is a habit of not letting any one number stand on its own: decompose the receivables, read the order book for durability, and watch unbilled against billed for the lever.