Evaluation criteria

Proposals are scored on three criteria. EXCELLENCE asks whether the concept is sound, ambitious and beyond the state of the art, and whether the methodology will actually deliver it. IMPACT asks who will be better off, by how much, and through what route. IMPLEMENTATION asks whether this consortium, with this work plan and this budget, can do it.

Each is scored from 0 to 5 with one decimal place, so 3.5 and 4.0 are meaningfully different marks. Half a point across three criteria decides funding in a crowded topic.

Impact is where most proposals lose. Researchers write it as a description of what they will publish and who might be interested, when the criterion asks for a causal chain: this result, reaching these users, through this route, producing this measurable change, with these barriers and these ways past them. A credible modest impact scores higher than an implausible transformative one.

Implementation is where most proposals lose points cheaply and needlessly. Work packages with unclear ownership, a Gantt chart that ignores dependencies, a risk table listing only risks that will not happen, a partner with 4 person-months and no visible role. Each is an easy deduction for an evaluator who must justify a score.

The criteria carry equal weight in a research or innovation action, but ties are broken differently: Excellence decides for a RIA and Impact decides for an IA. Other instruments reweight them, and the ERC discards two of the three. Its assessment is excellence alone.