Sawmill economics: cost structure, KPIs and profitability

A sawmill is simple to describe and hard to run profitably. It buys a heterogeneous, price-volatile raw material, converts it with expensive equipment into products whose prices it does not set, and sells the by-products to whoever needs them. Nearly every managerial question comes back to three levers: what you pay for logs, how much sellable value you get from each log, and how fully you use the capital already sunk. This article reviews the cost structure reported in open studies, the margin squeeze that periodically hits the industry, the KPIs that make the levers visible, and the arithmetic of break-even, working capital and investment. It closes with a worked example built from deliberately invented round numbers.

A note on method. Every real number below comes from a source that was opened and read for this article, and it is given with the country or region and the year the source reports. Where the article does its own arithmetic it says "own calculation". Where something could not be verified, it says so. Open data on sawmill costs is thinner than most people assume, because mill-level cost data is commercially sensitive, so several sources are old or narrow. They show structure well, but no single number below is "the" benchmark for today.

Why sawmill economics look the way they do

Three structural features explain most of what follows.

First, the mill is a price taker on both sides. Logs are bought on regional markets, often from a few large suppliers. Sawn timber is sold into construction, packaging, pallet and joinery markets that are international and cyclical. The margin is the difference between two numbers the mill does not set.

Second, the process is a volume conversion with a physical yield. Every cubic metre of log that does not become sawn timber becomes chips, sawdust, bark or shavings, which are worth much less per cubic metre. Yield is therefore a money question, not only a technical one.

Third, the plant is capital-intensive and long-lived. Sawline, kilns and log yard are paid for whether or not they run. This creates operating leverage: at high utilisation each extra cubic metre carries a large contribution, and when utilisation falls, fixed cost per cubic metre climbs fast.

Cost structure: what open studies show

Think of costs as a waterfall: revenue from sawn timber and residues, then raw material, then conversion costs (labour, energy, tools, maintenance), then overheads and capital costs. Shares depend on country, product mix, year and the denominator used (total cost, production value or turnover), so the studies below are not directly comparable.

Sweden: a survey of sawmill costs

A Swedish master's thesis from SLU (Lindholm, with Sandvik AB) surveyed Swedish sawmills. The thesis is dated 2006 and cites 2005 statistics; the opened text does not state the year of the cost data. It reports this average structure per cubic metre of sawn timber:

Source: Lindholm, SLU thesis, survey of Swedish sawmills (survey year not stated in the opened excerpt). Average total cost is reported as 1,554 SEK per m³; the items sum slightly higher through rounding.

Average net sales revenue for sawn timber was 1,556 SEK per m³, against that total cost of 1,554. That is a near-zero average result (own calculation: about 2 SEK per m³, roughly 0.1% of cost), though the excerpt does not say how by-product revenue was treated, so read it as indicative. The thesis also reports contribution margins: about 616 SEK per m³ after raw material (range 492 to 959) and about 355 SEK per m³ after variable overheads and grinding (range 111 to 708). The second figure must cover administration, depreciation and profit. The spread matters: between the best and worst surveyed mill, the margin available to pay for everything else differed by a factor of more than six (own calculation).

Germany: raw wood share of production value, 2006 and 2010

The Thünen Institute (Seintsch and Rosenkranz, Holz-Zentralblatt, January 2015) analysed German saw, planing and impregnation works (NACE 1610) from official statistics. The raw wood (Rohholz) share of gross production value rose from 35.1% in 2006 to 41.5% in 2010, and all wood-based inputs from 57.4% to 60.7%. The gross production margin fell from 5.4% to 2.1% of gross production value. Gross production value fell from 4.86 to 4.44 billion euro, and employment from 16,559 to 15,081.

The denominator is production value, not total cost, so 41.5% cannot be compared with the Swedish 63%. The document does not break out personnel or energy. It does show the direction clearly: as output value fell, the raw wood share rose and the margin shrank by more than half.

Finland: panel data on 36 sawmills, 2012 to 2021

A 2025 Silva Fennica paper (Koskivaara, Lähtinen and Toppinen) analysed firm-level data for 36 Finnish sawmills and 34 wood construction element producers, 2012 to 2021. It finds that material and wage costs have a relatively strong negative effect on competitiveness, and that sawmills are more uniform than element producers. The sawmill median net profit ratio stayed between -0.6% and +3.4% in every year except 2021, when it was +14.8%. Sawmill liquidity was slowly strengthening over the period.

The paper's Table 3 gives mean material costs of 11.3 million euro and mean salaries of 1.6 million euro per sawmill-year, not as shares of turnover. Comparing the two means (own calculation, observations not necessarily identical), material was roughly seven times the salary bill. The key message is the thinness of profit: for most of a decade the median sawmill sat between a small loss and a low single-digit margin, and 2021 was a price spike, not a new normal.

Poland: dominant raw material cost, little published structure

A 2024 Folia Forestalia Polonica paper on pallet lumber (Stanula, Wieruszewski, Dynowska and Adamowicz) speaks of "the dominant share of raw material costs" but gives no percentage. It does give Polish annual average prices for pine pallet lumber: 442.5 PLN per m³ in 2019, 452.5 in 2020, 757.5 in 2021, 1,225.0 in 2022 and 775.0 in 2023, plus transport costs of 28 to 38 PLN per m³ over 2019 to 2023. From those prices (own calculation) the 2022 price was about 2.8 times the 2019 level and 2023 was about 37% below the 2022 peak. That is the volatility a contribution margin must survive.

In 2024 the director of the Polish Chamber of Wood Industry (PIGPD), as quoted by a trade outlet, said operating costs of Polish sawmills exceeded revenue from product sales and that wood in Poland was the most expensive among neighbouring countries. The article gave no figures, so this is a trade-association statement, not verified data. No open, quantified Polish sawmill cost-structure study was found.

Canada: relative costs in the BC Interior, 2018

A 2019 benchmarking report for the Council of Forest Industries (FEA, covering 2018) compared regions with indices. Delivered wood was the highest input cost in every region. Delivered wood in the BC Interior was about 50% above the US South (index 152 against 100; US Coast 210), and the BC Interior was the least profitable of the seven regions (profit index 42 against 100). The report cites rising delivered wood costs plus regulatory and environmental costs, including carbon taxes. A region with expensive wood can have lower costs elsewhere and still lose.

Margin squeeze: when log prices outrun lumber prices

To a first approximation, gross margin is the sawn timber price plus residue revenue, minus log cost per cubic metre of sawn output. That log cost equals the log price divided by recovery, so a log price rise is multiplied by the inverse of recovery. At 50% recovery, one euro more on the log is two euro more per cubic metre of sawn timber (own arithmetic).

A squeeze occurs when logs get dearer faster than sawn timber, typically when sawlog supply is tight while construction demand is weak. The recent European case is described by Fastmarkets (November 2025). Southern Swedish spruce sawlogs reached over 1,400 SEK per m³ by the second quarter of 2025, "more than doubling since 2019". Finnish stumpage (Luke, October 2025) was about 76 euro per m³ for pine and 79 for spruce, 8 to 9% below the June 2025 nominal highs but 14 to 16% above five-year averages. German spruce sawn timber (44 to 50 by 150 mm) traded at 310 to 325 euro per m³ in autumn 2025. The article describes widespread unprofitability despite continued investment, and reports the planned March 2026 closure of a 115,000 m³-a-year sawmill in central Sweden, citing wood costs more than doubled in five years. This is commercial market commentary, not audited accounts, but it matches the thin margins in the academic data.

For volume context, a December 2025 Dovetail Partners summary of the UNECE committee meeting reports that softwood lumber production in the UNECE region increased 4% in 2024 and was forecast to fall 2.2% in 2025. It contains no margin data.

The options in a squeeze are narrow: raise recovery, shift the product mix to higher-value sizes and grades, grow residue revenue, cut conversion cost, or curtail volume. Curtailment is rational when the marginal cubic metre has a negative contribution, but it costs fixed-cost absorption, skilled staff and customers.

Recovery and value recovery

Recovery is the headline productivity KPI: sawn volume divided by log volume. North American practice uses lumber recovery factor (LRF), board feet per unit of log volume, which is not comparable with a volume percentage. Three cautions apply. The log volume basis matters (under bark versus over bark, and the measurement system). Board-foot recovery uses nominal dimensions and differs from cubic recovery (a point in the 1984 US Forest Products Laboratory review by Steele). And volume is not value: a mill can raise recovery by sawing more low-grade or undersized product while value per log falls.

Value recovery addresses the last point: total sales value of all products from a log lot, per cubic metre of log or divided by log cost. It combines yield, grade, size mix and residues, and compares directly with the log price. A mill with lower volume recovery but a better mix can have higher value recovery.

Steele's review shows the size of the technical levers, using studies from the US. Reducing kerf from 12/32 to 9/32 inch raised board-foot yield by an average of 7% for small logs. Improving sawing accuracy from 0.2 to 0.1 inch added 6.0%. Optimal face opening raised recovery by an average of 21% over the poorest opening for 5 to 20 inch logs. Computerised edging captured 97% of the theoretical optimum, an estimated 12.9% above operator decisions. Offset sawing probably adds 5 to 10% over centred sawing. These are historical figures and modern scanning has shifted the baseline, but kerf, accuracy and optimisation remain worth several points each.

A contrast comes from a 2021 BioResources paper on small and medium sawmills in the Zambian Copperbelt (Mandiringana and co-authors): recovery rates of 21%, 32%, 36% and 51% across mills, push-bench and circular saws at the low end and narrow bandsaws near 50%, with the mills running below 50% of capacity and all respondents citing raw material shortage. The setting differs from Europe, but it shows how technology drives recovery and how supply and utilisation interact.

On definitions: the Swedish thesis reports a survey-average sawing yield of 60% (sawn per log volume) but cites 47% as the national average for mills above 5,000 m³ a year in 2000 (secondary data). The gap may reflect sample, measurement basis or product, which is exactly why a definition must precede any comparison.

Residue revenue: the second product line

The Polish pallet-lumber paper (2024) reports by-product shares of about 20 to 30% chips, 10 to 15% sawdust and up to 8% bark. Its model-based efficiency indicator including by-products was 194 to 220% in 2022 and 135 to 161% in 2023, showing how much the by-product contribution moves with prices. That indicator is a model construct, not an accounting margin.

Three points matter for residue economics:

• Residues are a revenue line with their own cycle. Chips go to pulp, board or energy, sawdust to board or pellets, bark to energy or horticulture. When those buyers curtail, residue prices and the mill's margin fall together.
• Residues are also an energy source. Swedish mills use about 12% of incoming timber as heat in their own processes, mostly drying (Anderson and Westerlund, Applied Energy, 2011). Residues burned in kilns are not sold; their value is the avoided energy cost.
• Scale and location matter. An old Silva Fennica study of Finnish sawmill waste wood (Rinkinen, published 1964, 1959 survey) found that waste per unit of output rose with mill size, that small mills often could not use or sell it, and that chip production became profitable above roughly 1,000 to 2,000 standards of annual output. The numbers are obsolete; the logic that residue logistics favour larger, well-located mills is not.

Capacity utilisation, throughput and uptime

Fixed costs are spread over output, so utilisation is the largest unit-cost driver after the log price. Three related KPIs should be kept apart.

• Capacity utilisation: actual output or log input divided by installed capacity. Define capacity (nameplate or practical under the actual shift pattern); a one-shift mill can look underused yet be fully used within its labour model.
• Throughput: logs or sawn volume per hour through the true bottleneck (primary breakdown, sorting or kilns). A faster saw is useless if the sorter or kiln limits output.
• Uptime or availability: share of scheduled time the line runs. Mills usually split lost time into planned stops, breakdowns, micro-stops and waiting for logs. Overall equipment effectiveness (availability times performance times quality) is the usual composite.

Two reference points were verified. In the United States the sawmill and wood preservation utilisation rate was 64.7% in the fourth quarter of 2024, on a four-quarter moving-average basis (National Association of Home Builders, using Census Bureau and Federal Reserve data). In the Zambian Copperbelt small and medium mills ran below 50% of capacity (2021). Definitions and industry structures differ, so neither is a benchmark for a European mill. No open, current European utilisation or uptime benchmark could be verified, so none is quoted.

Operating leverage is the key economic point. If fixed cost is 53 currency units per m³ at a given volume, raising volume 10% cuts fixed cost per m³ by about 9% (own arithmetic), roughly 5 units per m³. That is why mills keep running through a weak market while the contribution is positive, and cut shifts when it is not.

Energy and labour

Labour. Production personnel were about 10% of total cost in the Swedish survey, administration another 6%. In the Finnish panel, mean salaries were 1.6 million euro per sawmill-year against 11.3 million euro of materials. Labour matters but does not dominate, so automation pays mainly where it also raises recovery or uptime. Labour is semi-fixed in the short run, which feeds operating leverage.

Energy. The main energy cost is kiln-drying heat, plus electricity for saws, handling and debarking. Anderson and Westerlund (Applied Energy, 2011) report that Swedish sawmills, producing 16.4 million m³ of lumber a year (95% from medium-to-large mills, lumber quotient 47%), used about 4.9 TWh of heat a year for drying, and that state-of-the-art techniques could cut this by roughly 2.9 TWh. Dividing gives about 300 kWh of drying heat per m³ of lumber (own calculation, averaging dried and undried product, indicative only). A 2025 Finnish thesis on a chamber dryer at a Hamina sawmill finds heat use high against modern dryers and proposes exhaust heat recovery, better airtightness and fan control, but its abstract gives no numbers. Mills with efficient kilns and a biomass boiler are better insulated against energy price spikes than mills buying heat.

Working capital in drying and stock

Cash sits in three places: the log yard (bought ahead to secure supply), work in progress (green lumber waiting for and going through drying) and finished goods (dried, graded packs awaiting dispatch). Carrying cost is the financing rate times stock value times time. In a business with a margin near 1%, interest on inventory can equal the whole profit, as the worked example shows. Stock also carries price risk: wood bought at one log price and sold at another produces inventory gains or losses that can dominate reported profit in volatile years, one reason Finland's 2021 median net margin (14.8%) looked so unlike the other years.

Open literature on this is thin. A University of New Brunswick report (Rathbone, 1988) on kiln drying for New Brunswick sawmills lists reduced inventory and lower freight among secondary benefits, and concluded that a 20% minimum acceptable rate of return could not be achieved at the time, since the return depends on the price spread between dried and green lumber and on savings elsewhere. It is dated, but the logic holds: a kiln pays through the spread and the cash it releases, not through technical elegance. No open data on typical inventory days in European sawmills was verified, so no benchmark is given.

Break-even logic

Break-even volume equals fixed costs per period divided by contribution per m³, where contribution is revenue per m³ (sawn timber plus residue credit) minus variable cost per m³ (logs, energy, consumables, outbound logistics). Dividing by capacity gives break-even utilisation. It shows the margin of safety, the sensitivity of profit to prices, whether a volume drop is survivable, and what an investment must deliver.

Fixed and variable are not clean categories (labour is semi-fixed, maintenance partly variable), and log cost per m³ of sawn output depends on recovery, so recovery is itself a cost driver.

Worked example (illustrative, invented numbers)

This example uses invented round numbers. It is not data from any real mill, country or year and is not a benchmark. It shows only how recovery, prices, volume and financing interact.

Assume 100,000 m³ of logs a year, installed capacity of 125,000 m³ of logs (utilisation 80%), recovery of 50% (so 50,000 m³ sawn, capacity 62,500 m³ sawn), a delivered log price of 100 euro per m³, a sawn price of 260 euro per m³ and residue revenue of 24 euro per m³ of sawn output.

Source: invented illustrative numbers; arithmetic by the author.

At 50,000 m³ the result is 2 euro per m³, or 100,000 euro a year on revenue of 14.2 million euro, a margin of about 0.7%. That is deliberately close to the thin margins seen in the real studies, but it is an invented case.

Break-even. Fixed costs are 53 times 50,000, or 2.65 million euro. Divided by a contribution of 55, break-even is about 48,200 m³ sawn: 96% of planned output but about 77% of capacity. The margin of safety is under 4% of volume.

Sensitivity 1, log price. A rise of 5 euro per m³ of logs (5%) adds 10 euro per m³ sawn (5 divided by 0.50). Contribution falls to 45, the result becomes a loss of 8 euro per m³ (400,000 euro a year), and break-even rises to about 58,900 m³, or 94% of capacity. A 5% log price rise turned a small profit into a large loss. This is the squeeze in numbers.

Sensitivity 2, sawn price. A fall of 10 euro per m³ in the sawn price (3.8%) has exactly the same effect on contribution as the log price rise. Both together bring contribution to 35.

Sensitivity 3, recovery. Raise recovery from 50% to 52% on the same 100,000 m³ of logs. Output rises by 2,000 m³. Each extra cubic metre sells for 260, costs about 29 more in energy, consumables and logistics, and displaces about 24 of residue revenue. The gain is about 2,000 times (260 minus 29 minus 24), or 0.41 million euro a year, about four times the base profit, from two points of recovery. This is why recovery is usually the first KPI to attack.

Sensitivity 4, working capital. Take 60 days of finished stock at an invented cost value of 250 euro per m³. On 50,000 m³ a year that is about 2.05 million euro of stock (12.5 million times 60 divided by 365). At 6% financing it costs about 123,000 euro a year, more than the whole base profit. Every 10 days of stock is about 342,000 euro of capital, or 20,500 euro a year at 6%. (The 3 euro per m³ interest line in the table is smaller because it is an invented partial charge; in a real mill, stock and equipment financing would be examined together.)

Investment payback. Suppose a scanner and optimiser upgrade costs 1.0 million euro and adds one point of recovery (invented). By the same arithmetic one point is worth about 1,000 times (260 minus 29 minus 24), or 0.21 million euro a year, so simple payback is about 4.8 years. A falling log price or curtailment shrinks the gain; better grading enlarges it. A real appraisal would add maintenance, training, installation downtime and the risk that promised recovery gains are not achieved.

Investment decisions and payback

Appraisal follows normal capital budgeting: incremental cash flows, a risk-adjusted discount rate and sensitivity tests. Three practical points.

Payback is a screening tool, not a decision rule. It ignores cash flows after the payback date and the time value of money. With asset lives of 15 years or more in a cyclical industry, net present value and internal rate of return across a full price cycle are better guides.

Sensitivity can dominate the answer. A Swedish student report from SLU (Hoflund and Snögren, 2011) built a hypothetical new saw line from data supplied by a forest-owner association and stated that the investment is fictitious and the figures not literal. Assumptions: investment 150 million SEK, 15-year life, 10% discount rate, about 416,667 m³ of logs a year, 200,000 m³ of sawn output (yield 48%), logs at 700 SEK per m³, sawn timber at 1,600 and by-products at 400 SEK per m³. The result was an NPV of about 158 million SEK, an IRR of about 25.9% and discounted payback in year five. The point is the sensitivity: NPV falls to zero if the log price rises by only about 49.8 SEK per m³ (about 7% over baseline, own calculation), if the sawn price falls by about 104 SEK per m³, or if both product prices fall about 5.1%. An investment that looks very profitable can hinge on the price spread, which is the lesson of the whole article.

Match the investment to the constraint. The best investments target the bottleneck, recovery and value (scanners, optimisers, thin kerf, sorting) or energy (kilns, heat recovery). Capacity expansion is justified only if log supply is secured at a price leaving a positive contribution, as the Zambian evidence and the European closures citing wood costs imply. Investment follows low wood costs: ResourceWise reported in 2019 a Beck Group estimate of more than 2.5 billion dollars flowing into southern yellow pine sawmills in the Southern US.

Benchmarking

Benchmarking turns KPIs into a management tool at three levels. Internal benchmarking compares shifts, lines and periods within one mill: cheap, with few definition problems. Peer benchmarking compares mills in a closed group; the Swedish survey shows what it reveals (contribution after variable overheads from 111 to 708 SEK per m³), and the BC Interior report does the same for regions with an index against a reference region. Public statistics (national offices, Eurostat structural business statistics, UNECE, national forest institutes) give averages by country and size class. A search for Eurostat sawmilling (NACE C16.1) figures did not yield a number that could be opened and verified, so none is quoted.

Four rules help. Define each KPI exactly (log volume basis, shift pattern, residue treatment). Compare like with like (product mix, log size, species). Look at distributions, not averages. And turn results into a gap analysis: what would closing half the gap to the best quartile be worth in euro per m³?

Small versus large mills: economies of scale

Larger mills tend to have lower unit costs: fixed costs are spread, equipment is more efficient, purchasing is stronger and residue logistics are easier. The evidence supports this but not without limit.

A Swedish study (Månsson, Journal of Forest Economics, 2003) estimated scale elasticity with an input distance function. Average scale elasticity was above 1, indicating economies of scale; most units could gain from expanding and a minority from shrinking. The opened page does not state the data years. In North America, ResourceWise (2019) found a strong inverse relation between production volume and cash manufacturing cost, but the largest mills (300 to 500 million board feet a year) were not substantially cheaper than other large mills (175 to 300 million).

In practice, scale helps up to a point, after which the log supply radius, haulage cost and single-site risk offset it. Small mills survive on cheaper local log access, niche products, direct sales at higher prices or captive residue use. They are most exposed where large competitors set the log price and the product is a commodity.

Financing and support schemes

Sawmill investments are financed from retained earnings, bank loans, leasing, supplier credit and sometimes public support. Lenders look at debt service in a downturn, not an average year, so a plan that works only at the top of the price cycle is not bankable.

On public support, only a limited amount could be verified. One example is Wallonia, Belgium: under the Walloon CAP programme for 2023 to 2027, investment support for forest harvesting and primary wood processing SMEs has a base aid rate of 20% of the standard amount, rising with bonuses to a maximum of 40%, with a cap of 500,000 euro, for machinery and equipment and for storage or shelter buildings (Interreg Europe good-practice page). That page names harvesting equipment, chippers and forestry machinery, not sawmill equipment, so it is no general sawmill grant. No specific current Polish scheme for sawmill investment, and no EU fund eligibility rule for Polish sawmills, could be verified from the sources opened, so none is claimed. Check current call documents and state-aid rules for your firm size and region.

Sources used

The sources below were opened and read for this article.

• Lindholm, G., SLU master's thesis on the cost and income structure of Swedish sawmills (dated 2006): https://stud.epsilon.slu.se/11299/1/lindholm_g_171010.pdf
• Koskivaara, A., Lähtinen, K., Toppinen, A. (2025), Longitudinal analysis of financial performance of sawmills and wood construction element producers in Finland, Silva Fennica 59(2): https://www.silvafennica.fi/article/24033
• Seintsch, B., Rosenkranz, L. (2015), Thünen Institute, wood-based value added, Holz-Zentralblatt 5: https://www.openagrar.de/servlets/MCRFileNodeServlet/timport_derivate_00029135/dn054878.pdf
• Stanula, Z. et al. (2024), Comparative efficiency of roundwood processing into pallet lumber, Folia Forestalia Polonica 66(4): https://www.ibles.pl/wp-content/uploads/2024/12/10.2478-ffp-2024-0023.pdf
• Timber Industry News, Polish sawmill industry and profitability (PIGPD statement, 2024): https://timberindustrynews.com/?p=13924
• FEA (2019), BC Interior sawmilling cost and profit benchmarking, for COFI: https://www.cofi.org/wp-content/uploads/FEA-BC-Interior-Public-Final-August-2019.pdf
• Fastmarkets (November 2025), European sawn timber market trends and outlook: https://www.fastmarkets.com/insights/european-sawn-timber-market-trends-and-outlook-november-2025/
• Dovetail Partners (December 2025), UNECE region market summary: https://www.dovetailinc.org/upload/tmp/1767192107.pdf
• Steele, P. H. (1984), Factors Determining Lumber Recovery in Sawmilling, USDA Forest Products Laboratory GTR FPL-39: https://www.fpl.fs.usda.gov/documnts/fplgtr/fplgtr39.pdf
• Mandiringana et al. (2021), factors limiting performance of small to medium sawmills in the Zambian Copperbelt, BioResources 17(1): https://bioresources.cnr.ncsu.edu/resources/analysis-of-the-factors-limiting-the-performance-of-small-to-medium-scale-sawmills-in-the-copperbelt-of-zambia/
• Rinkinen, I. (1964), Waste wood of the Finnish sawmill industry, Silva Fennica 115: https://www.silvafennica.fi/article/4718
• Anderson, J.-O., Westerlund, L. (2011), Surplus biomass through energy efficient kilns, Applied Energy 88(12): https://ideas.repec.org/a/eee/appene/v88y2011i12p4848-4853.html
• Pilhjerta, J. (2025), chamber dryer energy consumption thesis (Finland): https://www.theseus.fi/handle/10024/895682
• National Association of Home Builders (2025), US sawmill production and capacity in 2024: https://eyeonhousing.org/2025/03/u-s-sawmill-production-capacity-constant-in-2024/
• Rathbone, H. J. (1988), An economic analysis of kiln drying for New Brunswick sawmills, University of New Brunswick: https://unbscholar.lib.unb.ca/handle/1882/34002
• Hoflund, P., Snögren, J. (2011), Investeringskalkyl för en ny såglinje, SLU: https://stud.epsilon.slu.se/2752/11/Hoflund_Snogren_110606.pdf
• Månsson, J. (2003), Economies of scale in the Swedish sawmill industry, Journal of Forest Economics 9(3): https://nowpublishers.com/article/Details/JFE-0028
• ResourceWise, How are lumber manufacturing costs changing the industry (2019): https://www.resourcewise.com/market-watch-blog/how-are-lumber-manufacturing-costs-changing-the-industry
• Interreg Europe, Investment support for forest harvesting and primary wood processing SMEs (Wallonia CAP 2023-27): https://www.interregeurope.eu/good-practices/investment-support-for-forest-harvesting-primary-wood-processing-smes

Limits of this article

• No open, verified data was found for Baltic sawmill cost structure, Eurostat sawmilling statistics, current European utilisation and uptime benchmarks, or inventory days; none is quoted.
• Several cost-structure sources are old (Swedish thesis 2006, German data 2006 and 2010, Steele 1984, Rathbone 1988, Rinkinen 1964) and illustrate structure rather than current levels. The survey year of the Swedish cost data was not stated in the opened text.
• Cost shares from different studies use different denominators and cannot be compared directly. Trade-press figures (Fastmarkets, Timber Industry News) are not audited data; the Polish chamber statement was unquantified.
• The worked example uses invented round numbers and is not data. The Hoflund and Snögren investment case is a stated fictitious calculation. Own calculations are labelled.
• Polish and wider EU support schemes for sawmills could not be verified; only the Walloon example is cited, and it does not name sawmill equipment.