Construction Revenue

A construction company completes 60% of a highway project in the first year. The contract is worth INR 500 Cr. Under over-time recognition, the company reports INR 300 Cr of revenue this year. Under point-in-time recognition, it reports zero until it hands over the completed road, possibly two years later.

Same project. Same economics. Same cash outflows. But the P&L tells a fundamentally different story depending on which recognition method applies. For an FP&A team building quarterly forecasts and explaining variances to the board, this is not an academic distinction. It determines the shape of every revenue line in the model.

Over-Time versus Point-in-Time: The Core Test

ASC 606, IFRS 15, and Ind AS 115 all use the same framework. You recognise revenue over time if the contract meets any one of three criteria.

Criterion 1: The customer simultaneously receives and consumes the benefit. This applies to routine services (cleaning, security, ongoing maintenance) but rarely to construction. Nobody “consumes” a half-built bridge.

Criterion 2: The company’s performance creates or enhances an asset the customer controls. Most construction contracts satisfy this criterion. When you build on the customer’s land, using the customer’s permits, the asset belongs to the customer as the company constructs it. The customer controls the work-in-progress, so over-time recognition applies.

Criterion 3: The asset has no alternative use and the company has an enforceable right to payment for performance completed to date. This matters for build-to-spec manufacturing and bespoke construction where nobody can redirect the finished product to another customer. If the contract includes enforceable payment-for-work-done clauses (and Indian construction contracts often do), this criterion stands independently.

If none of these criteria apply, you default to point-in-time. That means zero revenue until the performance obligation is fully satisfied and control transfers. For a five-year infrastructure project, the difference between over-time and point-in-time recognition is the difference between a smooth revenue curve and a cliff: years of zero followed by a single massive recognition event.

Measuring Progress: Output versus Input

Once you establish that over-time recognition applies, the next question is how to measure progress. The standard offers two families of methods, and the choice is consequential.

Output methods measure progress based on what the team has delivered relative to what the contract promises. Physical completion surveys, milestones achieved, or units produced. The appeal is intuitive: if 60% of the bridge is built, you recognise 60% of the revenue. The challenge is measurement cost. Engineering surveys are expensive and infrequent, which creates lumpy recognition tied to survey dates rather than actual progress.

Input methods measure progress based on resources consumed relative to total expected resources. Cost-to-cost is the most common: costs incurred to date divided by total estimated costs. This is the successor to the old “percentage of completion” method, and most Indian construction companies default to it because cost data is available monthly from their ERP systems.

The risk with cost-to-cost is that it assumes the team incurs costs proportionally to progress. That assumption breaks in predictable ways. Early-stage material procurement distorts the ratio (buying all the steel in Month 1 does not mean you are 40% done). You need to exclude or adjust uninstalled materials unless they represent a distinct performance obligation. And any change in estimated total cost directly changes the cumulative revenue recognised to date, which creates the “catch-up” adjustments that make construction P&Ls so volatile.

The FP&A Problem: Forecasting Revenue That Moves Backward

This is the challenge I want to focus on because it is where accounting choices become operational headaches for finance teams.

Under the cost-to-cost method, revenue in any given quarter is a function of two variables: costs incurred this quarter and the total estimated cost to complete. The first variable is relatively predictable. The second is a judgment call that project managers revise continuously.

When the estimated total cost increases (scope changes, material price inflation, labour shortages), the percentage complete drops retroactively. The standard requires a cumulative catch-up: you recalculate total revenue recognised to date under the new estimate and book the difference in the current period. If that difference is negative, you have a quarter where revenue goes down even though the team did real work and incurred real costs.

For an FP&A team, this creates three specific problems.

Variance analysis becomes noisy. Your monthly actuals review shows a revenue miss on a project. Did work slow down, or did the project team revise the estimate-to-complete upward? Those are very different stories with very different implications, but they show up as the same number in the revenue line. Your variance bridge needs to separate “progress variance” (how much work got done relative to plan) from “estimate revision variance” (how the total cost assumption changed). Most construction company FP&A teams I have seen do not make this split, and their CFO narratives suffer for it.

Rolling forecasts require project-level granularity. You cannot forecast construction revenue at the segment level alone. Each project has its own cost curve, its own margin profile, and its own estimate-to-complete trajectory. Your rolling forecast needs a bottom-up build from project-level data, aggregated to segment and company. If your forecast model works at the segment level with average margins, it will miss consistently because the project mix shifts every quarter.

Stakeholder reporting needs a dual lens. Your board sees GAAP revenue. Your operations team sees physical progress. Your cash flow depends on billing milestones that may not align with either. A single project can show 70% revenue recognition (cost-to-cost), 55% physical completion (engineering survey), and 80% billing (milestone-based invoicing) all at the same time. As a Finance BP, your job is to present all three views and explain why they diverge, not to pick one and hope nobody asks about the others.

Contract Modifications: When the Scope Changes Mid-Project

Construction contracts change constantly. Variation orders, scope additions, design revisions, and client-directed changes are the norm, not the exception. The standard provides two treatments for modifications.

Treated as a separate contract: If the modification adds distinct goods or services at a price that reflects their standalone selling price, you account for it as a new, separate contract. The original contract continues unchanged. This rarely happens in construction because modification pricing is almost never at standalone rates.

Treated as part of the existing contract: If the modification is not a separate contract, you fold it into the original. You update the transaction price, update the total estimated cost, recalculate percentage complete, and book a cumulative catch-up adjustment. This is where a single variation order can move reported revenue by crores in the quarter it is approved.

The FP&A implication: your forecast needs a “modification pipeline” alongside the base contract forecast. If your project managers are negotiating a INR 20 Cr variation order that has an 80% probability of approval, that changes your revenue outlook materially. But you cannot recognise the revenue until the client approves the modification and meets the standard’s criteria. Your management commentary needs to flag the upside (or downside) from pending modifications separately from the base forecast.

Onerous Contracts: When the Whole Project Goes Negative

Ind AS 37 (and IAS 37) requires a provision when the unavoidable costs of meeting a contractual obligation exceed the expected economic benefits. In construction, this happens when total estimated costs exceed the contract price. You must recognise a provision for the expected loss immediately, not spread it over the remaining project life.

You continue to recognise revenue over time based on progress, but the loss provision accelerates the entire expected loss into the current period. The result is a quarter where you recognise revenue (because work progressed) and simultaneously book a loss provision (because the project economics deteriorated). Your FP&A commentary needs to separate these two effects clearly: how much revenue came from progress, how much loss provision was booked, and what the expected total outcome of the project is at completion.

Putting It Together: A Forecasting Checklist

Build the forecast bottom-up from individual projects. Each project carries its own contract value, estimated total cost, costs incurred to date, percentage complete, and margin. Aggregate from project to segment to company.

Separate the estimate-to-complete revision from the progress forecast. Your base case assumes the current estimate holds. Your sensitivity analysis shows what happens to revenue if total estimated costs increase by 5% or 10% on each material project.

Track pending modifications as contingent revenue alongside the base forecast. And monitor the gap between cost-to-cost progress, physical progress, and billing progress. When these three metrics diverge significantly on a project, it usually signals either an estimate-to-complete problem or a billing dispute, both early warning signs that deserve attention before they become quarterly surprises.

Where This Connects

The over-time versus point-in-time distinction is one application of Step 5 in the five-step model I covered in The Free iPhone Illusion. The same principle applies to SaaS contracts (where I explored upfront fees and subscription bundling), but construction introduces unique complexity through estimate revisions, modifications, and onerous provisions that other industries rarely encounter.

If your FP&A team is building a revenue forecast for a project-based business, or if you are working through the variance analysis challenges that come with cost-to-cost recognition, I would enjoy hearing how you approach it. Every construction company has a slightly different relationship between project controls, accounting, and FP&A, and the handoff points between those teams are where most forecast errors originate. Let’s connect.