The interquartile range on that same sold set ran 2.04x–3.59x SDE (Source: BizBuySell). A shop outside that gravity well still needs a reason — certificates, backlog, or a floor — not a hope.
Which print attaches to which earnings base?
SDE multiples attach to owner-operated cash. EBITDA multiples attach to a file that already carries a replacement manager. SDE versus EBITDA is the fork. How much a manufacturing business is worth is the sister question.
Census still counted 284,452 plants in 2023, 68.4% under twenty employees (Source: Census CBP). Most of those shops live nearer the BizBuySell median than a tombstone.
| What it measures | Misuse to refuse | |
|---|---|---|
| 2.75x median SDE | Main Street manufacturing sold set, 2021–2025 | Treating it as EBITDA |
| 2.04x–3.59x IQR | Middle half of that same SDE set | Calling the top of the IQR “the market” |
| ~$268k median SDE / ~$1.13M revenue | Typical sold-file scale in that recap | Scaling it to a $16M search close |
| Stanford ~6.2x EBITDA on a ~$16M median | Search-fund study neighborhood | Pasting it on a job shop |
| GF Data 7.2x / 5.9x under $25M EV | Larger-EV manufacturing recap | Quoting it as a Main Street turn |
Stanford’s 2024–25 search median of about $16 million at ~6.2x EBITDA on a ~25% margin profile is a different animal (Source: Stanford GSB). Quote it only when the file looks like that study.
Why a “state manufacturing multiple” is not a thing on this site
Valuation keeps national size-band math. States keep statute and CBP overlays. A closing in Ohio does not mint an Ohio turn. Do not invent one because a competitor whispered a number at a show.
SBA 7(a) still caps standard change-of-ownership debt at $5 million with a 10% injection (Source: SBA). A multiple that produces a note the box cannot hold is fiction.
How buyers actually apply a multiple after they hear 2.75x
They recast. They test add-backs. They haircut concentration. They compare earnings value to orderly liquidation. They ask whether utilization and backlog support the story. They do not wire 2.75 times last year’s Schedule C.
Buying a manufacturer is the buyer pillar if you are writing a buybox. The FAQ directory is only the map.
ISM’s August 2026 PMI of 54.6 and Fed utilization of 76.0% in July 2026 are climate, not a turn (those prints sit in research; they do not replace a recast). Climate can support a narrative. It cannot invent a cell.
What happens when two buyers quote different bases in the same week?
One individual writes 2.75 times SDE. A searcher writes a lower EBITDA multiple on a smaller base after subtracting a replacement manager. Both can be internally consistent. They are not the same offer. Add the dollars, not the turns, before you pick a winner.
Ask each party to show the bridge. If the individual cannot explain add-backs, the 2.75x is theater. If the searcher cannot name the replacement package, the EBITDA is theater. A process that forces both bridges onto one page is how you stop mixing rooms in your own head.
Climate prints do not break the tie. An ISM PMI of 54.6 and Fed utilization of 76.0% can support a narrative about demand. They cannot tell you whether to accept the SDE offer or the EBITDA offer. The peg, the iron floor, and the certificate stack can.
If both offers ignore a 32% customer, both offers are unfinished. If both offers ignore a Phase I, both offers are unfinished. A higher turn on a file that cannot close is not the better multiple. It is the longer exclusive.
Write the two offers as cash at close, not as turns. Subtract the peg true-up you already fear. Subtract the holdback the 32% customer will attract. Subtract the appraisal if OLV is the floor. The surviving dollar is the comparison. Sellers who pick the higher multiple and then donate cash in the true-up did not pick the higher offer. They picked the prettier slide.
If only one party can finance inside the $5 million 7(a) box, the other party’s turn is a conversation, not a close. Ask who has a preferred lender before you fall in love with a cell.
What should you do with the number you just heard at lunch?
Write the earnings base next to it. Write the size band. Write the sub-sector. If any of those three are missing, you do not have a multiple. You have a slogan.
Jason Taken is a business broker. He will place a file in a band and name the discounts. He will not sign an appraisal. Write the earnings base next to every turn you hear this month. If the speaker cannot name the base, they have not given you a multiple.
Educational information only. The conversion path after you read the owner URL is a confidential intro call.
If two people quote “about three times,” ask three times what. SDE and EBITDA are not cousins you can introduce at the same dinner. Keep them in separate chairs. A multiple without a base is a slogan, and a slogan is not a close.