Eighty-nine suppliers, five worth approaching

Ninety days to decide what to buy

Sawmilling is an old industry with new machines in it. A log arrives, it is measured, debarked, broken down, edged, trimmed, sorted, dried, graded and stacked. Every one of those steps is a machine, and most of those machines are bought from a specialist supplier rather than built by the mill.

Two constraints shaped everything. The first was that the market number had to be defensible to a board, which meant it could not be a top-down percentage of an industry figure. The second was that most of the candidates were private, which meant the evaluation had to work on partial financial information without pretending otherwise.

Our client, a global industrial equipment group, wanted to grow its position in that supply market and had decided the fastest route was to buy rather than build. The question was what to buy. Nobody inside the company could say with confidence how large the equipment market actually was, and the supplier landscape was a long tail of private, family-held engineering firms with no published accounts and no analyst coverage.

There was a board strategy session ten weeks out. The ask was a shortlist with reasoning attached, not a market report.

Output is public, spend is not

Sawnwood production is one of the best documented industrial statistics in the world. The Food and Agriculture Organization has published national production volumes for decades. You can pull a clean series for every producing country back to the 1960s in an afternoon.

The number you can find is not the number you need, and the gap between them is the whole job.

None of that tells you what the equipment market is worth. Production volume measures what came out of the mills last year. Equipment spending is driven by something else entirely: how much capacity exists, how old it is, and where in its replacement cycle it sits. A region can produce less every year for a decade and still be the largest equipment market on earth, because it has more mills, they are older, and they are being rebuilt.

On the supplier side the problem is the opposite. There is no shortage of information about what these companies make, because they publish it for customers. There is very little about what they earn. Of the eighty-nine suppliers we eventually catalogued, the large majority were privately held, several sat inside larger groups that do not report the segment, and a few had ownership structures that had changed twice in five years.

Production volume is the wrong denominator

The first instinct in a market like this is to take production volume and apply a percentage. It is fast, it is defensible-sounding, and it is wrong, because the two move independently.

Output tells you how the industry is feeling. Capacity, and its age, tell you what it is about to buy.

Read the chart as a demand signal and it says the industry shrank. Global output was 482 million cubic metres in 2020 and 445 in 2024, a fall of about 8% in four years. Energy costs rose, new-build housing weakened, log supply tightened, and China’s housing correction pulled the largest region down with it.

Now read it as an equipment signal. It says almost nothing. A mill that runs at 60% of capacity still replaces its blades on the same schedule, still takes its annual outage, and still has a primary breakdown line that is fifteen years old and due. Falling output can even accelerate equipment spending, because the mills that survive a downturn are the ones that modernise, and the ones that do not modernise are the ones that close.

The projection matters more than the history for an acquirer, and it is modest: recovery to 459 million cubic metres by 2030 and 488 by 2035, which only returns the industry to roughly where it was in 2020. Anyone buying into this market on a growth story is buying the wrong story. The case has to be made on share, on aftermarket, and on replacement.

Size the market at the machine, not the industry

So we built it from the bottom. One mill, fully specified, then multiplied.

Both of these are identified market size. They describe the spend that exists, not revenue anyone has won, and not a forecast of what an acquisition would capture. The gap between the two numbers, roughly €2 billion, is not lost opportunity either. It is Asia Pacific and Latin America, where the company had no service presence, and where an acquisition would have to bring one.

The move that makes this work is annualising. A sawmill does not buy a debarking and sawing line every year; it buys one every fifteen years or so. Spreading that purchase across the cycle converts a lumpy, unpredictable event into a rate per cubic metre of capacity, and a rate can be applied to any region for which capacity is known.

The second move is narrowing the scope before sizing rather than after. The client could supply debarking and sawing. It could not supply the buildings, the kilns or the energy plant. Sizing the whole mill and then taking a share of it would have produced a larger number and a worse decision.

The total addressable market came out at €5.7 billion a year in 2024, projected at €6.0 billion by 2030. The serviceable market, meaning the regions where this company actually had people who could sell and service, was €3.7 billion, projected at €3.8 billion.

The served half is where the old machines are

Splitting the serviceable market by region produced the finding that changed the shortlist.

One market buys replacement. The other buys capacity. A supplier that is excellent at one is not automatically useful in the other.

Europe and North America are close in production volume, 118 against 110 million cubic metres, and close enough in spend, €2.2 billion against €1.4 billion. On every other measure they are different industries.

Europe has 2,707 mills against North America’s 886. It has 189 million cubic metres of installed capacity but produced 118, which is 62% utilisation. And its average mill is 59 years old.

North America has fewer, larger, younger mills running at 92% of capacity, with an average age of 37. Its capacity is rising while its mill count stays flat, because large operators are closing several small mills to open one large one. Capacity is also moving, out of western Canada and into the US South where the timber is.

Europe is a modernisation market: a large, old, underused fleet where the purchase is a line rebuild in an existing building, sold to an operator deciding whether to reinvest or close. Between 2020 and 2025 European mill openings added about 4.0 million cubic metres of capacity while closures removed only about 0.8 million, so the fleet is consolidating rather than shrinking. Operators closing mills said so plainly: high log costs, low lumber prices, weak construction demand, and an investment need they were not willing to fund on an old asset.

North America is a project market: fewer, larger decisions, greenfield or major reconfiguration, sold to a corporate capital committee. The equipment is the same. Almost nothing else about the sale is.

Aftermarket is not a footnote, it is 43% of the spend

Of the €3.7 billion serviceable market, €2.1 billion was capital equipment and €1.6 billion was service and spare parts. Aftermarket is 43% of the money, and it behaves nothing like the other 57%.

What this changes in practice. It moves installed base from a nice-to-have into a scored criterion, and it makes the question during diligence not “how good is the technology” but “how many of these machines are running, where, and who services them”.

Capital spend is cyclical, competitive and won on price and specification. Aftermarket spend is annuity-like, largely uncontested once the machine is installed, and tied to the installed base rather than to this year’s order book. In an old fleet running below capacity, the aftermarket half is also the more reliable half.

That is why the evaluation scorecard gives 15% of its weight to the share of a supplier’s revenue that comes from service and spare parts, with the top band set at 40% or more. A supplier with a large installed base and a high aftermarket share is buying you an annuity and a customer list. A supplier with strong engineering and 5% aftermarket revenue is buying you a project business, which is a different asset at a different multiple.

Only about half of a sawmill is the sawmill

One number kept the scope honest.

For a mid-scale mill of 150,000 cubic metres a year, costing around €45 million to build, the process line and its kilns are 49.6% of the capital cost. Buildings, infrastructure, installation and the energy plant are 49.1%. The rest is rounding.

This is why the sizing excluded everything outside debarking and sawing, and why it should have. Had we sized the mill and taken a share, the market would have looked roughly twice as large and the shortlist would have been drawn from a different, wider pool of suppliers, most of whom build things this client had no intention of building.

Scale changes the arithmetic but not the lesson. A small commercial mill of 50,000 to 150,000 cubic metres costs €5m to €15m. A mid-scale industrial mill costs €20m to €60m. A large supermill above 400,000 cubic metres costs €80m to more than €200m, and comes with heavy scanning, a large kiln battery, an integrated energy plant and full digitalisation. The supermill is where the industry is heading, in both served regions, for the same reason: the same capacity at a lower operating cost.

Eighty-nine suppliers, and the eight that survived

With the market understood, the supplier work became tractable. The long list came from desk research, industry databases, trade exhibitions and the client’s own internal list, which is worth including because a sales organisation has usually met most of the market and never written it down.

The first filter is deliberately crude: a revenue floor of roughly €10 million, solid EBIT, and relevance to the equipment in scope. It exists to stop the expensive work being spent on companies that cannot pass, and it removed 71 of the 89.

The second filter is the scorecard, and the design choices in it are the transferable part.

Strategic fit and revenue scale carry the same weight. Both at 20%, more than any other criterion. A perfect strategic fit at €12 million of revenue does not move the needle for a group of this size, and a large supplier in the wrong part of the line is an expensive distraction. Neither survives alone.

Aftermarket share is weighted as heavily as profitability. Both at 15%, for the reason set out above.

Brand and ownership carry only 5% each, and ownership is the one that actually kills deals. It is scored low because it is close to binary, and a binary criterion distorts a weighted average. In practice a supplier whose owners have no intention of selling scores four points out of a hundred and is worth zero, so ownership was handled as a gate in the conversation rather than as weight in the model. That is a limitation of scorecards generally, and it is better to know it than to weight around it.

Five to approach, three to keep watching

The final steering committee session took the eight to five priority candidates and three high-potential ones. Which companies those were, and the synergies and red flags attached to each, are the client’s commercial property and are not reproduced here.

A scorecard ranks. It does not decide. The gap between the ranking and the recommendation is where the judgement lives, and it should be written down rather than hidden inside the weights.

What can be said is the shape of the answer, because the shape is reusable. The five were not the five highest scorers. Two companies with strong scores were moved to the high-potential list because their ownership made a near-term approach unrealistic, and one lower scorer was promoted because it filled a specific gap in the line that nothing else on the list filled.

The board pack carried, for each of the five: a company profile, the strategic rationale, the synergies, the red flags stated plainly, and an engagement sequence with a timeline to due diligence and the decision gates along it. The three high-potential names came with the condition that would have to change for them to move up.

What this study could not settle

Private company financials are estimates. Most of the eighty-nine were private. Where filings existed they were used; where they did not, revenue and margin were triangulated from headcount, installed base, project announcements and interview evidence. Those figures are good enough to rank candidates against each other. They are not good enough to price one, and the study says so.

The market model is a capacity model, not an order model. It says how much equipment spend the installed fleet implies at a normal replacement rate. It does not predict the timing of any individual purchase, and mills defer.

Willingness to sell was not tested. Nobody was approached. Ownership was assessed from structure and public signals, which is a weaker instrument than a conversation and was flagged as the largest single source of error in the shortlist.

Two customer interviews did not happen inside the window. One was rescheduled and one fell outside the ten weeks. Both would have added to the European operator view, which is the view the modernisation thesis rests on.

One source chart did not reconcile. The regional legend on the production chart contradicted the values and growth rates plotted beside it. The corrected reading is used here and the discrepancy is noted on the exhibit, because a figure that cannot be reconciled should be shown with its problem attached rather than quietly cleaned up.

What the numbers rest on

Every figure on this page is either published by the institutions named above, calculated from them by a method set out in the article, or supplied by the client. Where a number is an identified market size rather than a result, it says so on the line where it appears.

On anonymity. The client is not named. Neither is any company that was screened, shortlisted or recommended, and no financial figure attributable to a named supplier appears anywhere on this page. The market analysis is built from public data and is reproduced in full. The target evaluation is described by its method only.