You call it recipe testing. The statute calls it a process of experimentation.
Formulation, shelf-life, scale-up, and packaging development can clear the four-part test, once it is separated from preference testing and routine quality control.
Photo by elevatebeer on Unsplash
A formulation is a formula; a production route is a process. Both are named in the statutory list of business components, so both are eligible subjects in their own right - the process qualifies even when the product coming off the line is unchanged.
The work qualifies when the information available to you did not establish whether the result could be achieved, or how, and you resolved that through a systematic process of experimentation: designing alternatives, running them, measuring, and changing the formula or the parameters on the result. That is what a development log already records, and it is tested against the IRC §41 four-part test activity by activity.
This industry underclaims for a specific reason: the vocabulary. “Recipe testing”, “trial run”, and “panel” do not sound like research, so the hours and the ingredients never reach a preparer. Meanwhile the one thing that genuinely is excluded - asking consumers which version they prefer - happens in the same building, which makes the line worth drawing carefully. See documentation & substantiation for what the record needs to carry.
General and illustrative; eligibility depends on your facts.
Not activity categories - situations. If any of these read like a project you actually ran, there is very likely a qualified business component underneath it.
A supply change or a label commitment removed a functional ingredient. Replacement systems were built and tested for texture, stability, and process behaviour, and two were abandoned on the data.
Why it can qualify
Functional uncertainty about whether and how the component could be made to perform, resolved by evaluating alternatives.
Shelf-life and stability trials were run across formulations, packaging, and storage conditions until the failure mode was understood and designed out.
Why it can qualify
Improving reliability and quality is a permitted purpose, and a designed trial series is a process of experimentation.
Shear, heat load, hold times, and equipment geometry changed the product. Process parameters were developed and re-run at pilot scale before the formula was released.
Why it can qualify
Scale-up uncertainty about the process, which is its own business component under §41(d)(2)(B).
A new thermal, high-pressure, or aseptic route was developed to reach the same quality target, and the process window had to be established rather than looked up.
Why it can qualify
Method uncertainty in the physical and biological sciences, resolved by systematic trial rather than by supplier guidance.
Barrier, seal integrity, and migration behaviour were tested across materials and formats, and the structure was changed on the results.
Why it can qualify
Packaging development that resolves a technical performance question is qualified research; a graphics refresh on the same structure is not.
Changeover sequence, cleaning regime, and line parameters were trialled against scrap and giveaway data until the process held at rate.
Why it can qualify
Process development on a business component you never sell. Improving how you make it is squarely inside the statute.
A compliant system had to be developed from scratch, with functionality and sensory performance re-established through repeated bench and plant trials.
Why it can qualify
A real constraint that the available information did not tell you how to meet, resolved experimentally.
Illustrative situations, not client work. Whether any of them qualifies for you depends on your facts, your agreements, and your evidence.
Development groups in this industry keep better records than they think: formulation sheets with version numbers, trial protocols, stability pull schedules and their results, plant trial run cards, and the batches that were destroyed and why. Read as evidence rather than as housekeeping, that is a business component narrative already half written.
What usually has to be rebuilt is the money. Trial ingredients disappear into cost of goods, plant trial hours into production labour, and the pilot line into a capital project. Tying those back to the trials they belong to is where most of the recoverable value on this vertical actually sits.
Representative activities we see meet the four-part test across product development, process engineering, and packaging.
What spending counts toward the credit - and why the supply line matters more here than in almost any other industry.
| Expense category | What goes into the base |
|---|---|
| Technical wages§41(b)(2)(A)-(B) | W-2 wages for food scientists, process and packaging engineers, QA technicians on development work, and the plant staff whose time directly supports a trial. |
| Supplies§41(b)(2)(C) | Ingredients, packaging, and materials used and consumed in bench, pilot, and plant trials - including product that cannot be sold because it was a trial. Not depreciable equipment. |
| Contract research (65%)§41(b)(3) | 65% of amounts paid to U.S. contract laboratories, pilot facilities, and development partners for qualified research performed on your behalf under a pre-existing agreement. |
| Computer rental§41(b)(2)(A)(iii) | Amounts paid for the right to use computers in qualified research. Usually a small line here, and software seats are not a §41 expense category. |
A directional shape for a development-active manufacturer, not a benchmark. The supply line is unusually large here, and it is also the line most often left out entirely.
Pilot equipment, tanks, and line hardware you keep are property subject to depreciation, so they are outside the supplies category however central they are to development. What is used up in a trial - ingredients, packaging, the batch that went to waste - is the part that belongs in the base.
The constraint is almost always records rather than law. Trial runs need to be separable from production runs in whatever system the plant already keeps: a work-order type, a trial batch code, a cost centre. If nobody can tell the two apart afterwards, an examiner will make the same observation.
Food and beverage claims fail on a short and predictable list, and the first item on it happens in the same building as the qualifying work.
§41(d)(4)(D)
Market research, consumer preference panels, sensory testing run to choose between options that all work, and efficiency or management studies are expressly excluded. The distinction is the question being asked: “will this system deliver the texture at all” is technical, “which of these three do shoppers like more” is not. Trained descriptive panels used to measure a technical outcome sit closer to the first question, and the record has to show which one was being answered.
§41(d)(4)(A), (B)
Once a product is in commercial production, work on it is outside the credit. So is adapting an existing product or process to a particular customer’s requirement - a private-label version of a formula you already make, a pack size change, a flavour line extension using a system you have already proved.
§41(d)(4)(D)
Ordinary testing and inspection for quality control, routine data collection, and running an established HACCP or monitoring programme are excluded however technical the instruments are. Developing and validating a genuinely new control step is a different activity and is a candidate.
§41(d)(3)(B)
Research that relates to style, taste, cosmetic, or seasonal design factors is outside the permitted purposes. A new label, a holiday pack, a colour change, or a flavour tweak chosen on preference with no technical problem underneath it is not research. Holding a flavour profile while solving a technical constraint, such as a sugar or sodium reduction that breaks the formula, is a different question, analysed on its own facts.
§41(d)(4)(F), (H)
Development a customer paid for is funded research, and excluded, unless you both bore the risk of failure and kept substantial rights in the results. Research performed outside the United States is excluded whatever the arrangement. Both cut in either direction depending on which side of the co-man relationship you sit on, which is why the agreement gets read first.
None of these is a reason not to claim. They are reasons to separate the trial from the panel, and the development batch from the production batch, before the number is filed - which is what Audit Protection and a documented substantiation file are for.
Food and beverage manufacturing is plant-bound, which is exactly the profile states legislate for. Most states with a corporate income or franchise tax run their own R&D credit, each with a different base, rate, cap, carryforward, and refundability - and a federal number multiplied by a state rate is not a state credit. Ricerca computes both from one substantiated research base.
If you are a newer producer, test §41(h) as well: a qualified small business may elect to apply up to $500,000 of credit per year against payroll taxes rather than income tax. It turns on gross receipts under $5 million in the credit year and on not having had gross receipts before the five-year window ending in that year, it is claimed on Form 8974 with your quarterly employment tax return, and it cannot be made for more than five tax years.
Summary only - the statutory definition and the aggregation rules decide it. We test them explicitly.
A hypothetical scenario to show how the pieces fit together. It is not a quote, projection, or promise of results.
Plus the full §174A first-year deduction on domestic research and experimental costs.
Illustrative only. Figures are hypothetical and rounded; the federal credit commonly works out to roughly 6-10% of QRE depending on method, filing history, and the §280C election. Your result depends entirely on your facts, and on whether your trial spend can be separated from production. This is not a quote or a promise of any result.
IRC §174A restores immediate, full expensing of domestic research and experimental costs for tax years beginning after December 31, 2024. Its definition is broader than the §41 credit’s, so development spend that does not clear the four-part test can still qualify for the deduction. A study run properly captures both.
Tell us what you formulate and how you trial it, and we’ll map the qualifying development to the four-part test - separating trial spend from production, and capturing §174A - reviewed and finalized by R&D experts and backed by Audit Protection. Contact us for pricing tailored to your study.
We typically reply within one business day.
