What Should Small Companies Consider Before Making an R&D Tax Credit Claim?

If you're running a small company and thinking about your first R&D tax credit claim, or you've claimed before but never really stress-tested the process, a handful of decisions made before you file are what separate a straightforward claim from one that gets capped, trimmed, or rejected outright. This guide walks through those decisions in the order they tend to matter.

Which scheme applies to you: merged RDEC or ERIS?

Since April 2024, the old SME scheme and RDEC have been replaced by a single merged scheme: a 20% gross expenditure credit, worth around 15p per £1 of qualifying spend once tax is accounted for. That's the default for every company.

But if you're loss-making and R&D-intensive (your qualifying R&D expenditure is at least 30% of your total expenditure for the period), you can elect to claim through the Enhanced R&D Intensive Support scheme (ERIS) instead. ERIS pays a 14.5% credit on your surrenderable loss, which works out as a materially higher cash return than the merged scheme.

If you meet the criteria for it, use ERIS. There's little reason for an eligible loss-making small company to settle for the merged scheme's lower rate.

Have you filed the Advance Notification Form?

Before any of the above can matter, first-time and lapsed claimants need to clear an earlier hurdle: the Advance Notification Form (ANF). If your company has never claimed R&D tax relief before, or hasn't claimed in the last three years, you need to tell HMRC within six months of the end of the accounting period that you intend to claim.

This is a common trip-up for small companies making their first claim, precisely because it falls so early, months before the claim itself is even close to ready. There's no exemption in play, and the claim simply cannot be made for that period if you miss the deadline, regardless of how strong the R&D was. Check this as soon as your accounting period ends, not when you sit down to prepare the claim.

Are your projects actually qualifying R&D?

Innovative isn't the same as qualifying. Small companies often assume that because something is new to them, building a first version of a product, solving a problem no one on the team has solved before, it automatically counts. HMRC's test is narrower: the work has to resolve genuine scientific or technological uncertainty and produce an advance in the field, not just deliver something new to your business.

R&D begins when work to resolve the uncertainty starts, and ends when that uncertainty is resolved or work on it stops, not when the wider commercial project wraps up. Get this wrong and every number downstream, apportionment, cost tracking, the cap, is built on the wrong base.

Company A is building a new inventory management platform. Working out how to reconcile stock data across three incompatible legacy systems in real time was genuine R&D: the uncertainty took four months to resolve. The remaining eight months went on standard reporting screens and user permissions, ordinary development work, not R&D. Only the first four months' costs belong in the claim.

Before assuming a project qualifies, check the following:

• The uncertainty is scientific or technological, not commercial uncertainty (e.g., whether the product will sell).
• You can identify the point work started on resolving that uncertainty, not the point the wider project was scoped or funded.
• You can identify the point the uncertainty was resolved (or abandoned), not the point the product shipped.
• Each project is assessed on its own terms. A single commercial build can contain both qualifying and non-qualifying phases side by side.

How the PAYE/NIC cap could limit what you receive

The PAYE/NIC cap limits the payable credit an SME can receive, based on the PAYE and NIC the business has actually paid on its employees. It's £20,000 plus 300% of your PAYE/NIC liability for the period. Anything your calculated credit produces above that isn't paid out in that period.

Company A has a PAYE/NIC liability of £10,000 for the period. Its cap is £20,000 + (3 × £10,000) = £50,000. If its R&D calculation produces a payable credit of £72,000 before the cap applies, only £50,000 is actually paid.

Under the merged scheme, the remaining £22,000 carries forward to the next claim. Under ERIS, it doesn't.

For start-ups and small teams leaning on subcontractors rather than employees, this cap can bite. Confirm your PAYE/NIC liability and run the calculation before you assume the full credit will land, and check whether subcontractor or externally provided worker PAYE/NIC can be brought into the calculation. It sometimes can, depending on the arrangement.

Are your staff costs apportioned defensibly?

The PAYE/NIC cap is applied to a number built from your staff costs, so getting the apportionment right matters twice over: it sets the size of the claim, and it's the part HMRC scrutinises hardest if it asks questions.

Your apportionment also defines whether you qualify for ERIS, as this is how you work out how much of your total expenditure is R&D compared to non-R&D.

HMRC accepts reasonable, evidence-based estimates. Very few people spend 100% of their time on qualifying work, so some apportionment is normal. What matters is that the method is consistent and defensible, not mathematically perfect.

A developer spending three days a week on a genuine technical problem and two days a week on routine maintenance justifies a 60% apportionment, backed by sprint records, timesheets or a simple weekly log.

Keep the evidence behind each percentage (timesheets, project records, documented management estimates). A figure with no supporting rationale is a weak point in any claim.

Do your records back up the numbers?

This is where the apportionment percentages above actually come from. The most common reason claims are weak isn't that the R&D wasn't real, it's that nobody recorded it at the time, and reconstructing a year's technical detail from memory afterwards is unreliable.

Build tracking into the year rather than treating it as a year-end task. The simplest fix for a small team: one person working on the project pulls together three or four bullet points from the R&D project every month or quarter, covering what was attempted, what was resolved, and what cost went where.

Are you inside the two-year claim deadline?

This area has no flexibility at all. An R&D claim must be made within two years of the end of the accounting period it relates to. You can submit through your original CT600 or through an amendment. Miss the deadline, miss your claim.

Note the deadline as soon as an accounting period ends, and treat "is there R&D to claim this year?" as a standard year-end question alongside your usual accounts and tax prep, not something to figure out later.

DIY, software, or an adviser?

For a small company with £50,000 to £200,000 of R&D spend, the claim route you choose affects both cost and accuracy. Preparing a claim entirely in-house is possible, but it asks a lot of a team that might not have R&D tax expertise sitting in-house already. A full consultancy service gives you that expertise, but the fee can feel disproportionate to the size of the claim. Guided software sits between the two: it structures the process, applies the rules consistently, and gives you a reviewed claim without the cost of a full advisory relationship.

Which of the three suits you depends on how complex your R&D activity is and how confident you are in your own record-keeping. For more information on Tax Cloud, check out our guide: How Does Tax Cloud Work?

Key takeaways

• Check your scheme before you assume the merged rate applies. If you're loss-making and R&D-intensive, ERIS usually pays out more.
• The Advance Notification Form comes first. For first-time or lapsed claimants, this six-month window closes long before the claim itself is ready.
• Innovative isn't automatically qualifying. Pin down the scientific or technological uncertainty and the exact dates work on it started and stopped.
• The PAYE/NIC cap can restrict what you receive. Calculate it early and check whether subcontractor or EPW contributions can be included.
• Apportionment needs to be reasonable and evidenced, not uniform across every project and person.
• The two-year claim deadline has no flexibility. Treat it as a fixed date in the calendar.

Getting a small company's first R&D claim right is mostly about sequencing: knowing which of these checks to do first, and not skipping ahead to the number before the groundwork underneath it is solid. If you'd like help working through any of these steps for your business, get in touch and we'll walk you through it.