Every finance team can tell you what was spent on payroll, rent, and travel last quarter. Far fewer can tell you what was spent on R&D. The difficulty comes from a structural mismatch: The financial reporting systems are designed around transactions, departments, and accounting periods, while R&D activity follows a different logic.
Two different maps of the same territory
An accounting system organises the business around payroll runs, invoices, departments, and reporting periods. R&D activity is organised around experiments, prototypes, technical decisions, failed approaches, and work packages. Both ultimately describe the same underlying activity, just from different perspectives.
Apply a technical lens and the relevant questions concern the scientific or technological uncertainty, what was investigated, and what the team learned. Apply a financial lens and the questions change. Which resources were consumed? How should those resources be attributed? Which costs qualify under the relevant funding or tax regime? The different relevant questions of each lens produce different views of the same work.
Many businesses discover this gap only when a grant auditor, tax reviewer, or board member asks for a defensible account of R&D expenditure or the return generated by innovation spending.
Why "just track it better" doesn't fix it
The first response is often to treat this as a tracking problem. Better tagging and a cleaner chart of accounts can all help, but the deeper issue remains. A meaningful share of R&D cost categories can't be classified with certainty in the first place.
Some costs are unambiguous. Equipment or a researcher entirely dedicated to R&D, for example. But most real R&D spend sits somewhere between clear categories. A cloud subscription may support both production and development, a senior engineer may divide their time across several projects, capital equipment may begin in an R&D environment before moving into commercial production. In these cases, apportionment requires active judgement.
This uncertainty is a structural feature of R&D financial measurement. Standard accounting platforms, project-costing tools, and job-management systems don’t always reflect the rules used to assess R&D qualification.
The challenge is especially strong in product-led and software businesses, where technical work may cut across products and teams rather than clearly defined projects.
The scale of the disconnect
Even at the country level, R&D expenditure figures depend on interpretation. The OECD developed the Frascati Manual (the internationally recognised standard for defining and measuring R&D) in an attempt to address the difficulty of consistent measurement across countries and sectors. Even with specialist statistical agencies behind the effort, estimates still involve assumptions and professional judgement.
The same challenge appears within individual businesses, where R&D qualification often doesn't sit clearly within anyone's job description. Finance owns the numbers but may have limited visibility into the technical work. Engineering owns the work but may have little training in the financial rules that determine which activities and costs qualify.
What this means in practice
None of these complications mean R&D expenditure can't be measured. It can be, it just requires a different approach from the one used for much of the profit and loss account. A reliable process begins by recognising the mismatch between technical activity and financial reporting.
The categories that create the most friction (staff time, capital equipment, subcontracted work, materials) each carry distinct rules and areas of judgement under R&D tax credit and grant regimes around the world. Recognising these challenges as structural is the first step towards producing an R&D expenditure figure that is consistent, transparent, and defensible.