Is This Job Worth Tracking? A Four-Question Sorting Tool for Custom Manufacturers

TL;DR

• Most custom jobs are good, routine work. A smaller number may involve a real technical unknown, and those are the ones worth a closer look for the R&D credit.
• Four questions can help you sort a job. Was the method or capability uncertain? Did you test more than one approach? Who carries the risk and owns the result? Had the job moved into routine production?
• Sorting jobs early is not the same as claiming them. It helps you know where to look when it is time to size a credit.

The problem with “we do custom work, so we must qualify”

Owners of custom manufacturing shops tend to land in one of two camps. Some assume that because every job is a little different, everything qualifies. Others assume that because the shop is “just making what the customer drew,” nothing does. Neither assumption holds up well.

The federal research credit under IRC Section 41 is built around activity, not around industry or job type. Under Treasury Regulation 1.41-4, research generally has to be aimed at eliminating uncertainty about the capability or method of developing or improving a product or process, and it has to involve a process of experimentation. A job can be unusual without being uncertain. A job can also look ordinary from the outside and still contain a stretch where your team genuinely did not know how to hold a tolerance, fixture a part, or get a material to behave.

That is why a sorting tool is more useful than a blanket rule. The four questions below are not a legal test, and a qualified professional would look at the full facts. They are a practical way to decide where your attention is best spent.

Question 1: At quote time, did you know how you would make it?

Think back to when the job was quoted. If your team looked at the drawing and knew the machines, the fixturing, the toolpaths, and the material behavior, the job is likely routine for your shop, however complex the part. Skilled execution of a known method generally does not involve the uncertainty the regulation describes.

If instead the team said “we think we can hold that, but we are not sure how,” the job may be worth a closer look. The regulation looks to whether the information available to you did not establish the capability or the method. Examples that often prompt this question in a shop include a thin-wall geometry that moves when cut, a tolerance stack that existing fixturing cannot hold, a material or coating the shop has not worked with before, or a process change meant to hit a cycle time that current methods cannot reach.

Question 2: Did you try more than one way?

Uncertainty alone is not enough. The next question is what your team did about it. A process of experimentation generally means evaluating one or more alternatives to reach a result: a different fixture design, a revised toolpath or feed and speed strategy, an alternate material lot or heat treatment, a changed sequence of operations.

A useful gut check is to ask whether anyone would have a scrap tag, a revised program, a modified fixture, or a note that says “tried X, did not work, moved to Y.” Teams that evaluated alternatives often leave this kind of trail without trying. Teams that picked a method and ran it start to finish usually do not, and that is a fair signal the job was execution rather than development.

The regulation does not require the effort to succeed. A fixture concept that failed and was replaced can still be part of the story, depending on the specific activities and documentation.

Question 3: Who carried the risk, and who owns the result?

This is the question many shops skip, and it can matter a great deal. Research that is funded by another person is generally excluded from the credit, and whether a customer-paid job counts as funded depends on contract terms, such as whether payment was contingent on the work succeeding and who holds the rights to the results. Those rules are separate from the activity test and are applied contract by contract.

You do not need to resolve this yourself. You do need to know which jobs have unusual terms. As a practical step, flag any job where the customer paid a flat engineering fee regardless of outcome versus any job where payment depended on the part passing, and note who owns the tooling, programs, and drawings. A short list like that makes any later review faster.

Question 4: Had the job moved into routine production?

Activities conducted after the beginning of commercial production of a business component are generally not qualified research. For a shop, that line often sits somewhere between the first article and the release to repeat production. Work to figure out how to make the first good part may be a different category from running the hundredth identical part.

Ask when the job was released to production, and separate what happened before that date from what happened after. Wages and supplies tied to the earlier stretch are the ones most likely to matter. Cost reductions and routine improvements after release generally fall on the other side of the line.

Putting the four questions together

You can score each job by how many questions point toward “worth a closer look.”

• Three or four signals. The job had a real unknown, you tried alternatives, the contract does not obviously exclude it, and the activity was before routine production. These jobs are the best candidates to keep records on and discuss in a review.
• One or two signals. Something was uncertain, but the trail is thin or the contract is unclear. Worth noting, and worth asking a specialist about before you decide.
• Zero signals. Good, profitable work that is probably routine for your shop. There is no need to spend effort here, and that is a fine outcome.

Two cautions. First, size does not decide this. A small job with a genuine unknown can matter more than a large job that ran on a known process. Second, the credit is calculated from qualifying wages, supplies, and certain contract research tied to qualifying activity. The sorting exercise helps you find the activity, and a full study is what supports the numbers.

What this takes from your team

Sorting jobs this way should not turn your floor into a paperwork department. The questions can be answered from information your team already has: quotes, programs, scrap tags, and the memory of the lead machinist or engineer. Strata runs a structured, guided process in which most of the work sits with us rather than with your team, and a typical engagement runs four to eight weeks depending on the specific activities, documentation, and facts. At the end, you receive a documentation package built to support the claim, which you then hand to your CPA, who files it with the return. Individual results vary.

Where to go from here

If you want a quick view of which of your jobs deserve attention, our 15-minute review is a good place to start, and the $0 initial assessment can help you see whether a credit is likely to be meaningful for your shop before you commit to anything. Book your 15-minute review at stratataxgroup.com/contact.

For more on how the credit is defined, see IRC Section 41, Treasury Regulation 1.41-4, and the IRS instructions for Form 6765.

This post is for informational purposes only and does not constitute tax or legal advice. Consult a qualified tax professional regarding your specific circumstances.