The two tests at the core
The definition of R&D for tax relief comes from the DSIT guidelines, the statutory guidelines that replaced the older 2004 version and apply to accounting periods beginning on or after 1 April 2023. Everything else builds on two requirements that have to be met together.
First, the project must seek an advance in science or technology. Second, it must do so by resolving scientific or technological uncertainty. An advance with no uncertainty is not R&D. Uncertainty with no attempt at an advance in the field is not R&D either. Both, together, in the same project.
An advance in science or technology
New knowledge or capability for the field, not just for your company
Scientific or technological uncertainty
A competent professional could not readily work out the answer
“Advance in science or technology”
An advance means an increase in overall knowledge or capability in a field of science or technology, not an increase in your own company’s knowledge alone. This is the test that catches the most claims out. Building something that is new to you, but that a competent professional elsewhere could already build, is not an advance in the field. It is catching up.
The advance can take different forms. It can extend knowledge, create a process or product that represents an increase in the field’s capability, or make an appreciable improvement to something that already exists through scientific or technological change. It can also include adapting knowledge from another field where doing so was not readily deducible. The outcome can be tangible, such as a more efficient process, or intangible, such as new knowledge, and a project still counts even if it fails to achieve the advance it was aiming for.
“Scientific or technological uncertainty”
Uncertainty exists when the answer is not readily available in the public domain and not readily deducible by a competent professional working in the field. That covers both whether something is scientifically possible or technologically feasible, and how to achieve it in practice.
The word doing the heavy lifting is “readily”. A problem you can solve by looking up published information, by buying in a known specialist solution, or by routine work, is not a qualifying uncertainty, however much effort it takes. Improvements, optimisations and fine-tuning that do not materially change the underlying science or technology do not count either.
The “competent professional” test
Both tests above are judged through the eyes of a competent professional. That means someone who knows the current baseline in the field, by qualification or experience. They can say whether a genuine advance was being sought, and whether the uncertainty was real rather than routine.
This is why a credible claim needs the right person to stand behind the technical case. If a competent professional in your field would say the answer was already known or readily worked out, the claim does not hold, regardless of how the project felt internally.
It has to be a project, and planned
R&D for tax purposes happens within a project: a set of activities following a method or plan to achieve the advance. Getting the boundaries of that project right matters. A commercial project can contain qualifying R&D and a great deal that is not. Only the parts that directly contribute to resolving the scientific or technological uncertainty (plus certain qualifying indirect activities) belong in a claim.
The work also has to be deliberate. R&D that seeks to resolve a known uncertainty in a planned way is very different from stumbling on something by accident and badging it as R&D afterwards. Describing a project as “trial and error” tends to signal the latter, and HMRC often asks for records created at the time to show the work was planned and methodical.
When your work is not R&D
The honest part, and the part most adviser content skips. Plenty of genuine, skilled, valuable work is not R&D for tax purposes. The following are not qualifying activities in themselves:
The answer was already out there
Work whose solution was readily available or readily deducible by a competent professional.
Cosmetic and aesthetic changes
Changes that do not depend on a scientific or technological advance.
Taking a product to market
The commercial and financial steps of bringing a product to market.
Production and distribution
The production and distribution of goods and services.
Admin and support functions
Business administration and general support functions.
Routine use of existing technology
Applying existing technology in a routine way, even if it is new to your business.
If your project sits here, the right answer is that there is no claim. Being told that early, by someone who will say it, is worth more than a claim that falls over under scrutiny.
Frequently asked questions
Does my work qualify if it was difficult and took a long time?
Not on its own. Difficulty is not the test. The question is whether you were seeking an advance in the field and resolving an uncertainty a competent professional could not readily resolve. Hard routine work is still routine work.
We built something new to our business. Is that R&D?
Possibly not. The advance has to be new to the field, not just to you. If a competent professional elsewhere could already do it, you were catching up, which does not qualify.
Does it count if the project failed?
Yes. A project that genuinely sought an advance and addressed real uncertainty can qualify even if it did not succeed. Failure does not disqualify a claim; absence of a real advance or uncertainty does.
Does HMRC approve my claim when it pays it?
No. Payment is not approval. HMRC can and does check claims after paying them. A claim being paid tells you nothing about whether it was correct.