Yes. R&D tax relief rewards the attempt to resolve scientific or technological uncertainty, not the outcome. A project that never reached a working result can still qualify in full, and the legislation has always worked this way: the test is whether the work sought an advance through resolving uncertainty that a competent professional could not readily resolve, not whether the advance was achieved.
Why failure can strengthen a claim
Failed iterations are often the clearest evidence that the uncertainty was genuine, but the inference only runs one way. Failure on its own qualifies nothing. A project can miss its target because the money ran out, because the specification moved, or because the work was done badly, and none of those is a scientific or technological uncertainty. What decides the claim is why the work failed.
A prototype that missed its performance target because the underlying behaviour could not be predicted, a process that could not be scaled because it behaved differently at volume, an approach abandoned once testing showed the route would not work: each records an attempt to resolve something a competent professional in the field could not readily answer, which is what HMRC needs to see. Where that is the case, we find the abandoned branches of a project are frequently easier to defend than the branch that eventually worked, because nobody can argue the solution was obvious.
The same logic applies within successful projects. Most R&D that ends in a working product passes through failed attempts on the way, and those attempts are part of the qualifying activity. Leaving them out of a claim both understates the qualifying work and weakens the story of uncertainty the claim depends on.
What you need to be able to show
The relief attaches to the attempt, but the attempt has to be evidenced. The Guidelines put it directly: even where the advance sought is not achieved or fully realised, the R&D still takes place — what matters is being able to describe what was attempted, why existing knowledge was insufficient, what was tried, and what was learned, including from the failures. The Additional Information Form requires a project description covering exactly this ground, and HMRC checked around one in six claims in 2023-24, its latest published figure, so the description needs to hold up under questioning.
Records do not need a statutory format. HMRC’s guidelines for compliance accept that some R&D costs will be an estimated proportion of known expenditure, provided the estimate is arrived at using evidence and reason, and that where nothing was written down at the time a detailed explanation provided later may be acceptable in some cases. That latitude is not a first-claim concession — HMRC publishes no such thing, and it applies to a fifth claim as much as a first. Nor is it a reason to leave records to the year end: keep project documents, test results and staff time records as you go, because evidence created at the time the work was done is far more persuasive than reconstruction after the event.
Commercial failure and technical failure also pull in different directions here. A product that flopped in the market may rest on work that qualifies in full, and a project that ran into serious difficulty may qualify for none of it. Only the statutory definition settles which you have, and our guide to what counts as qualifying R&D sets out where the line falls.
Where to go next
If you shelved a project and assumed the spend was lost, it may still be claimable: the window for amending a return generally runs two years from the end of the accounting period. We will tell you honestly whether the work qualifies before any claim is prepared.
Sources
- Guidelines on the meaning of R&D for tax purposes — paragraph 10: R&D still takes place even where the advance sought is not achieved; paragraphs 6 and 13, the advance and uncertainty tests that decide whether a failed project qualifies at all.
- Check if a project qualifies as R&D for tax purposes — the definition the outcome does not change.
- GfC3: Recommended approach to claims and record keeping (part 5) — estimates arrived at using evidence and reason, and explanations provided later where written records do not exist; the latitude is not limited to first claims.
This page describes the rules as they stood at the review date above, as general information rather than advice on your circumstances. For how that distinction works, see our terms; for an answer on your own facts, talk to us.