How industrials companies are valued

Why EV/EBITDA is the default, and where it quietly lies to you

Every candidate can say EV/EBITDA is capital-structure neutral. Fewer can say why that matters more in industrials than almost anywhere.

The reason is that industrials comparable sets mix leverage widely. Inside one comp set you can have a family-controlled manufacturer that has run with almost no debt for decades, a sponsor-owned competitor that recapitalized itself twice, and a diversified conglomerate carrying pension obligations and a captive finance arm. Equity value multiples on that group tell you almost nothing. Enterprise value strips the financing decision out and lets you compare the operating business itself. EBITDA also neutralizes depreciation policy, and industrials companies make genuinely different choices about useful lives and about what counts as capitalized versus expensed maintenance, so two plants of similar age can carry very different charges on nothing more than accounting convention.

Now the nuance most candidates miss, and the one a good interviewer will hunt for. EBITDA flatters capital-intensive businesses precisely because it ignores the cost of keeping the asset base alive. A railroad or a foundry does not get to skip the money it spends replacing track and relining furnaces. Depreciation is an imperfect estimate, but in a heavy manufacturer it is a rough proxy for a real, recurring, unavoidable cost. An asset-light distributor has almost no such cost. Put both in the same comp set on EV/EBITDA and you have systematically made the asset-heavy company look cheap when the difference may be nothing more than capital intensity.

That is the moment to reach for EV/EBIT. By charging each company for the consumption of its own asset base, EBIT puts the heavy manufacturer and the light distributor closer to the same footing.

The correct interview posture is not to pick one metric and defend it to the death. It is to say which one you are leading with, name the distortion it carries, and say which second metric you would run to check yourself.

The mid-cycle problem

This is the heart of valuing anything in this sector, and where most candidates get taken apart.

Cyclical companies report peak earnings at the top of a cycle and trough earnings at the bottom. Everyone knows that. What catches people is that multiples move inversely to earnings across the cycle. At the top, the market can see that record margins are not sustainable, so it refuses to capitalize them at a full multiple and the company trades at its lowest multiple of the cycle. At the bottom, earnings collapse toward zero while the market prices in an eventual recovery, so the multiple looks enormous, or becomes meaningless entirely when EBITDA approaches nothing and the multiple approaches infinity.

Which means the classic error is a double count. A candidate takes the last reported year, which happens to be a peak year, applies the multiple the sector traded at in good times, and produces a valuation that assumes both peak earnings and a peak multiple simultaneously. That answer is wrong in the same direction twice.

The fix is to build a mid-cycle view. Three steps, in order.

First, normalize margins. Look across a full cycle rather than a favorable stretch of it, and ask what an average operating margin looks like for this business through the whole thing. Then adjust for anything structural that genuinely changed: a closed plant, a divested low-margin segment, a permanent shift in aftermarket mix. Structural improvement is real and you should keep it. Cyclical strength is not, and you should take it out.

Second, normalize volumes. Margin is only half of it. A machinery maker at the top of a replacement cycle is shipping units at a rate that reflects a fleet age bulge, not steady-state demand. Ask what a normal shipment level looks like, anchored on installed base and replacement rates rather than on the most recent order book.

Third, apply a through-cycle multiple rather than the multiple observable at whatever point in the cycle you happen to be standing.

Here is what that looks like in a hypothetical. Assume a machinery business posted $300M of EBITDA in a peak year, and your normalization work says mid-cycle EBITDA is $200M. Assume a through-cycle multiple for the peer set of 8x, a peak-period trading multiple of 6x, and a trough-period optical multiple of 13x.

Notice that the two internally consistent methods, rows two and three, land closer to each other than either does to the naive answer in row one. You may pair peak earnings with a compressed multiple, or normalized earnings with a normalized multiple, but never a peak number with a good-times multiple. The same logic runs in reverse at the bottom, where a trough year at a trough multiple buries a recovery the market is already paying for.

Reading where you actually sit in the cycle, from order rates and book-to-bill rather than guesswork, is covered in backlog, book-to-bill, and cyclicality. You cannot normalize a cycle you cannot locate yourself in.

What actually explains a multiple gap

Take two industrials companies with the same revenue and growth rate trading at meaningfully different multiples. An interviewer asking you to explain that gap is testing whether you understand business models or only spreadsheets. Growth is rarely the answer here. Quality of the revenue is.

The first two rows deserve emphasis because they drive the widest gaps in the sector. The economics of aftermarket content and why a large installed base changes the entire valuation conversation are worked through in aerospace and defense banking. The asset-heavy versus asset-light split, which shows up most starkly between an operator that owns the network and a broker that owns none of it, is covered in transportation and logistics.

Free cash flow conversion, the metric investors actually obsess over

Industrials investors talk about EBITDA in comps and about cash conversion in diligence. Conversion is how much of EBITDA survives the trip to free cash flow, and the bridge leaks more here than in most sectors.

Walk it in order. Start at EBITDA. Subtract capital expenditure, splitting maintenance from growth. Subtract the change in working capital. Subtract cash taxes, which can diverge widely from the book tax line when accelerated depreciation is in play. Subtract cash interest. Then subtract the restructuring and severance costs that get added back as one-time items year after year in a business perpetually rationalizing its footprint. Be skeptical of those add-backs out loud.

Working capital deserves its own attention, because it is where the sector is genuinely different. Industrials businesses hold real inventory and extend real receivables to distributors and to large customers who dictate payment terms, and long-cycle manufacturers carry partially built units for quarters before anything ships. The consequence is counterintuitive and worth saying explicitly in an interview: a strong growth year can consume cash. Volumes rise, inventory and receivables build ahead of collection, and free cash flow falls while EBITDA rises. That divergence is not a red flag by itself. Failing to notice it is.

The maintenance versus growth capex distinction matters most in a leveraged context. Maintenance capex keeps current capacity running and is not optional; growth capex funds new capacity and can be deferred. In an LBO the sponsor is underwriting whether cash flow after mandatory spending can service the debt, so the maintenance figure is what constrains the capital structure, and an aggressive assumption there is the fastest way to build a model that breaks in a downturn. Management teams rarely disclose the split cleanly, so you triangulate it against depreciation and against trough-year spending. The financing vocabulary around all of this sits in leveraged finance terms.

Enterprise value bridge landmines

Getting from equity value to enterprise value is supposedly mechanical. In industrials it is where careless work shows.

Each of these has a financing dimension as well as a valuation one, which is why a sponsor sizing debt against an industrial balance sheet treats them as claims competing with debt service rather than as footnotes.

Sum-of-the-parts for multi-segment companies

Industrials contains more genuine conglomerates than any other coverage group, and a single blended multiple on a company with an aerospace aftermarket business, a commodity components business, and a building products business is not analysis. It is an average of things that do not belong in the same average.

Sum-of-the-parts values each segment on its own comparable set, using that segment's own EBITDA or EBIT and the multiple its natural peer group commands, then adds the pieces. Two mechanical points decide whether it is credible. First, unallocated corporate overhead has to go somewhere: push it into the segments pro rata, or capitalize it at a multiple and carry it as a negative stub. Silently ignoring it inflates the total by the capitalized value of an expense the company genuinely bears. Second, be explicit about what a buyer of a segment actually takes on, including stranded costs and shared services that do not disappear when the segment leaves.

SOTP is also the analytical spine of the separation pitch. When the sum exceeds the trading value of the whole, the gap is the conglomerate discount, and quantifying it is what turns a strategic opinion into a board conversation. That argument is developed in full in industrials M&A and spin-offs.

The DCF in a cyclical business

The mechanics are the same here as anywhere, so the sector-specific question is narrower: what does the cycle do to your assumptions.

Two things. The terminal year must be a mid-cycle year. Terminal value carries most of the valuation, so growing a peak year into perpetuity capitalizes an unsustainable margin forever, and growing a trough year does the reverse. If your final projected year lands on a cyclical extreme, insert a normalization step so the perpetuity starts from a through-cycle margin and a through-cycle level of reinvestment.

Second, forecast horizons sometimes get extended, because a standard five-year projection may cover only part of a long capital equipment cycle. Running the explicit forecast longer, so it spans a full peak-to-trough-to-recovery arc, means terminal value starts from a genuinely normalized base rather than from wherever the fifth year happens to land. Say that out loud and you have separated yourself from every candidate reciting a generic DCF walkthrough.

Practice question

A cyclical industrial company just reported record EBITDA. How would you value it?

The first thing I would do is refuse to value it off that number. Record EBITDA usually means the company is at or near a cyclical peak, and peak margins are not the earnings base a buyer is actually acquiring. So I would build a mid-cycle view: look at operating margins across a full cycle, keep any structural improvement like a closed plant or a richer aftermarket mix, and strip out the purely cyclical strength. Then I would normalize volumes against installed base and replacement demand rather than the current order book.

The mistake I want to avoid is double counting the cycle. Cyclicals trade at their lowest multiples on peak earnings, because the market already knows those earnings will not last, so applying a full through-cycle multiple to a peak number gets the answer wrong in the same direction twice. If mid-cycle EBITDA is $200M against a reported $300M, an 8x through-cycle multiple on the normalized figure gives me a defensible anchor. I can cross-check by applying the compressed multiple the market is actually paying on those peak earnings to the reported number and seeing whether the two triangulate.

I would also look at EV/EBIT, since this is capital intensive and EBITDA ignores the cost of maintaining the asset base, and check free cash flow conversion after maintenance capex and working capital.

What the interviewer is listening for: Whether you recognize that multiples and earnings move inversely across a cycle, which is the specific trap the question is built around. They also want to hear you normalize both margin and volume rather than waving at the word normalized, and a candidate who volunteers the EV/EBIT and cash conversion cross-checks without being prompted is signaling they have actually thought about capital intensity.