Pre-Engineered Buildings
The percentage margin on a PEB contract is close to meaningless, because it moves with the steel price rather than with performance. What actually matters is the absolute rupee spread earned on every tonne fabricated — a number that a well-run operator holds roughly steady whether steel is cheap or expensive.
Tonnage times realisation, less tonnage times cost, gives contribution per tonne. Apply operating leverage to that and you have the business — which is why utilisation, not the order book, is the number to watch.
What a PEB company actually sells.
The core product is fabricated steel, and the value added around it: design, engineering, cutting, welding, painting, and erection on site. A PEB company buys steel at one rupee-per-tonne figure and sells a fabricated, engineered, erected structure at another. Everything between the two — design capability, fabrication efficiency, erection execution, overhead absorption — is the business.
Against conventional cast-in-place construction, the pre-engineered route offers speed, high dimensional accuracy and labour savings. The relevant comparison for any specific project is against precast and cast-in-situ alternatives, and the extent of the advantage varies by building type rather than being a fixed premium.
Adjacent methods worth distinguishing
- PrecastComponents cast off-site and assembled. Faster and often cheaper than cast-in-situ, though the differential varies by project.
- Cast-in-situConventional on-site casting. Slower, more labour-intensive, more weather-exposed.
- PPVCPrefabricated volumetric construction — complete modules built off-site. The most industrialised end of the spectrum.
Contract length is a risk control. PEB contracts typically carry price-variation clauses linked to steel indices, and projects usually run three to nine months. That short cycle is itself a hedge: less time between pricing and delivery means less exposure to a steel move. It is one reason percentage margins fluctuate while rupee spreads hold.
Customised tooling has a lead time. Creating a customised steel mould typically takes four to six weeks, which sits inside any build-to-order schedule and should be reflected in delivery commitments.
Per-tonne contribution, and the leverage on top of it.
A worked example
Other variable and allocated fixed cost
(labour, power, consumables, design, overhead): ₹15,000 / tonne
Total cost: ₹75,000 / tonne
Customer billing: ₹85,000 / tonne
Per-tonne delta: ₹10,000
Margin on revenue: ≈ 11.8%
Now let steel spike or collapse. If the operator reprices to preserve roughly ₹10,000 per tonne, the percentage margin changes materially while the economics of the business do not. Judging this sector on margin percentage will therefore produce the wrong conclusion at both ends of the steel cycle.
The revenue bridge is simple and worth building explicitly. Tonnage sold multiplied by realisation per tonne gives revenue; tonnage multiplied by cost per tonne gives cost of goods; the difference is contribution per tonne. Fixed overheads are then applied to arrive at EBITDA and EBIT. Every question worth asking in this sector attaches to one of those four terms.
Fixed costs are high, so operating leverage is pronounced. At low utilisation, EBITDA per tonne shrinks quickly because the fixed base is spread across too few tonnes. Above the threshold, margins improve sharply. Utilisation is therefore the single most important operating metric — more important than the order book, which tells you about future tonnage but nothing about whether the plant is busy now.
Order book quality varies more than order book size. Public sector work and private work carry different payment behaviour, different retention terms and different execution risk. Within private work, the end sector matters — warehousing, manufacturing and infrastructure each behave differently through a cycle.
Spread protection, utilisation, and execution.
Holding the per-tonne spread
Pricing discipline and index-linked variation clauses that genuinely reprice through a steel cycle. The operator that keeps its rupee delta intact when steel doubles is the one worth owning.
Utilisation
With a high fixed-cost base, EBITDA per tonne is a function of how full the fabrication capacity is. Filling the plant matters more than winning the largest possible order book.
Execution on schedule
Speed is the product's core claim against conventional construction. Delivery delays undermine the value proposition and tie up capital in incomplete projects.
What to answer before underwriting.
- →Contribution per tonne. The actual rupee delta, and how it has moved over the last several steel cycles. This is the business in one number.
- →Realisation per tonne. Current level, and the specific strategy to improve it — higher-value structures, more design content, better mix.
- →Price variation clauses. Are contracts genuinely index-linked, and how quickly does a steel move reprice in practice rather than in theory?
- →Fabrication capacity and utilisation. Tonnes per year installed, tonnes actually produced, and the utilisation level at which margins inflect.
- →Order book quality. Public sector versus private, and within private, the end sector. Split by value and by tonnage.
- →Project cycle length. Typical duration, and whether it is genuinely in the three-to-nine month band that limits steel exposure.
- →Delivery record. Average time for a build-to-order project of a given size, including lead time and final inspection. Any pattern of delays.
- →Tooling lead time. Time to create a customised steel mould, and whether that sits inside quoted schedules.
- →Erection capability. In-house or subcontracted, and what that means for margin capture and schedule control.
- →Receivables and retention. Payment terms by customer type, and retention money held against completion.
What to monitor, quarter by quarter.
| KPI | Calculation | Benchmark or read-through |
|---|---|---|
| Contribution per tonne | Realisation/tonne − cost/tonne | The core metric. Should hold roughly steady across steel cycles |
| Realisation per tonne | Revenue ÷ tonnes sold | Rising realisation with a stable delta means mix is improving |
| Cost per tonne | COGS ÷ tonnes sold | Split steel from conversion cost — they move independently |
| Tonnage sold | Tonnes dispatched per period | Volume half of the revenue bridge |
| Capacity utilisation | Tonnes produced ÷ tonnes/year capacity | Drives EBITDA per tonne through operating leverage |
| EBITDA per tonne | EBITDA ÷ tonnes sold | Contribution per tonne net of fixed cost absorption |
| Order book (tonnage) | Contracted tonnes, not just value | Value moves with steel; tonnage is the real volume signal |
| Order book mix | PSU vs private, by end sector | Determines payment behaviour and execution risk |
| Execution cycle time | Days from order to handover | 3–9 months typical; slippage erodes the speed proposition |
| Receivable days | Debtors ÷ revenue × 365 | Public sector work stretches this materially |
| Retention money | Held against completion | Locked-up capital that does not show in headline receivables |
| Steel index vs realisation | Track the two together | Tests whether variation clauses are working as claimed |
How the thesis breaks.
- !Judging the business on margin percentage. The most common analytical error here. A falling percentage during a steel spike can accompany an unchanged or improving rupee spread.
- !Utilisation shortfall. High fixed costs mean a volume miss compresses EBITDA per tonne quickly, with nothing to offset it.
- !Variation clauses that do not work in practice. Index linkage on paper is not the same as timely repricing with a customer who resists it.
- !Project delays. The proposition is speed. Slippage damages both the commercial argument and working capital.
- !Public sector receivables. Government-weighted order books stretch payment cycles and lock capital in retention money.
- !Capacity added into a soft cycle. Fabrication capacity resets the fixed base immediately; demand for industrial and warehousing construction is cyclical.
- !Order book value flattered by steel. A backlog quoted in rupees rises when steel rises without any increase in real work. Ask for tonnage.
The figures, and where they stand.
| Metric | Value | Note | Basis |
|---|---|---|---|
| Illustrative steel cost | ₹60,000 / tonne | Worked example — the dominant input | Illustrative |
| Illustrative conversion cost | ₹15,000 / tonne | Labour, power, consumables, design, overhead | Illustrative |
| Illustrative billing | ₹85,000 / tonne | Total cost ₹75,000 / tonne | Illustrative |
| Per-tonne delta | ₹10,000 | ≈ 11.8% on revenue — the percentage moves, the delta shouldn't | Illustrative |
| Project cycle | 3–9 months | Short cycle limits steel price exposure | Research note |
| Customised mould lead time | 4–6 weeks | Sits inside build-to-order schedules | Research note |