Rebuilding the Assessment System: An Account from Inside Innovate UK

Each year, thousands of businesses apply to Innovate UK (IUK) for innovation funding, and every application is judged through a broadly similar assessment process of questions, criteria, scores and guidance that has been in place, with only minor modification, since the early 2000s. This report provides an insider account of how IUK redesigned that process between 2017 and 2026, viewed from street level: through the experiences of the administrators who develop and oversee assessment, and of the research team that advised the agency across the period. An audit of 34,869 assessor evaluations found little indication of favouritism, by location, gender or other applicant characteristics, in how applications were scored. It did show that IUK’s ten assessment criteria were functioning as a single evaluative judgement rather than ten, and that the equal weighting of criteria was a design choice with real consequences for which projects were funded. Benchmarking of fifteen international agencies showed that most operate with four or five criteria. On this evidence, and through workshops with its own staff, the agency developed a five-area framework covering why, what, how, commercial intent, and finance and value for money, which was then tested in practice: assessors re-evaluated applications under the new questions, and paired surveys of assessors and applicants were embedded in the Women in Innovation Awards. Both groups broadly welcomed the new approach.

The report treats the redesign as a case of how an innovation agency learns about, and changes, its central allocation device. It documents the choices involved in building an assessment system, the difficulty of altering a shared system within a federated organisation, and the refinements the live testing surfaced. It closes with the challenge the redesign leaves open: with four in five applicants now using generative AI to prepare their bids, the cost of producing a plausible application has fallen sharply while the cost of assessing one has not, unsettling both the economics of the system and what assessment scores measure. For those who design or run grant assessment, the report offers evidence on how criteria were diagnosed, redesigned and tested over a decade; for researchers advising public institutions, it draws lessons on sustained engagement, planning for non-linearity, and claiming contribution rather than outcomes.

Download the full brief using the button on the right-hand side of this webpage.