Routine work that does not qualify

Every R&D scheme defines qualifying work in positive terms and then rules out the work around it, in a statutory list, in published guidance, or both. The form differs but the content converges: in every one of the seven countries, routine development with known methods, configuration and integration, quality control, cosmetic changes, market research and commercial production sit outside the claim. Reviewers know these categories well, and a claim that has not carved them out invites a reduction.

This guide follows the line drawn in Qualifying versus non-qualifying activities and goes into the excluded side in more detail.

What this guide covers

• The categories of work every scheme leaves out, and why.
• Each scheme's exclusions in its own wording, with links to the country pages.
• The two kinds of testing, and why one qualifies and the other does not.
• A method for carving routine work out of a project.
• What good carving out looks like, and what goes wrong.

The categories every scheme leaves out

Routine development with known methods

Applying established techniques, tools and methods to build something is development, but it is not R&D, even when the result is commercially new. The test is the same as for uncertainty itself: could a competent professional in the field have done this from existing knowledge, without experimenting? If yes, the work is routine. The software guidance published in six of the seven countries says this in nearly identical terms, and the United States reaches the same result through the uncertainty and experimentation parts of its test.

Configuration, installation and integration

Configuring purchased software or equipment, installing it, and integrating systems through documented interfaces following the supplier's instructions is routine. Integration becomes a candidate only where the combined behaviour could not be predicted and had to be discovered by structured testing, and then only the discovering qualifies.

Quality control and routine testing

Testing divides in two. Testing done to find out, where the result decides an open design question, is part of the work done to resolve the uncertainty. Testing done to confirm, against established requirements or a standard, that something works as intended is quality control. Regression, acceptance and compliance testing, calibration and inspection fall on the routine side. The purpose of each test run decides which side it is on.

Cosmetic and stylistic changes

Changes to appearance, style, colour, finish, taste or seasonal design are excluded in every scheme, however much work they take. A change that looks cosmetic but raises a technical question (a finish that must survive a new chemical environment) is judged by the technical question.

Market research and commercial work

Market research, market testing, sales promotion, user research into preferences, commercial feasibility, pricing, and the legal and financial steps needed to bring a product to market are outside every scheme. They may be essential to the project; they are not R&D.

Commercial production

Once the uncertainty is resolved, producing saleable output on the settled design or process is commercial production, not R&D, even for a new product. Some schemes allow the experimental portion of a production run where the production itself is the experiment; some exclude trial runs and tooling up altogether.

Internal administration software, where excluded

Several schemes single out software developed for the business's own administration (payroll, accounting, human resources) and either exclude it from eligible activities or apply a higher test. Elsewhere it is judged by the ordinary test of whether it resolved a technical uncertainty.

Other common exclusions

Research in the social sciences, arts and humanities; management studies and efficiency surveys; prospecting for minerals; routine data collection; and the legal and administrative work of patenting appear in several of the lists below.

Carving routine work out of a project

The method starts from the uncertainty, not from the project. A project is a commercial unit; the claim is built on the work done to resolve a technical unknown inside it.

1. State the uncertainty first. One or two sentences, in the technical lead's words, dated. Without it there is nothing to carve towards.
2. List the project's activities. Use the structure the team already has: epics, work packages, work orders, batch records.
3. Mark each activity. Does it resolve the uncertainty? Does it directly support resolving it (a test rig built only for the experiments)? Or neither? Only the first two are candidates, and the second is subject to each scheme's rules on supporting work.
4. Apply the time boundary. Within a candidate activity, work before the uncertainty was being addressed and after it was resolved or abandoned comes out.
5. Apportion shared costs honestly. Where a person or a rig served both sides, use the time records to split the cost. A round percentage applied to a whole team is the sign of a claim that has not been carved out.
6. Record the exclusions with reasons. What was left out, and why, is as valuable under review as what was claimed.

What good looks like: a project where 40 percent of the cost is claimed and the other 60 percent is listed by activity with a one-line reason for each exclusion. What goes wrong: claiming the whole project because part of it was uncertain; claiming the whole payroll of a team called "R&D"; treating every test run as experimentation; and claiming the launch because the product was new.

Illustrative example

Illustrative example. Halvard Instruments, a hypothetical company, spends 500,000 developing a handheld gas analyser. The technical uncertainty is whether sensor drift can be compensated across the required temperature range, where the published compensation methods failed in early tests and a competent instrumentation engineer could not say whether it could be done. The carve-out below is what the finance lead and the technical lead agree, activity by activity. The claim is built on 180,000, and the benefit is calculated at that country's rate. Whether the test rig counts depends on the scheme's rules on supporting work.

Country differences

The exclusions each scheme states, in its own words; the country pages give the detail and the exceptions.

Related pages

• Qualifying versus non-qualifying activities: where the line really sits
• R&D tax incentives for manufacturers: where the qualifying work hides
• How to run a project identification workshop with your technical team
• R&D Tax Fundamentals, Module 2: What counts as R&D
• Eligibility hub

Educational material, not advice. Eligibility and treatment depend on the scheme and the facts.