Why does size band change the multiple?
Size changes who can close. Individuals and SBA lenders live on Main Street SDE files. Searchers and independent sponsors live in the first institutional EBITDA band. Add-on platforms and strategics thicken as earnings clear a few million. Platforms and larger strategics dominate once adjusted EBITDA is into eight figures.
GF Data’s Q1 2026 manufacturing drilldown, as recapped, averaged 7.2x EBITDA, printed 5.9x under $25 million enterprise value, and 8.8x in the $100–$250 million enterprise-value bracket (Source: Iconic GF Data recap). Those are larger-EV observations. They are not a license to quote 7x on a $1.1 million-EBITDA metal fab shop.
IBBA and M&A Source Market Pulse Q1 2026 plotted LMM multiples as SDE below about $2 million enterprise value and as EBITDA in the larger bands. The $5–$50 million sparkline printed near 4.5x–5.1x EBITDA (Source: IBBA highlights). Treat that as advisor-survey climate. It is not a sub-sector quote and it is not GF Data.
Pepperdine 2025, as secondarily cited, put PE-style marks near 5.5x EBITDA around $10 million of earnings. Stanford’s 2024–25 search-fund median sat near 6.2x on a roughly 25% margin file (Source: Stanford GSB). Each source has a room. How each buyer type prices a manufacturer is the companion.
What do the published EBITDA columns actually say?
The site publishing table for adjusted EBITDA, going-concern, cash-free and debt-free, is synthesized from the sources in /research/valuation-benchmarks.md. On-page language: advisor-survey and reported-deal ranges, not an appraisal.
| Earnings band | Commodity fab, print, and similar | Precision machining and plastics | AS9100 aero and ISO 13485 medical |
|---|---|---|---|
| $1–3 million adjusted EBITDA | 4.0x–5.5x EBITDA | 5.0x–6.5x EBITDA | 6.0x–8.0x EBITDA |
| $3–10 million adjusted EBITDA | 5.0x–6.5x EBITDA | 5.5x–7.5x EBITDA | 6.5x–9.0x EBITDA |
| $10 million-plus adjusted EBITDA | 6.0x–8.0x EBITDA | 6.5x–9.0x EBITDA | 7.0x–11.0x EBITDA |
Read left to right only after you have confirmed the earnings base is adjusted EBITDA with a replacement general manager already subtracted. If the owner is still the estimator and the “EBITDA” still includes $240,000 of owner salary you refused to replace, you are not in this table. You are on the SDE versus EBITDA page.
The $10M+ column is where GF Data’s higher brackets start to rhyme with the site table. The $1–3M column is where search math and small-add-on math live. Do not drag an 8.8x $100–$250 million observation into a $2.2 million-EBITDA electronics EMS file.
How should a $1–3 million EBITDA shop read the table?
This band is the first institutional room. Buyers will ask for a quality-of-earnings once add-backs get aggressive; the research file flags QoE as common around ~$2 million+ EBITDA and sometimes earlier. They will also ask whether the top customer sits above about 25–30% of sales, whether the last surveillance audit is clean, and whether quoting dies when the owner is out.
Illustrative only, precision machining, EBITDA $1–3M: $2.1 million adjusted EBITDA, no customer above 18%, current ISO 9001, 2016–2022 CNC mix. The precision column is 5.0x–6.5x. Illustrative enterprise value $10.5–$13.7 million before the working-capital peg and excess cash. If orderly liquidation of the cells is $4.8 million and inventory is clean, earnings still win. If the same earnings sit on 1990s manuals and a 38% customer, you do not get the top of the cell.
Illustrative only, commodity stamp, same earnings: the general-fab column is 4.0x–5.5x. The iron is easier to replicate. Customers shop. A strategic might still pay for a scarce die book. A financial buyer will not pay an aerospace cell for a stamp house.
CTA Acquisitions’ synthesis of BizBuySell Insight and IBBA Pulse (2024–Q2 2026) places some cleaner small industrial files in a wider observed SDE band. That synthesis does not move a $2 million-EBITDA shop into the SDE column. Size band first.
What happens in the $3–10 million and $10 million-plus columns?
At $3–10 million of adjusted EBITDA, add-on platforms and strategics thicken. IBBA Pulse Q1 2026 reported that 83% of deals over $5 million attracted three or more offers (Source: IBBA). That is a process fact. It is not a promise that your automotive IATF supplier clears 9x.
The aerospace / medical column in this band is 6.5x–9.0x. That cell assumes current AS9100 or ISO 13485, transferable audits, and a customer mix a second chair can hold. A lapsed certificate is a discount, not a footnote. See how quality certifications affect value.
Above $10 million of adjusted EBITDA, the larger-EV series becomes more relevant. GF Data’s 8.8x observation lives in the $100–$250 million enterprise-value bracket, which is not the same as “we do $11 million of EBITDA, so we are 8.8x.” You still pick the process column. You still test owner dependence and backlog quality.
Why GF Data is not a Main Street quote
Main Street manufacturing still speaks SDE. BizBuySell’s 2021–2025 manufacturing set shows a 2.75x median SDE multiple (Source: Iconic recap of BizBuySell). That median belongs on owner-operated shops, typically under about $1 million of discretionary earnings. It is the wrong instrument for an EBITDA conversation and the wrong instrument for a $6 million-EBITDA food co-packer.
The reverse error is just as common. A seller hears 7.2x on a podcast and applies it to a $900,000-SDE wood millwork shop. That is not size-band work. That is brochure math. Manufacturing valuation methods is the sequence. The calculator is the fence.
Time-to-close is also size-sensitive. BizBuySell’s 2025 year-in-review put manufacturing at a 223-day median close, the slowest sector in that review (Source: BizBuySell). Lower-middle-market processes often run intro-to-close on a longer clock once a CIM is in market. Neither figure is a guarantee. Both argue against treating a multiple as a week-six event.
How to move inside a size-band cell
Size picks the column. Process picks the row — see multiples by sub-sector. Inside the cell, concentration, certificates, backlog, owner dependence, equipment age, union withdrawal liability, and environmental RECs move you. Those are broker rules of thumb, labeled as such in the research file. They are not a published turn for every plant.
ISM August 2026 PMI 54.6 and Prices 71.1 (Source: ISM) belong in the CIM as operating context. They do not bump a shop from the $1–3M column into the $10M+ column. Fed G.17 utilization 76.0% in July 2026 belongs in capacity diligence, not as a multiple escalator.
Sellers: sell my manufacturing business. Buyers: buy a manufacturing business. Worked files: manufacturing valuation worked examples. Jason Taken is a business broker, not an appraiser. No cell on this page is a quote on your plant.
What a CIM should state in the first ten lines
State the sub-sector, the earnings base, the size band, and the cell. “Specialty packaging, adjusted EBITDA $3–10M, published range 5.5x–7.5x, advisor-survey / reported-deal, not an appraisal.” Then state the two or three facts that move you inside the cell. Do not open with a GF Data average and hope the reader never asks which enterprise-value bracket it came from.
State cash-free, debt-free, and a working-capital peg in the same breath as the multiple. A reader who thinks inventory is extra will misread every column on this page. State whether real estate is in or out. Out is the default on this site.
State the replacement manager if you are speaking EBITDA. A first page that still includes $250,000 of owner salary inside “EBITDA” is an SDE page wearing the wrong label. SDE versus EBITDA exists so that sentence can be short.
How lenders quietly re-cut a size-band story
A lender does not have to accept your column. Collateral, DSCR, and the 7(a) $5 million cap can push a file back toward SDE language or toward an asset conversation even when the CIM is written in EBITDA. Ask the lender which earnings definition they will underwrite. Do not discover the answer in week eight.
Search-fund credit boxes often want a minimum EBITDA after the GM. If your $1.3 million “EBITDA” becomes $1.05 million after a real package, you may still be in the $1–3M column — or you may fall out of the searcher’s box entirely. The column did not lie. The add-back did.
R&W insurance, more common as enterprise value rises, does not raise the multiple. It changes how risk is parked. Do not treat a policy conversation as a reason to sit in the $10M+ column when earnings are $2.2 million.
Size-band mistakes that show up after a letter of intent
The seller who accepted a $1–3M column and then “found” $400,000 of add-backs in exclusivity. The buyer who used the $10M+ column as a negotiation ceiling on a $3.4 million-EBITDA auto file. The CIM that printed both SDE and EBITDA multiples and invited the reader to pick the larger. All three are size-band failures.
CTA’s wider observed SDE band for some cleaner small industrials is a reminder that medians are middles. It is not permission to drag a $2.8 million-EBITDA plant back into SDE so the multiple looks familiar. If you want a teaching walkthrough, use worked examples. If you want the fence on one page, use the calculator.
How a first-page CIM block should read when the column is honest
Specialty packaging, adjusted EBITDA in the 3 to 10 million band, published range 5.5x to 7.5x, advisor-survey and reported-deal, not an appraisal. Then the two or three inside-the-cell facts. Then cash-free, debt-free, peg, and real estate out. Then the replacement manager if you are speaking EBITDA. That block is the whole discipline. A GF Data average without an enterprise-value bracket is not a block.
Lenders may re-cut the column. Search boxes may drop you out after a real GM. R&W insurance does not raise the multiple. Finding 400000 of add-backs after a 1 to 3 million letter is a size-band failure. So is using the 10 million-plus column as a ceiling on a 3.4 million EBITDA auto file.
Educational information only. Not an appraisal and not tax, legal, or engineering advice. Confirm with M&A counsel, a CPA, and qualified appraisers. No value is guaranteed.