The 5 Biggest Challenges When Making an R&D Tax Relief Claim
HMRC scrutiny is higher than ever, and claim errors can trigger costly enquiries. Here are the five most common R&D claim challenges — and how to avoid them.
Making an R&D tax relief claim is not technically complicated, but it is surprisingly easy to get wrong. Since 2022, HMRC has significantly increased its compliance activity, and claims that would have passed without comment five years ago are now routinely subject to enquiry. The cost of an enquiry — in management time, adviser fees, and uncertainty — can easily exceed the value of a modest claim. Understanding where claims go wrong is the first step to making sure yours does not.
Here are the five challenges that cause the most problems for UK businesses.
1. Identifying Qualifying Activities
The first and perhaps most fundamental challenge is simply working out what qualifies. The HMRC definition of qualifying R&D — advancing science or technology by resolving a scientific or technological uncertainty — is precise, but it does not map neatly onto the way engineers describe their own work.
When you ask a software engineer whether they did R&D last year, the answer is almost always “no, we just built features”. Ask the same engineer to describe a time they had to solve a problem that had no obvious solution, tried multiple approaches before finding one that worked, or built something for which there was no available library or framework — and suddenly a very different picture emerges.
The challenge is translating technical reality into HMRC criteria, and the risk runs in both directions: companies that claim too broadly (including work that is just standard professional practice) and companies that claim too narrowly (leaving genuinely qualifying work off the table entirely). Both are costly errors.
Menlo addresses this by analysing activity directly in your source control history. Rather than relying on engineers to self-identify qualifying work from memory, Menlo maps actual commit activity and project history to HMRC’s criteria — surfacing qualifying work that would otherwise go unnoticed, while filtering out activity that does not meet the threshold.
2. Writing the Technical Narrative
HMRC requires a written technical narrative for every qualifying project included in a claim. This narrative must explain: what scientific or technological advance was sought, what the specific uncertainties were, why those uncertainties could not be resolved using existing publicly available knowledge, and how your team worked to overcome them.
In practice, most technical teams find this genuinely difficult. Engineers are not trained to write in HMRC’s register. The instinct is either to be too brief (“we built a machine learning model”) or to produce highly technical documentation that an HMRC inspector cannot evaluate. Neither approach serves the claim well.
A good technical narrative sits in the middle: specific enough to demonstrate genuine technical substance, accessible enough that a non-specialist reviewer can understand what was attempted and why it was non-trivial. Getting this balance right is a skill that takes time to develop, and the stakes are high — a vague or unconvincing narrative is one of the most common triggers for an HMRC compliance check.
Menlo auto-generates structured technical narratives from your repository data. The output is designed specifically to meet HMRC’s requirements — covering advance, uncertainty, and resolution in clear, reviewable prose. Your team reviews and approves rather than writing from scratch.
3. Increased HMRC Scrutiny and Enquiries
HMRC’s approach to R&D compliance has changed materially since 2022. The volume of compliance checks roughly doubled between 2021/22 and 2023/24, driven by a significant increase in the number of erroneous or fraudulent claims during the pandemic years. HMRC has invested heavily in specialist R&D compliance resource and is applying that resource across claims of all sizes — not just the very large ones.
An HMRC enquiry into an R&D claim is not necessarily a disaster — if the claim is well-documented and defensible, it will be resolved in your favour. The problem arises when a claim has been prepared carelessly: vague project descriptions, costs that are difficult to reconcile, narratives that do not clearly explain the qualifying activity. In those cases, enquiries drag on, require significant management and adviser time, and often result in claims being reduced or withdrawn.
The best protection against an enquiry is a claim that is so clearly documented it is not worth HMRC’s time to challenge. Clear project descriptions, a precise mapping of costs to qualifying activities, and a well-evidenced technical narrative all contribute to that outcome.
4. The Cost of Traditional Consultants
The traditional R&D tax advisory market is dominated by consultants who charge on a contingency basis: typically 15–25% of the refund or tax reduction. At the lower end of the market, this fee structure makes economic sense — the consultant’s incentive is aligned with the size of the claim. In practice, however, it creates several problems.
First, the cost is simply very high. On a £60,000 claim, a 20% contingency fee costs £12,000 — money that reduces the net benefit to the business significantly. For smaller claims, this can make the whole exercise borderline uneconomical. Second, the percentage model creates an incentive for consultants to maximise the gross value of the claim, which is not always the same thing as creating a well-evidenced, defensible one. Third, the process typically involves significant time from your engineers, who are interviewed at length to generate the raw material for narratives that the consultant then writes.
Menlo charges a flat fee, not a percentage of your claim. The cost is fixed and transparent at the outset, so the full value of your relief stays with your business. See our pricing page for details.
5. The Time Burden on Engineering and Finance Teams
Even when a business is fully committed to making a claim, the practical burden of preparing one falls on the people who are least able to absorb it. Engineering leads are pulled into lengthy interviews to explain what they worked on months ago. Finance teams spend hours reconciling payroll records, apportioning staff time, and pulling together subcontractor invoices. Senior managers review and approve documents that took weeks to produce.
This time cost is real and it recurs every year. It is one of the reasons many businesses — particularly those with limited finance resource — decide not to claim despite being clearly eligible. The effort is perceived as disproportionate to the return, especially if the claim is modest.
Menlo was designed specifically to address this. By pulling data directly from source control — where the evidence of qualifying activity already exists — Menlo dramatically reduces the amount of manual input required from your team. What would previously take weeks of preparation time is compressed into a structured review-and-approve workflow.
Menlo Was Built to Solve These Problems
Every one of the five challenges above is a direct consequence of trying to prepare an R&D claim using tools and processes that were not built for the job: spreadsheets, interviews, manual narrative writing, and percentage-based consultants. Menlo takes a different approach — starting from the source of truth (your codebase and engineering activity) and using that data to automate the parts of the process that are most burdensome. The result is a faster claim, a more defensible claim, and a substantially lower cost.