How do you class inventory and WIP before you trust the peg?
Buying a manufacturing business without those classes is how the working-capital peg becomes a fight after the truck arrives. The buyer-side peg explainer is the method. This page is the stock.
Four classes and the ownership question
Raw. Steel, resin, components, inserts, chemicals. Age and alloy or lot matter. ISM's August 2026 Prices Index of 71.1 (Source: ISM) is a reminder that last year's cost may not be this year's value if the shop cannot pass through.
WIP. The fight. Job-cost WIP and percentage-of-completion need a method before LOI, not after. Valuing an incomplete job at quoted margin is hope. Valuing it at cost incurred, with a test that the job can finish without the seller's tribal knowledge, is inventory.
Finished. Saleable, not rejected, not "waiting on a customer who ghosted." Customer-owned finished goods sitting in your cage are not yours.
Consigned / customer-owned. Tooling, consigned resin, vendor-managed components. If you put them in the peg, you bought someone else's property.
Ask who owns the dies, molds, fixtures, and programs. On many plastics and stamping files, tooling ownership is the business. Counting a customer-owned mold as an asset is a red flag dressed as a balance sheet.
| Class | What to request | Typical write-down trigger | Peg treatment |
|---|---|---|---|
| Raw | Age, alloy/lot, consignment flag | No movement in your finance band; obsolete spec | Include if owned and usable |
| WIP | Job list, % complete method, hours to finish | Jobs that cannot finish without the seller | Cost, not quoted margin |
| Finished | Age, quality hold, customer-owned flag | Holds, returns, dead SKUs | Include if saleable and owned |
| Tooling | Ownership schedule, steel remaining | Customer-owned; worn out of spec | Usually excluded if not yours |
| E&O reserve | Policy and last physical | Policy weaker than your lender | Apply before the average |
ABL market practice — label it as such — often advances 40–60% of eligible inventory, with a deeper haircut on WIP, versus 70–85% of eligible AR. Job shops with lumpy WIP get less help than they expect. SBA 7(a) still caps at $5 million with a 10% injection (Source: SBA). The peg still has to be countable.
The count is the deal
The close count should be physical. If the seller refuses, walk or hold back enough cash that a miss does not become your problem. Closing is where the signed count lives next to the wire.
Write the protocol in the purchase agreement: who counts, how WIP is measured, what happens to receipts and shipments in the window, and how disputes are decided. A quality of earnings can test whether the reserve policy matches history. It cannot count the cage for you.
Job costing and WIP are the same conversation. If the ERP cannot produce a job list with cost incurred and hours to complete, budget someone who can rebuild it from travelers and packing lists. That cost belongs in your model.
Sub-sector wrinkles without inventing SKUs
EMS and harness shops carry customer-owned designs and component risk. Food and co-pack carry shelf life and allergen holds. Medical and aero carry traceability. Automotive carries release-based inventory that can reverse. The classes stay. The aging bands tighten.
Fed G.17 utilization 76.0% in July 2026 (Source: Fed G.17) does not tell you whether the finished-goods rack is saleable. Book-to-bill climate from ISM (backlog 51.8 in August 2026) does not either. Your target's travelers do.
Main Street manufacturing in BizBuySell's 2021–2025 set cleared at 2.75x median SDE on about $268k SDE (Source: BizBuySell via Iconic). On a file that size, $80,000 of obsolete inserts is not a rounding error. It is a third of a turn. BizBuySell 2025 put manufacturing at a 223-day median close (Source: BizBuySell). Use some of those days to age the stock, not to argue about a spreadsheet.
Stanford's $16 million / 6.2x / ~25% margin median (Source: Stanford GSB) describes companies that usually have inventory controls. It is a search-fund median, never a job-shop multiple, and never a reason to skip the physical on a 14-person shop.
How to sample a cage without living there
You do not need to count every insert in diligence. You need a designed sample: oldest raw lots, largest WIP dollars, finished goods on hold, and anything the seller calls "consigned" without a tag. Photograph tags if the NDA allows. Tie a traveler to a location. If the traveler and the location disagree, the report is a story.
Cycle-count accuracy claims need the last three count adjustments, not a speech. A shop that adjusts $40,000 every quarter is telling you the peg average includes fiction. Put that in the QoE scope. On a Main Street file near ~$268k SDE (Source: BizBuySell via Iconic), $40,000 is not noise.
Customer-owned finished goods should have a shipping plan. If they have sat 180 days, ask why. You may be storing someone else's obsolete program and paying to insure it. That is not an asset. It is a landlord problem.
Scrap bins and quarantine cages are inventory too. If they are full and the P&L scrap line is thin, job-cost is leaking. Job costing will show the same lie from the other side.
Sub-sector inventory clocks you should not invent
Do not publish a national days-on-hand number. Publish the target's. Then compare it to common sense for the process: EMS components go obsolete with packages; food has shelf life; medical has lot traceability; aero has life-limited parts; automotive has release reversals; chemicals have expiration and segregation. The classes stay. The clock changes.
SBA 7(a) still caps at $5 million with a 10% injection (Source: SBA). Lenders will want classed inventory with E&O. They will not accept "it's all good steel." Combined 7(a)+504 can reach $10 million under the 2026 notice (Source: SBA). It still wants a countable cage.
A fourth heading belongs here because buyers skip it: tooling as inventory's cousin. House molds and dies can be assets. Customer-owned tools are not. Worn tools that cannot hold tolerance are deferred capex, not inventory. Steel remaining in a mold is an appraiser-adjacent question. Do not drop a house mold into raw materials to puff the peg.
Stanford's $16 million / 6.2x / ~25% margin median (Source: Stanford GSB) is a search-fund profile with, often, better inventory hygiene. It is never a job-shop multiple and never a reason to skip the sample on a 14-person shop. Census CBP's 68.4% under-20-employee share (Source: Census CBP) is where spreadsheet inventory lives.
IBBA's 83% three-offer statistic (Source: IBBA) does not reward buyers who waive the physical to look easy. It rewards buyers who freeze a method and still close. Fed utilization 76.0% (Source: Fed G.17) does not count a rack.
Close-count roles so the physical is not a brawl
Buyer operations lead. Seller operations lead. Neutral counter if the dollar risk is large. Rule for receipts and shipments after 2 p.m. Rule for jobs that finish during the count. Rule for holds. Rule for customer-owned. Calculator both sides can audit within five days. That is the close intersection with this page.
SBA 7(a) still caps at $5 million with a 10% injection (Source: SBA). A light count is an equity problem. Combined 7(a)+504 can reach $10 million (Source: SBA). Same problem. Main Street 2.75x / ~$268k SDE (Source: BizBuySell via Iconic) feels a $70,000 miss. Stanford $16 million / 6.2x / ~25% margin (Source: Stanford GSB) is a different scale and never a job-shop multiple. Count anyway.
ABL market practice still haircuts WIP harder than finished goods. Do not argue with a field exam using a CIM chart. Argue with travelers. ISM Prices 71.1 (Source: ISM) is why aged raw needs a reserve. Fed utilization 76.0% (Source: Fed G.17) is not a count method. IBBA's 83% three-offer statistic (Source: IBBA) is not a waiver.
Jason Taken is a business broker, not an inventory appraiser. Educational information only. Confirm peg language with counsel and a CPA. No outcome is guaranteed.
Questions the cage should answer before you freeze a peg
Which lots have not moved in the band you would finance. Which WIP jobs cannot finish without the seller. Which finished goods are on quality hold. Which resin, steel, or components are customer-owned. Which molds are house tools versus customer property. Whether the last cycle-count adjustments were material. If those answers are shrugs, the peg is a rumor.
Freeze method only after a sample, not after a CIM chart. Twenty raw lots, fifteen WIP jobs, and the hold rack will tell you more than a trailing-12 average that included hope. If the sample is ugly, size a holdback or walk. Do not "true it later" after the truck is gone.
A shop that adjusts inventory $40,000 every quarter is already telling you the average includes fiction. On a small SDE file that dollar amount is not noise. Write the protocol. Name who counts. Then close on a signature, not on a spreadsheet.
HedgeStone will not argue a peg on a CIM inventory total. Bring a classed file and a sample plan, or do not ask for exclusivity. House tools and customer-owned stock are different columns. Mixing them is how buyers overpay for someone else's property.
How to read cycle-count adjustments as a diligence file
Ask for the last four quarterly inventory adjustments, not just the CIM total. A shop that books a $40,000 swing every quarter is already telling you the perpetual is a rumor. Map each swing to a class: raw, WIP, finished, holds. If the swings sit in WIP, the travelers are lying or the percent-complete method is hope. If they sit in finished goods, the hold rack or the ship-hold list is missing from the CIM. If they sit in raw, either the steel desk is stealing or the receiving process never caught up to ISM-style price moves. Write the adjustment history into the peg memo as a reserve or a holdback. Do not average a year of fiction and call it a trailing-12. On a small SDE file those dollars are a turn. On a larger EBITDA file they are still a credit finding because ABL field exams will haircut the same classes.