How-to guides
What an evaluator reads first, and what they skip
No public document says in which order an evaluator reads a proposal. What is public is the set of instructions evaluators receive: the criteria, the scoring scale, the evaluation form they fill in, and a list of things they are told not to take into account. This post summarises those instructions for Research and Innovation Actions (RIA) and Innovation Actions (IA) under the Work Programme 2026-2027.
The process
The Commission's standard briefing for experts (version 14.0, 15 April 2026) describes five steps.
- EU staff check admissibility and eligibility.
- Individual evaluation: at least three experts assess each proposal separately and each writes an Individual Evaluation Report.
- Consensus group: the same experts agree on common comments and scores. They are told to agree comments before scores.
- Panel review: checks consistency across proposals, resolves cases where no agreement was reached, and orders proposals with the same score.
- The Commission or agency draws up the final ranking list.
The consensus report is the basis of the Evaluation Summary Report sent to applicants, and the briefing says the two are identical in most cases. General Annex F gives an indicative timing of around 5 months from the deadline to the result.
Criteria, scores and thresholds
General Annex D sets three award criteria: Excellence, Impact, and Quality and efficiency of the implementation. Scores are given per criterion, not per aspect within it.
| Rule | Full proposal | First stage of a two-stage call |
|---|---|---|
| Criteria scored | All three | Excellence and Impact only |
| Scale | 0-5 per criterion, in steps of 0.5 | 0-5 per criterion, in steps of 0.5 |
| Threshold per criterion | 3 | 4 |
| Overall threshold | 10 out of 15 | Set after evaluation, normally expected at 8 or 8.5 |
| Page limit for RIA and IA | 40 pages, 45 with lump sum funding | 10 pages |
For IA, the Impact score is weighted by 1.5, but only to establish the ranking, not to decide whether thresholds are passed. A score of 3 means "good" with a number of shortcomings, 4 "very good" with a small number, and 5 "excellent" with only minor ones. Specific topics can set different thresholds or weights.
What evaluators are told to disregard
- Excess pages. After the deadline, pages over the limit are made invisible automatically and are not considered.
- Hyperlinks that extend the proposal. The Part B template states that experts are instructed to ignore links to information designed to expand the proposal and get around the page limit.
- Potential. Proposals are evaluated as submitted, not on what they could become after changes. Experts explain shortcomings and reflect them in a lower score, and do not recommend changes to partners, work packages or budget.
- Work outside the topic. Aspects are assessed only to the extent that the work is within the scope of the topic. A partly out-of-scope proposal gets a lower score.
- Gender equality plans. These are checked by staff and are not part of the criteria.
- Ethics. Experts in the scientific evaluation do not need to comment on ethics, because successful proposals go through a separate ethics review.
- Use of generative AI by applicants. This may not be treated as a reason to penalise a proposal.
- The same weakness twice. One weakness should not be marked down under two criteria.
What the evaluation form shows
The evaluation form for RIA and IA (V4.0, 12 November 2025) starts with the project data, the list of participants and the abstract taken from Part A. Then come three comment boxes, one per criterion, each headed by the aspects from General Annex D. After the scores come further questions, including whether the proposal is in scope. The briefing requires comments on every aspect of each criterion, and only on those. From the Work Programme 2026-2027 onwards, it notes, the scale and significance of contributions are no longer considered under Impact.
Two more public rules
For a resubmission made within 2 years and declared in the forms, experts must justify comments and scores that differ significantly from the previous report, if conditions were comparable. An evaluation review can be requested within 30 days, on procedural grounds only, not on the merits.