Buying a machine shop: diligence on programs, setups and maintenance logs

Which records show whether a machine shop is repeatable?#

The records that show whether a machine shop is repeatable are the ones that let a new owner rerun a past job without guesswork: proven programs, setup sheets, tool data, maintenance history and job costing. Equipment and customer lists describe what the shop owns. These records describe how it actually makes parts.

Search funders and lower-middle-market operating partners usually inherit a founder or lead programmer whose knowledge holds the shop together. Diligence on records measures how much of that knowledge is written down, where it lives and whether it transfers with the business.

How do you test the proven-program library?#

The proven-program library is tested by picking repeat parts from the order history and asking the shop to produce each program, its revision and evidence that it ran cleanly. A library that answers quickly from the CAM system or a DNC server is a transferable asset; one that sends people hunting through controls is not.

Ask who maintains the post processors, whether CAM seats such as Mastercam, Esprit or Fusion are licensed to the company, and whether edits made at the control get saved back to the master file. Edits that stay on the machine are a quiet source of scrap after a change of ownership.

• 1Pull repeat part numbers from the ERP and request the matching programs.
• 2Confirm revision control: one master per part revision, not several drifting copies.
• 3Check that proven programs are marked and that prove-out notes exist.
• 4Confirm CAM license ownership, maintenance status and the vendor's transfer process.
• 5Identify programs stored only on controls or personal devices.

Setup sheets and tool libraries measure key-person risk#

Setup sheets and tool libraries measure key-person risk because they capture the decisions an experienced setup machinist makes without thinking: workholding, work offsets, tool order, stick-out and first-article checks. When those details are written down, a second shift or a new hire can run the job.

Walk the floor with recent jobs and compare each setup sheet to what operators actually did. Gaps are normal; the question is whether the shop updates the sheet when the process changes. A shared tool library tied to programs is a stronger sign than programmers each maintaining their own.

Thin records do not kill a deal. They change the plan: budget for documenting the top repeat jobs early in the ownership period, and build the seller's transition role around that work.

What maintenance logs reveal about the asset list#

Machine maintenance logs reveal whether the equipment on the asset list will perform at the value assigned to it. A dated history of preventive tasks, alarms, crashes, spindle rebuilds and service calls tells you more than an appraisal of make, model and year.

Shops keep this history in a CMMS, in spreadsheets, on paper tags or only in service vendors' invoices. Where logs are missing, reconstruct what you can from invoices and control alarm histories. Look hardest at repeat failures on the machines that run the highest-margin work, because those carry both capital and delivery risk.

Good maintenance records keep their value after close. Linked logs of symptoms, diagnoses and repairs are the operational history AI developers study, which is one more reason to preserve them rather than discard them during a CMMS change.

Quote hit rate and estimated versus actual hours#

Quote hit rate and estimated versus actual hours show whether the shop prices work it can deliver profitably. A shop that wins nearly everything it quotes may be underpricing; one that rarely wins may be estimating from stale rates or chasing the wrong work.

Ask for quotes from the ERP or estimating tool linked to the resulting jobs, and for job costing that captures hours by operation. Shops on systems such as JobBOSS, E2 Shop System, ProShop or Epicor often have this data if operators clock in and out of operations. Where hours are booked only to the job total, you can see overruns but not where they happen.

• 1Which jobs consistently run over estimate, and at which operation?
• 2When were estimating rates last reviewed, and by whom?
• 3How are lost quotes recorded, and is a reason captured?
• 4Which customers drive repeat work that runs on proven programs?

Customer-owned data and export-controlled work#

Customer-owned data and export-controlled work need their own diligence track, because many machine shop records are derived from material the customer owns. Models, drawings and specifications usually belong to the customer, and programs built from them may carry the same confidentiality obligations under purchase order terms or an NDA.

Defense-related work adds federal rules. ITAR defines technical data to include information required for the manufacture, assembly, repair, testing or maintenance of defense articles, including blueprints, drawings and instructions, and BIS describes EAR technology as information necessary for the production, use or maintenance of other controlled items. Whether a record is controlled is assessed with export counsel, and controlled work stays out of any reuse plan.

Illustrative: a search fund evaluates a precision shop#

Illustrative: a fictional search fund is evaluating a precision machine shop that serves industrial equipment makers. The equipment appraisal looks solid, but diligence shows most programs live in the lead programmer's personal folders and maintenance history exists only as vendor invoices.

Programs for the steadiest repeat parts come back quickly from the CAM system; others must be pulled from machine controls. The buyer negotiates a longer seller transition focused on building a revision-controlled library, adds a CMMS rollout to the first-year plan, and confirms that a small defense-related program is segregated under the shop's export compliance procedures.

After close, the cleaned maintenance logs and job costing become documented history owned by the company, available for internal use, a future exit and, where rights allow, licensing.

How SourceX looks at a machine shop's records#

SourceX helps owners of acquired manufacturers assess whether internal records such as maintenance logs, quality decisions and job histories can be licensed to AI developers. The work follows the SourceX five-step transaction, Supply, Rights, Preparation, Approval and Delivery, and the first fit check uses metadata only.

Customer-owned designs and export-controlled work are carved out before any other step. The SourceX Enterprise Data Value Framework weighs drivers such as domain expertise, human-generated signal, recency and rights against preparation cost and privacy burden, so an owner knows which record families matter before anything is prepared.