Definition
The BIS Guidelines, published at gov.uk, require that R&D seeks a scientific or technological advance beyond what is already known. The baseline of knowledge is therefore the aggregate of what is publicly known: published research papers, open-source implementations, commercial software and hardware, academic textbooks, and documented industry practice. It is not limited to what the claimant's team knew. Nor is it limited to what competitors use in their products. It is the state of the public knowledge base in the specific sub-field.
HMRC's CIRD manual addresses this at CIRD81200, where it states that the advance must be beyond what is known or knowable in the public domain. Knowledge that is proprietary to another company and not publicly disclosed does not raise the baseline. If two companies independently grapple with the same problem and neither has published their solution, both may be doing R&D even though each is working on substantially the same challenge.
How the baseline is established in a claim
A well-prepared claim narrative establishes the baseline explicitly. It describes: what published approaches existed at the project start date; what commercial tools or libraries were available; and where those tools fell short of the project's requirements. This framing positions the subsequent investigation as genuine advance rather than implementation of known techniques.
The baseline is field-specific. A company developing a novel algorithm in a niche subdomain of image processing should establish the baseline for that subdomain, not for image processing generally. Using a too-broad baseline makes the uncertainty appear cosmetic; using a correctly focused baseline reveals where genuine unknowing begins.
Example
A company is building a model to predict equipment failure in a specific category of industrial machinery. At the project start, published methods for predictive maintenance exist, but they rely on sensor data types that are not available for this class of equipment, and the failure modes are not documented in open datasets. The baseline of knowledge in this sub-field does not contain a documented approach to the company's specific prediction problem. That gap between what is publicly known and what the company needs is where the R&D sits.
Common mistakes
The most common mistake is describing the baseline vaguely. Phrases like "nothing suitable was available in the market" are not the same as "the published literature in this sub-field as of [date] did not contain a documented approach to [specific technical problem] because [specific gap]." HMRC reviewers look for specificity.
A second mistake is setting the baseline at the start of the project without actually documenting what the team looked at before concluding that investigation was needed. Good practice is to retain records of the literature review, the commercial products evaluated, and the reasons they were ruled out. These records support the baseline claim if HMRC later queries it.
If you want to test whether your project narrative clearly establishes the baseline, the free assessment process is a useful sense-check.