Skip to content
R&D TaxPartners
Book a call

Guide

R&D tax relief examples: what qualifies and what does not

Examples of R&D projects that do and do not qualify for tax relief across software, manufacturing, construction, food and engineering, tested against the rules.

A project qualifies for R&D tax relief if it sets out to achieve an advance in science or technology, and the work involved scientific or technological uncertainty that a competent professional could not readily resolve. Routine development, cosmetic changes and commercial risk do not qualify, however new they are to your company. The best way to see the line is through examples, so this guide sets out pairs of projects, one that qualifies and one that does not, in five sectors.

The test in brief

The Government's guidelines on the meaning of research and development for tax purposes (opens in a new tab), last updated on 7 March 2023, set three ideas at the centre:

  • An advance. Progress in overall knowledge or capability in a field of science or technology, not just something new to your company. An appreciable improvement to an existing process, material, device or service can count.
  • Uncertainty. Whether something can be done, or how, is not readily available or deducible by a competent professional working in the field. This includes system uncertainty, which arises from how known components behave together.
  • Direct contribution. The activities must directly contribute to resolving that uncertainty. Routine analysis, copying or adaptation of an existing product does not count.

Our page on what qualifies goes through each in more detail.

The examples below are illustrations written by us. They are not client cases, and your facts will differ. Where an example follows one in the guidelines, we say so.

Software

Qualifies. A company needs a mobile app to keep data consistent across devices that spend long periods offline, with strict limits on battery and data use. Known synchronisation methods fail under those limits, and the team does not know if any approach can meet them. They design and test several new approaches, most of which fail. The advance is a technique that works within the limits, and the uncertainty is whether one exists. The failed attempts are part of the evidence.

Does not qualify. A company builds a customer portal using a standard web framework and off-the-shelf components. The features are new to the business, but a competent software professional could specify the approach in advance. This is ordinary development.

The guidelines apply the same criteria to software as to any other field (example B1), and treat software to analyse market research data, with no advance in software expected, as not R&D (example D2). Our software sector page has more.

Manufacturing

Qualifies. A manufacturer wants to make a component from a recycled polymer blend. Existing moulding conditions produce warping and weak points, and nobody can say if the required tolerance is achievable with this material. The company runs trials to change the process and the blend. The uncertainty is technological, and the work directly addresses it.

Does not qualify. The same manufacturer, having proved the process, scales it up to a second production line, switches supplier, and tunes machine settings within known limits. Scaling a proven process is production, not R&D. The guidelines make the same point for testing: further testing after a design is settled is not R&D (paragraph 39 and example G1).

See our manufacturing sector page.

Construction

Qualifies. A builder develops a prefabricated wall panel that must meet thermal and structural targets together. Published data does not predict how the new combination of materials will behave under load and moisture, so the team builds and tests prototype panels and revises the design. This is an appreciable improvement where the outcome was uncertain.

Does not qualify. A contractor applies a standard design to a new site, adjusts it to meet building regulations using established methods, and manages the programme. Project management, site conditions and compliance with known standards are not scientific or technological uncertainty. Our construction sector page covers the line in more detail.

Food and drink

Qualifies. A producer wants to remove an ingredient from a chilled product without shortening its safe shelf life. Whether the product can stay microbiologically stable with a different formulation is not known in advance. The company screens alternatives, tests stability and revises the formulation. The guidelines treat systematic ingredient and formulation testing to resolve this kind of question as R&D (example E2).

Does not qualify. The same producer launches a new flavour using established methods, changes the packaging design, and runs taste panels to see which version consumers prefer. The guidelines treat assessing what will appeal to consumers as not R&D (example E3), and cosmetic or aesthetic qualities alone are outside R&D (paragraph 42). If achieving a particular appearance needs a technological advance and involves real uncertainty, that part can qualify (example H2 deals with a fabric). Our food and drink sector page has more.

Engineering

Qualifies. An engineering firm designs a compact heat exchanger for a harsh environment. Each component is understood, but how they behave together under vibration and temperature cycling cannot be predicted from the literature. This is system uncertainty. The guidelines say that combining components can involve uncertainty even when the principles are known (paragraphs 29 and 30), and a similar example (J2) treats designing an engine that integrates new components as R&D until a functionally final prototype is tested.

Does not qualify. The firm produces a detailed design for a client from handbook calculations and a standard specification. The work takes skill and time, but the answer was readily deducible by a competent engineer. Assembling components to an established pattern involves little or no uncertainty. See the engineering sector page.

What this means for your claim

Look at your own projects with three questions:

  1. What did we not know how to do at the start?
  2. Why could a competent professional in our field not simply have looked it up or worked it out?
  3. What did we try, and what happened?

If you have clear answers, with records made at the time, you probably have a claim. If the honest answer to the first question is "how to sell it", "how it should look" or "how much it would cost", that is commercial uncertainty and does not count. When a project has both kinds, only the technological work qualifies.

Talk to us

Describe a project in a few lines and we will give you an honest first view. Book a call. We reply the same working day.

This article is general information, not advice. The examples are illustrations, and whether a real project qualifies depends on its facts.

FAQs

Frequently asked questions

What is an example of a project that qualifies for R&D tax relief?

A project that sets out to do something a competent professional could not say was possible, or how to do, using the knowledge readily available in the field. For example, developing a way of moulding a recycled material to a tolerance that existing methods cannot hold, where nobody knows in advance if it can be done. A failed attempt can still qualify.

Does a project have to succeed to qualify?

No. The Government's guidelines say that a project which fails, or achieves only a marginal gain, can still be R&D, because what counts is that it sought an advance and involved uncertainty. This is why the technical records of what was tried, including the failures, are important evidence.

Is building new software always R&D?

No. The guidelines apply the same criteria to software as to any other field. Building features with known tools and techniques is ordinary development. Software work qualifies when it seeks to achieve something that a competent professional could not readily say was possible, or how, such as a new technique or an untested combination of components.

Do design, styling or customer appeal count?

Not by themselves. The guidelines treat cosmetic or aesthetic qualities, and testing what appeals to consumers, as outside R&D. The exception is where achieving a cosmetic effect needs a scientific or technological advance and involves uncertainty, in which case that work can qualify.

Does adapting an existing product or process qualify?

Only if the result is an appreciable improvement and getting there involved uncertainty. The guidelines exclude routine analysis, copying or adaptation of an existing product, and treat minor or routine upgrades as outside R&D. The question is what you had to work out that a competent professional could not simply have looked up.

Does assembling known components qualify?

Sometimes. The guidelines recognise system uncertainty, where the difficulty lies in how components behave together, even when each is well understood. But assembling components to an established pattern involves little or no uncertainty. The claim has to show what could not be predicted about the system.

When does the R&D stop?

When the uncertainty has been resolved. The guidelines give the example of testing a prototype: testing that helps resolve uncertainty is R&D, while further testing after the design is settled is not. Scaling up, routine production and market testing come after that point.

Can you tell me if my project qualifies?

We can give a first view. Describe the project in a few lines: what you were trying to achieve, what you did not know how to do, and what you tried. We will tell you plainly whether we think it meets the test. If it does not, we will say so rather than recommend it.

Related advice

Related services

Talk to an R&D tax specialist.

A confidential first conversation about your claim or your HMRC enquiry. We reply the same working day.

Book a callEmail us

Or write to taxadvisory@aswatax.co.uk

Chartered Tax Adviser
Message us on WhatsApp (opens in a new tab)