R&D tax credit for failed projects: a practical UK guide for startups

The Foundational Mindset Shift: It’s About the Attempt, Not the Outcome

Many founders mistakenly believe that R&D tax credits are reserved for successful, market-ready products. This is the single biggest misconception. The UK government provides this relief to encourage companies to take risks and push technological boundaries. Failure is an inherent part of that process. HMRC's focus is not on your commercial success but on your genuine attempt to overcome 'Technological uncertainty'.

Technological uncertainty is the core requirement for any claim. It exists when your team of competent professionals cannot readily or easily know the outcome of a project from the outset. This is distinct from commercial uncertainty. For example, being unsure if customers will buy your new app is a commercial risk and does not qualify. Being unsure if you can build a core feature of that app because it requires a novel data-processing algorithm is a technological uncertainty and likely does qualify.

You might not know if your new algorithm can be developed to perform at the required speed, if a new biotech compound will be stable, or if different software systems can be integrated in a novel way. If you undertook systematic work to resolve that uncertainty, the associated costs are likely eligible, regardless of whether you ultimately succeeded. This is why you can claim R&D tax credit on failed projects; the scheme rewards the valuable learning that comes from the R&D expenditure on unsuccessful projects.

What Costs Still Count When a Project Fails?

When a project is abandoned, it’s easy to view all associated spending as a write-off. However, for the purposes of an R&D claim, many of those direct costs remain eligible. The key is isolating the expenditure directly tied to resolving the technological uncertainty. Indirect costs, like office rent or standard software licences, are not claimable. Here’s a breakdown of what still counts.

1. Staff Costs

This is often the largest part of a claim. You can include a proportion of the gross salaries, employer's National Insurance contributions, and employer's pension contributions for the staff who worked directly on the failed R&D project. For a developer who spent 70% of their time on the abandoned project and 30% on routine maintenance, you can claim 70% of their employment costs for that period. The same applies to a CTO who might split their time between management and hands-on technical problem-solving for the R&D project.

Time tracking data from tools like Jira can be invaluable here, but this is not a strict requirement. The reality for most pre-seed to Series B startups is more pragmatic, and well-reasoned estimates are also acceptable. These estimates can be based on project plans, staff interviews, and calendar entries to create a credible apportionment of time.

2. Subcontractor Costs

If you used external contractors to help solve a specific technical problem within the failed project, a portion of their costs may be eligible. It’s important to note the specific rules here. For instance, "For the SME R&D scheme, claimants can typically claim 65% of payments made to unconnected subcontractors." (HMRC SME R&D Scheme rules). This 65% cap exists because HMRC assumes subcontractors include their own profit margin, which is not a direct cost of R&D. The rules differ for connected parties or for claims under the RDEC scheme, so careful review is necessary.

3. Consumables and Software

This category covers items that were used up or consumed during the R&D process. This is a critical distinction from assets that have a lasting value. A standard laptop is not a consumable because it's a capital asset, but the specialised materials used in an experiment are. The eligibility hinges on the item being transformed or consumed by the R&D activity itself.

• SaaS-specific example: Imagine your team was trying to build a new AI-driven recommendation engine but abandoned the project. The costs for a standard cloud hosting service like AWS would likely not be eligible. However, if you paid for a highly specialised, high-cost machine learning model API specifically to test and build your failed prototype, those API usage fees would be considered a consumable and could be included in the claim.
• Biotech-specific example: A team attempting to develop a new diagnostic assay may run a series of experiments that ultimately prove the method is not viable. The costly reagents, specific cell cultures, and other lab supplies that were completely used up during these unsuccessful experiments are qualifying consumables. The fact that the experiments failed to produce the desired outcome is irrelevant to the eligibility of the costs incurred.
• Deeptech-specific example: A hardware startup attempting to create a new type of sensor might build and test several prototypes. The specialist polymers, electronic components, and rare-earth magnets that were purchased for and destroyed or rendered unusable during the prototyping phase are considered consumables. Their cost is eligible even though the final sensor design was never achieved.

Building a Defensible 'Failure Story' for HMRC

The most critical part of an unsuccessful R&D claim process is constructing a clear and credible narrative for HMRC. If your claim is ever reviewed, an inspector needs to understand the 'why' behind the failure and see evidence of a systematic R&D process. This doesn't require perfectly curated reports, especially for early-stage companies. The evidence is often spread across the tools you already use.

Your goal is a credible, evidence-backed narrative, not a flawless report. Start by gathering your contemporaneous evidence, meaning the records created *during* the project. This is far more powerful than trying to retrospectively justify activities months later.

Useful sources include:

• Project Management Tools (Jira, Linear): Tickets describing technical blockers, different approaches being tested, and comments explaining why a certain path didn't work are excellent evidence. Look for phrases like, "Attempted API integration via method X, but latency was too high. Now testing method Y."
• Communication Records (Slack): Key discussions in dedicated project channels can demonstrate the technical challenges and problem-solving attempts in real-time. A search for keywords like "blocker," "issue," or "failed test" can quickly surface relevant conversations.
• Documentation platforms (Notion, Confluence, Google Drive): Initial project proposals, technical specifications, and meeting notes can establish the project's baseline, its intended advance, and the uncertainties you identified at the start.
• Version Control (GitHub): Commit messages and branch history can provide a timeline of development, showing the iterative process of trying and discarding different technical solutions. A commit message like "Reverting v2 algorithm; memory leak identified under load" is compelling proof of R&D.

This evidence forms the backbone of your 'failure story'. For HMRC, this narrative should clearly articulate:

1. The Starting Point: What was the existing technological baseline in your field before you began?
2. The Goal: What specific scientific or technological advance were you trying to achieve?
3. The Uncertainties: What were the specific technical challenges you couldn't solve easily using existing methods?
4. The Work: What systematic activities and tests did you undertake to try and resolve these uncertainties?
5. The Outcome: Why was the project ultimately abandoned due to technical reasons?

For internal records, a simple project closure memo can summarise this. For instance, an effective summary sentence might be: "Project 'QuantumLeap' was terminated after concluding that the latency issues with the v3 algorithm were insurmountable with current processing architecture, failing to meet the sub-10ms response time required for market viability."

From Sunk Cost to Cash Injection: Estimating Your Claim

Once you’ve identified the qualifying R&D expenditure on unsuccessful projects, you can forecast the potential cash benefit. This is vital for managing runway. The calculation below applies to a typical loss-making SME for expenditure incurred from 1 April 2023 under the main R&D scheme. Note that rates for profitable SMEs, R&D intensive companies, and those claiming under the RDEC scheme will differ.

Let’s walk through a common scenario for a SaaS startup. The project was abandoned after six months.

• Qualifying Spend:
  • Developer Salaries (apportioned): £50,000
  • Specialised API Consumables: £10,000
  • Total Qualifying R&D Spend: £60,000

Here is the step-by-step calculation for the payable credit:

1. Calculate the Enhanced Deduction: The R&D scheme allows you to enhance your qualifying spend for tax purposes. "For expenditure from 1 April 2023, the 'Enhanced Deduction' for loss-making SMEs is 86% of qualifying R&D spend." (HMRC R&D Scheme rules).
2. £60,000 (Qualifying Spend) * 86% = £51,600
3. Determine the Total Relievable Loss: This is your initial spend plus the enhancement. You surrender this total loss to HMRC in exchange for a cash credit.
4. £60,000 (Qualifying Spend) + £51,600 (Enhanced Deduction) = £111,600
5. Calculate the Payable Tax Credit: A cash payment is calculated based on this surrendered loss. "For expenditure from 1 April 2023, the payable credit rate for the 'Total Relievable Loss' for loss-making SMEs is 10%." (HMRC R&D Scheme rules).
6. £111,600 (Total Relievable Loss) * 10% = £11,160

In this example, the £60,000 sunk cost from the failed project generates £11,160 in cash back for the company. It's also worth noting that for companies spending a very high proportion of their total costs on R&D, a more generous 'R&D Intensive' scheme exists. "The payable credit rate for the R&D Intensive scheme is 14.5%." (HMRC R&D Scheme rules).

Practical Takeaways for Your Next Project

Thinking about R&D tax relief for abandoned projects shouldn't be an afterthought. It should be part of your financial and operational strategy from day one. Here are immediate, practical steps you can take.

• Look Backwards First: You have a limited window to make a claim. "Claims can be made for projects from the last two completed accounting periods." Review your abandoned projects from this period now to identify potential missed opportunities for non-dilutive funding.
• Document As You Go: You don't need a complex system. For your next R&D project, create a dedicated Slack channel or Notion page. Post weekly updates on progress, blockers, and decisions. This simple habit builds a powerful, contemporaneous evidence trail that is far more credible to HMRC than a retrospective report.
• Tag Your Costs in Real-Time: In your accounting software like Xero, use tracking categories or project codes. Tag R&D-specific staff time, subcontractor invoices, and consumable costs as they happen. This makes calculating your qualifying spend at the end of the year significantly easier and more accurate.
• Conduct Post-Mortems with R&D in Mind: When a project is shelved, hold a post-mortem that explicitly discusses the R&D narrative. Document the initial technical goals, the uncertainties faced, the approaches tested, and the specific technical reasons for termination. This meeting's minutes can become a key part of your claim's supporting evidence.
• Plan for the Process: Remember that "Claims are filed with the Company Tax Return (CT9)." Once submitted, the timeline for receiving the cash is generally predictable, as "HMRC typically processes R&D tax credit payments within 40 days of submission." (HMRC guidance). Factor this timing into your cash flow projections.

Ultimately, every failed project contains valuable lessons. The UK R&D tax credit scheme ensures it can contain valuable cash as well, providing a vital lifeline for your next attempt at innovation. See the R&D claim documentation hub for preparing your submission.

Frequently Asked Questions

Q: What if only part of our project failed?
A: You can still claim for the R&D expenditure on the unsuccessful parts. A claim can be focused on specific workstreams within a larger project. The key is to isolate the costs associated with the activities undertaken to resolve specific technological uncertainties, even if the overall project was a commercial success.

Q: Does the reason for failure matter to HMRC?
A: Yes, it is critical. The project must have been abandoned for technical reasons, not commercial ones. For example, if you stopped because the technology did not work as intended, that qualifies. If you stopped because a competitor launched first or you couldn't find a market, that does not qualify.

Q: Is there a minimum spend to claim for a failed project?
A: No, there is no minimum R&D expenditure required to make a claim. Eligibility is based entirely on whether the project activities meet the definition of R&D, specifically the presence of technological uncertainty and a systematic process to resolve it, not on the amount of money spent.

Q: How long after a project is abandoned can we still make a claim?
A: You can claim R&D tax relief for costs incurred in your last two completed accounting periods. This means you must submit the claim within two years of the end of the accounting period in which the R&D expenditure on the failed project took place. It is a strict deadline.