Which mistakes cost manufacturing buyers the most after the wire?
The expensive ones: a PE multiple on an SDE shop, a tour before tax returns, a skipped lien column, and an earn-out modeled as lender cash. Buying a manufacturing business is underwriting. This page is the catalog of ways buyers refuse to underwrite.
Multiple and stack mistakes
Applying a GF Data 7.2x average or Stanford's 6.2x to a $400k-SDE general CNC shop is the classic. GF Data Q1 2026 manufacturing recap averaged 7.2x, 5.9x under $25 million EV, 8.8x at $100–250 million EV — a larger-EV series (Source: Iconic GF Data recap). Stanford's 2024–25 median purchase of $16 million at about 6.2x sat on a ~25% margin company (Source: Stanford GSB). Main Street manufacturing in BizBuySell's 2021–2025 set cleared at 2.75x median SDE (IQR 2.04x–3.59x) on about $268k SDE (Source: BizBuySell via Iconic). Size band × sub-sector stays on every multiple. Mixing cells is a mistake, not a strategy.
Writing an LOI a 7(a) lender cannot fit under the $5 million cap is the financing twin. SOP 50 10 8 wants a 10% injection. A seller note counts as equity only on full standby for the life of the loan and only up to half the injection (Source: SBA). Treating current interest as "standby" is how exclusivity dies in credit. Read SBA up to the cap before you promise a structure.
Modeling earn-outs as year-one DSCR is the third stack mistake. Most lenders ignore them until earned.
| Mistake | What you told yourself | What to do instead |
|---|---|---|
| PE multiple on an SDE shop | "Manufacturing clears 7x" | Open the size-band table |
| Tour before tax | "I need to feel the culture" | Evaluate on paper |
| No lien column | "We'll clean it up at close" | UCC + serials now |
| Customary NWC | "It's standard" | Write the peg |
| Skip Phase I | "It's been a shop forever" | ASTM E1527-21 when dirt or process is wet |
| Gov contracts assign themselves | "The CO likes us" | FAR 42.12 |
| Spouse stays for free | "They're family" | Budget a controller |
| Two-week training | "I'm a fast learner" | Stay plan or walk |
| Earn-out = cash flow | "The bank will get it" | They will not |
| Second plant first | "Platform premium" | Run plant one sick |
Diligence mistakes that look like speed
Touring before reading tax returns. Ignoring liens on the only 5-axis cell. Treating government contracts as assignable. Skipping Phase I. Assuming the owner's spouse will stay. Sending an LOI with no peg method in a process where IBBA Q1 2026 found 83% of deals over $5 million drew three or more offers (Source: IBBA). Speed without a freeze list is how you "win" a file other people priced.
Skipping a quality of earnings because you liked the tour is how add-backs become your salary. Hoping the lender "sees it differently" is not a close condition.
BizBuySell 2025 put manufacturing at a 223-day median close (Source: BizBuySell). Demanding thirty days on an SBA plant file is a mistake about calendars. So is camping in the lobby until employees learn.
People and platform mistakes
Replacing the estimator in your head before you have hired one. Announcing a sale to the second shift. Changing ERP in week two. Buying a second manufacturer before the first ERP can produce a concentration schedule. Prepaying a roll-up in the first multiple.
ISM Prices 71.1 in August 2026 (Source: ISM) makes "we'll raise price later" a margin mistake if the quoting file has no surcharge history. Fed utilization 76.0% (Source: Fed G.17) does not excuse a 95% overtime story you never tested.
Mistakes that show up after you think you won
Winning the LOI and then starving diligence is a sequel mistake. Exclusivity is not a trophy. It is a clock for appraisal, Phase I, and QoE. Buyers who "save money" by skipping the physical invent a true-up war. Buyers who accept a new peg definition at midnight invent a different price. The LOI guide already said that. Fatigue makes people forget.
Another sequel: announcing the deal to the floor before close, then wondering why the programmer has an offer. Plant-visit protocol exists after the letter too. A third: changing item numbers in week one so nobody can find a traveler. A fourth: paying the seller's personal credit-card AP without a cutoff. A fifth: shipping quality-hold stock because the cage was full.
SBA packaging mistakes continue after the letter. SOP 50 10 8.1 applies to loan numbers on or after 1 October 2026. Using the wrong SOP text in a condition memo is how a file restarts. Combined 7(a)+504 can reach $10 million under the 2026 notice (Source: SBA). Citing that number as if it were a $10 million 7(a) is a mistake that credit committees notice.
Environmental mistakes have long tails. Closing on a shrug because "Phase I is in process" can become a REC you own. ASTM E1527-21 / AAI is the practice standard. Walking past unlabeled totes because the tour was pretty is how you buy a tank farm.
Mistakes about who you are
Pretending you are a searcher with committed capital when you have a coffee and a spreadsheet. Pretending you are an operator when you need a stay for every cell. Pretending you are above the cap when you have a 5% down fantasy from a prior SOP. Sellers and brokers will ask. HedgeStone will not hide the answer.
Evaluation exists so you kill files before these mistakes get expensive. Census CBP 2023: 284,452 establishments, 68.4% under 20 employees (Source: Census CBP). Most mistakes on this page happen on small files where people think process is optional. Small files still have liens, WIP, and certificates.
Deloitte/MI's 3.8 million / 1.9 million workforce figures (Source: Deloitte/MI) make "I'll just hire" a labor mistake. BLS OEWS May 2025: welder median $53,750 (Source: BLS OEWS). Budget the hire or keep the person.
Foreign-person buyers who treat ITAR as a logo mistake § 122.4(b)'s 60-day DDTC notice (Source: DDTC). Domestic buyers who treat government contracts as assignable mistake FAR 42.12. Both are in the red-flag page. Repeating them after you read that page is a choice.
A worked file: how five mistakes stack
Composite, not a listing. A buyer likes a 16-person general CNC shop. Stated SDE $410,000. Asking EV $2.7 million — about 6.6x, which is a search-fund-shaped number on a shop that is not a ~25% margin / $16 million median file (Source: Stanford GSB). Mistake one: the multiple cell.
They tour first because a friend said culture matters. They miss that the only 5-axis has a lessor. Mistake two: paper last. They write an LOI with "customary NWC" and a seller note that pays interest, then tell the banker it is SOP 50 10 8 equity. Mistake three and four: peg and standby. They model a $200,000 earn-out as year-one cash. Mistake five.
A corrected path on the same composite: treat it as a 2.0x–3.6x SDE conversation in the BizBuySell manufacturing quartile band (Source: BizBuySell via Iconic). Read tax versus job-cost. Run UCC. Write the peg. If 7(a) fits under $5 million with a 10% cash-or-standby injection, write that honestly (Source: SBA). If it does not, stop. Count the physical. Keep the estimator.
That composite is not clever. It is the same five mistakes in one file. Most wrecked purchases are not exotic. They are this stack. Fed utilization 76.0% (Source: Fed G.17) and ISM Prices 71.1 (Source: ISM) will not save it. A pretty cell will not save it. Evaluation exists to break the stack before the letter.
Deloitte/MI's 3.8 million / 1.9 million workforce figures (Source: Deloitte/MI) add a sixth mistake if the buyer also planned to "replace the programmer later" with no wage. BLS OEWS May 2025 machinist median $58,750 (Source: BLS OEWS) is the floor for that conversation, not the ceiling in a scarce metro.
Jason Taken is a business broker. He will name mistakes before they become wires. He will not undo a multiple you invented. Educational information only. Confirm legal, tax, environmental, and lending questions with licensed people. No outcome is guaranteed.
Questions that prevent the five-mistake stack
Which size-band cell are you in. Did you read tax before the tour. Did you run UCC. Did you write the peg. Did you model the earn-out at zero for DSCR. If you cannot answer those on a pad, you are about to repeat the composite on this page.
Stop. Read the evaluation page. Then write the letter. Sequel mistakes after you "win" — starving diligence, announcing to the floor, changing item numbers in week one, paying personal cards without a cutoff — wreck files that survived the first stack. SOP text that governs the loan number still matters in Q4 2026. Using the wrong SOP is how a file restarts.
Foreign-person ITAR clocks and FAR novation are choices if you read the red-flag page and still handshake them.
HedgeStone will name the five-mistake stack on the intro call if that is what your last LOI looked like. Bring a size-band cell, a UCC plan, and a peg sentence. Winning exclusivity and then starving diligence is the sequel. Exclusivity is a clock, not a trophy.
Treating CIM backlog as contracted revenue
A CIM backlog chart is a sales slide until you can tie it to purchase orders, releases, and finishable travelers. Buyers who model the chart as year-one cash repeat the earn-out mistake with a different label. Ask which jobs are released, which are forecast, and which cannot finish if the seller leaves. Then model the forecast at zero for DSCR. A last-minute EV increase that breaks the five-million-dollar cap is the same family of mistake. If the letter no longer fits 7(a), say so and rebuild the stack. SOP text that governs the loan number still matters in Q4 2026. Using the wrong SOP is how a file restarts after you spent counsel. Sequel mistakes after you win — starving diligence, announcing to the floor, changing item numbers in week one — wreck files that survived the first stack. Exclusivity is a clock, not a trophy.