This page is about UK (HMRC) claims. You have US (IRS) selected. See US (IRS) guides or .

UK R&D Tax · Practice Efficiency

The Hidden Cost of Collecting R&D Evidence Retrospectively

Repeated client requests, technical interviews, searching project systems and reconstructing historical work can create substantial administrative costs for R&D advisers. Explore how practices can measure evidence-collection time, identify inefficient processes and evaluate continuous evidence preservation.

For:
Partners and managers at R&D tax practices
Published:
11 Oct 2026
Last reviewed:
11 Oct 2026, against the sources listed below
Prepared by:
Evidante. It has not been independently reviewed by a qualified tax adviser, and it is not tax or legal advice.

In brief

  • Collecting evidence after the period means requests, chasing, interviews and searching through systems, mostly by people whose time is not separately recorded.
  • Practices can measure this with a few weeks of time recording on a sample of claims. The worked example below shows the arithmetic; the numbers in it are assumptions.
  • HMRC does not require any particular system. Its guidance says it does not expect expensive systems, and that for smaller projects the records a business normally creates are often enough to identify a qualifying project and its costs later, supplemented by the recollections of the people involved.

What retrospective collection involves

When a claim is prepared after the accounting period, the adviser usually has to rebuild the story of each project from what can still be found:

  • sending requests for project information, then following them up;
  • interviewing technical staff about what was attempted, what failed and why;
  • searching ticketing systems, code repositories, shared drives and chat history for the period;
  • working out who was involved and for how much of their time;
  • matching activities to costs, and resolving the gaps and contradictions that appear.

Why the cost is easy to miss

Much of this time is not recorded as a separate task. It sits inside fixed fees, inside the time of the client’s engineers, or inside “claim preparation” as a whole. It also tends to rise when the people who did the work have moved on: HMRC’s GfC3 part 5 says missing records “can be a particular problem if staff who worked on the project are no longer available”.

How to measure it in your practice

1. Pick a sample

Choose a handful of recent claims of different sizes and clients.

2. Record time by task

For each, record time separately for evidence collection (requests, chasing, interviews, searching), technical writing, cost work and review. Ask clients, where they are willing, how much engineering time they spent.

3. Note what was reconstructed

Record which statements in the claim rest on reconstruction rather than on records made at the time, and what each was reconstructed from.

4. Work out the cost

Multiply hours by a fully loaded cost per hour, separately for your staff and the client’s, and scale by claims per year.

A worked example

The figures below are invented to show the method. Replace every one with your own.

Hypothetical inputs, for illustration only. They are not industry averages or results from Evidante customers.
Retrospective evidence task (one claim)WhoHours
Requesting, chasing and following up records from the clientAdviser4
Technical interviews to reconstruct what was tried and whyAdviser3
Engineers searching tickets, repositories and chat for the periodClient’s engineers5
Engineers attending interviewsClient’s engineers3
Matching records to projects and costs, and resolving gapsAdviser4
Adviser hours × £90 + client hours × £6011 × £90 + 8 × £60 = £1,470
Across 30 claims a year£1,470 × 30 = £44,100

In this example, the evidence-collection part of each claim costs £1,470 of staff time, £990 of it the adviser’s, and about £44,100 a year across 30 claims. If records collected during the period removed, say, half of the requesting, searching and reconstruction, the figure would fall accordingly; how much it actually falls is something to measure, not assume. Time saved is only a saving if it is used for other work. You can try your own numbers in the illustrative calculator on our home page.

What retrospective evidence cannot do

Official guidance

HMRC accepts that a project may be identified after it began or finished, and that “in the absence of written records, a detailed explanation provided later may be acceptable in some cases” (GfC3 part 5). It also says that without records “you may find it harder to identify there was a project and harder to show your expenditure qualifies for relief”.

Reconstruction can describe a project; it cannot create records that were never made. Where a claim rests on recollection, say so, and say what it was checked against.

Ways to reduce the effort

  • Agree with clients at the start of the period what they will keep: a short project plan, the uncertainties as they are identified, and test results. GfC3 says a page of bullet points can be enough for a small project.
  • Ask clients to tag tickets or repositories by project, so records can be grouped later without interviews.
  • Hold short check-ins during the period rather than one long reconstruction at the end.
  • Collect engineering records automatically from the tools the team already uses, so they exist whether or not anyone remembers to write them down.

Common practice

None of these is required by HMRC, and each has its own cost. The right mix depends on the client: a company running one small project may need very little, while one running many technical projects benefits more from collecting records as it goes.

What this means for your practice

  • Measure evidence-collection time before deciding to change anything.
  • Start with the clients where reconstruction is hardest: many projects, high staff turnover, or history of HMRC questions.
  • Whatever you change, keep the professional’s judgement and the cost methodology at the centre of the claim.

Evidence checklist

A practical list for organising records. It is a recordkeeping aid, not a test of whether work qualifies.

  • Time recorded against evidence collection separately from technical writing and cost work, for a sample of claims.
  • The number of requests sent to each client per claim, and how long each took to be answered.
  • Which records had to be reconstructed, and from what.
  • Gaps that could not be closed, and why (for example, people who had left).
  • The client’s own engineering time spent on the claim, where the client will share it.
  • A view of which clients would benefit most from collecting records during the period instead.

Where Evidante fits

Evidante is one way to collect records during the period: the client connects its engineering tools once and records of commits, pull requests, builds, ticket changes and team notes are kept as they happen, by client and project. It does not remove the need for the professional’s explanation, technical interviews where they add something, or cost work. Whether it saves time for a given practice depends on that practice’s clients and process, which is why we suggest measuring first.

Evidante does not decide whether work qualifies, does not calculate qualifying costs, does not replace the professional's judgement and cannot guarantee how HMRC will treat a claim.

Official sources

Grouped by what kind of source each is, because they do not carry the same weight. Checked on 11 Oct 2026.

HMRC guidance

HMRC's published view and process. Not law.

HMRC internal manual

Guidance to HMRC officers. Shows HMRC's approach; not law.

Official statistics

Published figures, for context.

Related UK guides

Educational information only. Not tax or legal advice. Rules change and depend on the company and accounting period; check the current HMRC sources and take professional advice before acting. Evidante is not endorsed by HMRC.