The R&D tax credit most small companies miss

Most small companies that qualify for the federal R&D tax credit never claim a dollar of it. Not because their work fails the test, but because they assume the credit belongs to someone else — a drug company, a chip designer, someone in a white coat. So they leave a five- or six-figure benefit on the table every year, quietly funding the government for work they were already doing.

The R&D tax credit is a dollar-for-dollar reduction in the tax you owe, granted for spending on qualified research. It has been in the code since 1981 and was made permanent in 2015. It is not a deduction, which only lowers taxable income. A credit cuts your bill directly, which makes it far more valuable per dollar.

It is not just for people in lab coats

The most expensive myth about this credit is that it requires a laboratory. It does not. The law rewards the development of new or improved products, processes, software, formulas, and techniques. In practice that sweeps in a large amount of ordinary work at ordinary companies.

If your engineers write software, build a feature that forces them to figure out how to make it work, design a manufacturing process, or build a prototype and iterate on it, you are likely doing qualified research. A company writing custom software to run its inventory, a firm designing a new industrial fixture, a food producer reformulating a recipe to extend shelf life — all of these routinely qualify. The question is never how impressive the work sounds. It is whether the work meets a specific four-part test.

The four-part test

To qualify, an activity has to clear all four of these hurdles. Treat them as a checklist you apply to each project, not to your company as a whole.

• Permitted purpose: the work aims to create a new or improved product, process, or software — better function, performance, reliability, or quality.
• Technological in nature: the work relies on a hard science such as engineering, computer science, physics, chemistry, or biology.
• Elimination of uncertainty: at the start you did not know whether you could achieve the result, how to achieve it, or the right design to use.
• Process of experimentation: you evaluated alternatives through testing, modeling, or trial and error to resolve that uncertainty.

Most legitimate product and software development clears this bar without anyone stretching. If your team ever wrote a feature, hit a wall, tried three approaches, and shipped the one that worked, that is a process of experimentation resolving technical uncertainty. That is the credit.

The payroll-tax offset that changes the math for startups

Here is the part that catches founders off guard. The usual complaint is fair: we are pre-revenue and losing money, so we owe no income tax, and a credit against income tax is worthless to us. For years that was true. It is not anymore.

A qualified small business — broadly, one with less than $5 million in gross receipts and no receipts more than five years old — can elect to apply the R&D credit against the employer portion of its payroll taxes instead of income tax. The cap is up to $500,000 per year, raised from $250,000 for tax years beginning after 2022. That means a company burning cash and paying salaries can convert its research spending into real payroll-tax savings, quarter after quarter. For a startup, that is not a rebate you wait years for. It is runway.

What it is worth, and what you have to prove

The benefit typically lands in the high single digits as a percentage of qualified spending. Under the simplified method most small companies use, the federal credit works out to roughly 6% to 10% of qualified research expenses, depending on the calculation and your history.

Qualified spending falls into a few buckets: wages for employees doing or directly supervising the research, supplies consumed in the process, and 65% of amounts paid to outside contractors for qualifying work. Cloud computing used to run the research can also count. What you have to document is straightforward but real:

• Time allocation — what share of each qualifying employee's hours went to research, backed by records rather than a guess.
• Project records — design docs, tickets, test results, and version history that show the experimentation actually happened.
• A clean tie from qualifying wages, supplies, and contract research back to your payroll and general ledger.

A worked example. Say you run a 12-person software company. Four engineers spend about 70% of their time building genuinely new functionality, and their combined salary is $600,000. That puts roughly $420,000 of wages into the qualified pool. Add $30,000 of qualifying cloud infrastructure and $50,000 paid to a contract developer, of which 65% — $32,500 — counts. Your qualified base is about $482,500. At an 8% effective rate, that is a federal credit near $38,600. As a qualified small business you can take that against payroll taxes and feel it inside the year.

Work with a specialist, not a hunch

This is one area where doing it yourself is a false economy. The rules on which activities and costs qualify are detailed, the IRS scrutinizes aggressive claims, and the documentation standard has tightened. A specialist — often working alongside your regular accountant — will scope the study properly, defend the numbers, and usually surface more qualified spend than you would find alone. The credit is generous by design. The mistake is assuming it was written for someone else.