EU Proposal Scoring Scale and What It Costs
A proposal can be technically credible, backed by experienced partners and still lose the call because it is read as a 3 rather than a 4. The EU proposal scoring scale is where that difference becomes visible. It is not a generic quality rating. It is the mechanism through which evaluators convert evidence in a fixed document into a ranked funding decision.
That distinction matters late in a bid. Once submitted, a weak impact pathway cannot be clarified in a meeting, an unconvincing role for a partner cannot be explained by email, and a missing delivery dependency cannot be repaired through goodwill. The evaluation summary report will record the consequence, but by then the score is final.
There is no single EU proposal scoring scale
Applicants often speak of the EU 0-5 scale as though it applies identically across programmes. It does not. Horizon Europe evaluation forms commonly use a 0-5 score for each award criterion, with half-point scores available. Digital Europe calls may use comparable arrangements but can set different thresholds, weightings and sub-criteria. Erasmus+ commonly works on a 100-point scale, with programme-specific minimum scores and category thresholds.
The published call documents, the applicable work programme and the evaluation form control. They should be treated as the scoring specification, not as background reading. A strong proposal against an assumed framework can be poorly aligned against the actual one.
In many Commission evaluation forms, the 0-5 descriptors follow a familiar pattern: 0 means the proposal fails to address the criterion or cannot be assessed because of missing or incomplete information; 1 is poor; 2 is fair; 3 is good; 4 is very good; and 5 is excellent. The wording attached to those bands matters. A 5 is not merely a section with no obvious errors. It normally requires that the proposal addresses the criterion successfully and has no significant shortcomings.
That is a demanding test. Evaluators are not obliged to infer what the consortium intended, fill gaps from their sector knowledge, or reward claims that appear only in an annex, a partner CV or a later work package. They score what is evidenced where the criterion asks for it.
How scores become a funding decision
Before award criteria are scored, proposals must pass admissibility and eligibility checks. Page limits, submission completeness, topic fit, legal entity conditions and consortium requirements can all sit outside the quality score. A well-written narrative does not cure an ineligible application.
For eligible proposals, evaluators score the stated criteria independently. In a typical Horizon Europe Research and Innovation Action or Innovation Action, those are Excellence, Impact, and Quality and Efficiency of the Implementation. Calls often set a threshold for each criterion and a total threshold, frequently expressed out of 15. Some actions also apply weighting, particularly to Impact, or use a distinct approach set out in the call conditions.
A threshold is a floor, not a competitive score. A proposal that receives 3.0 for every criterion may meet a 3-per-criterion and 10/15 total threshold, yet remain well below the scores needed in a heavily subscribed topic. Ranking is shaped by the final consensus scores, then by the call’s tie-break rules where needed. The practical question is therefore not simply, “Will this pass?” It is, “Which score band will an evaluator be able to defend?”
The cost of one point
One point can be lost in a single sentence or across an entire section. Consider Impact. A proposal may identify a credible market, cite policy relevance and promise measurable benefits. If it does not show a causal route from project outputs to adoption, deployment and quantified outcomes, an evaluator can reasonably score the pathway as incomplete.
The same pattern appears in Implementation. A Gantt chart may show a convincing sequence of work packages, but a 4 becomes difficult to justify if decision rights, interdependencies, contingency arrangements and resource allocations are only implied. An evaluator cannot award points for governance that the proposal appears to possess but does not describe.
The score band is also affected by scale. A minor ambiguity in a supporting task may be a shortcoming. A vague exploitation route for the project’s principal result, an uncosted pilot dependency, or a work package owned by a partner without relevant capacity may be significant. The distinction is not cosmetic. Significant weaknesses pull a score below the range in which the criterion is judged very good or excellent.
What evaluators scrutinise in each criterion
Excellence is not a reward for ambitious language. It tests the clarity and pertinence of objectives, the credibility of the methodology, and the extent to which the proposed work goes beyond the state of the art or current practice. For calls with social sciences, humanities, gender dimension, open science or other specific requirements, the evaluation form may place those matters inside the methodology rather than treat them as optional additions.
A common scoring loss occurs where objectives are presented as broad intentions rather than testable end states. “Develop an interoperable platform” says little about scope, performance, validation setting or acceptance criteria. The evaluator then has no stable basis for judging whether the methodology can deliver it.
Impact is where internal optimism is most expensive. Evaluators look for the logic between results, outcomes and the impacts requested by the topic. They test whether target users are specific, whether exploitation or uptake responsibilities are assigned, whether barriers are acknowledged, and whether KPIs measure outcomes rather than activity. A dissemination plan full of events and publications may be useful, but it does not by itself establish adoption.
Implementation concerns whether the plan can be executed. Evaluators inspect the coherence of work packages, tasks, milestones, deliverables, effort, resources, risk management, management structures and decision-making. Consortium capacity is often assessed here or through a related sub-criterion. The names on the participant list are not a score. Each partner’s role, relevant capability and contribution to critical work must be visible.
Where a call uses different criteria, such as an Erasmus+ relevance, design, partnership and impact structure, the principle remains the same: score against the published sub-criteria, not against a remembered Horizon Europe template. Reusing a successful prior proposal can be efficient. Reusing its evaluation logic without rebuilding it for the new action type is not.
Independent readings and adjudication matter
Real evaluation is not one reader’s impression. Evaluators work from briefs and forms, score independently and then reconcile their views in consensus. Differences are normal, particularly where a claim can be read as either evidenced or aspirational.
That disagreement is useful diagnostic evidence before submission. If one reader treats a governance description as adequate and another cannot locate authority for resolving a technical conflict, averaging the views conceals the issue. The correct response is to identify the passages each reading rests on, decide whether the evidence is genuinely present, and revise if it is not clear enough to survive both interpretations.
This is the logic behind a controlled pre-submission assessment. BidShark applies the official marking scheme transcribed for the relevant call type and form version. Six independent readings cover Excellence, Impact, Implementation, consortium capacity, unsupported claims and sector context. Where readings differ by more than a point on a criterion, the evidence is exchanged and adjudicated into one final 0-5 score. The report retains the disagreement rather than presenting a false appearance of certainty.
The scoring is automated, not a human evaluation panel verdict. Every finding is tied to the passage it relies on, so the proposal team can test the assessment against its own document. Where a team needs judgement from a person who has evaluated EU proposals, the separate Expert Q&A package provides written answers after that evaluator has read the report. Those are distinct activities and should not be confused.
Use the scale as a revision order
The useful output from a scoring exercise is not a decorative total out of 15 or 100. It is a defensible revision order. Start with any criterion at risk of missing its threshold. Then address significant weaknesses in high-weight or high-competition criteria, followed by shortcomings that recur across readers. Finally, remove claims that no section substantiates.
Do not attempt to improve every paragraph equally in the final days. A clearer abstract will not compensate for an impact section that lacks a credible uptake route, and a better diagram will not resolve a resource plan that does not match the work. Make each revision traceable to a criterion, sub-criterion and score-band concern.
The aim is not to predict the real panel’s exact score. No pre-submission process can do that, because evaluator expertise, competition and consensus remain outside your control. The aim is narrower and more valuable: find the weaknesses before the real evaluators do, while the document can still be changed.