How to Strengthen Proposal Evidence Before Submission
A proposal can contain excellent science, a credible product and experienced partners, then still lose points because its central claims are merely asserted. Knowing how to strengthen proposal evidence is therefore not a matter of adding more citations or making the document longer. It is the work of making every claim that matters verifiable against the call, the proposed method and the capacity of the consortium.
An evaluator has limited time, a defined marking form and no obligation to infer what the applicant meant. If the evidence is not where the criterion asks for it, a strong idea can read as an unsupported ambition. Once submitted, that reading is fixed.
Evidence is what allows an evaluator to award the score
Evidence is not a single category. In an EU proposal, it may be a baseline dataset, a pilot result, a technical benchmark, a cited regulatory requirement, a named partner responsibility, a signed access arrangement, a budget allocation or a measurable KPI. Its value depends on the claim it is meant to support.
The test is straightforward: could an evaluator point to a passage and say, “This is why I can award the higher score band”? If not, the text may be persuasive but it is not yet evidential.
This distinction matters because evaluation forms score defined award criteria, not the general quality of the consortium’s intentions. Horizon Europe proposals commonly face Excellence, Impact, and Quality and Efficiency of Implementation, with thresholds and weightings set out in the applicable call documentation. Digital Europe and Erasmus+ use their own forms, sub-criteria and scoring arrangements. Do not import a generic checklist from another programme. Start with the evaluation form published for the specific call and action type.
Admissibility and eligibility come first, but they are not award criteria. Meeting page limits, submitting on time and satisfying participation conditions prevents exclusion. It does not establish that the objectives are credible, the exploitation route is plausible or the work plan is deliverable.
Find the claims that carry the score
Proposal teams often review sections. Evaluators score claims within criteria. That is a materially different reading method.
Take an Impact section that states the project will improve adoption of a technology across European SMEs. The sentence contains at least four claims: the technology is adoptable, the target SMEs are reachable, the consortium can influence adoption, and the anticipated improvement can be measured. A dissemination plan alone does not prove any of them.
Work through the draft by identifying score-bearing claims, then attach the evidence needed for each. For a complex proposal, a claim-evidence register is usually more useful than a late-stage prose edit. It need not appear in the submitted document. Its purpose is to expose gaps before they become an evaluator’s significant weakness.
For each claim, record the relevant criterion and sub-criterion, the current supporting passage, the evidence type, the owner and the location where a reviewer can find it. Where the answer is “the consortium knows this”, treat that as a gap. Evaluators can only score what the proposal demonstrates.
Excellence: show that the method can answer the question
Under Excellence, vague methodological confidence is a recurrent weakness. “We will develop an innovative platform” does not establish the state of the art, the technical advance, the research hypothesis, the validation method or the limits of the approach.
Stronger evidence identifies the baseline against which improvement will be measured and explains why that baseline is relevant. It distinguishes an objective from a method. It specifies the data, trial environment, sample, test protocol or technical architecture that will produce the expected result. If the project relies on access to a facility, cohort, data space or demonstration site, say who controls that access and on what basis.
Novelty claims require particular discipline. A literature review or market scan may support them, but only if the proposal states the comparison clearly. “First-of-its-kind” is rarely useful without a defined comparator, a relevant geography or market, and an explanation of what is genuinely new.
Impact: evidence the route between output and outcome
Impact sections frequently fail in the middle. They move from project outputs to broad European benefits without showing the mechanism between them.
An evaluator needs to see a route from deliverable to use, from use to outcome, and from outcome to the expected impact specified in the call. That route may include validation, standardisation, regulatory engagement, procurement, licensing, training, integration by a named user or a route to market. Which elements matter depends on the topic, but the chain must be explicit.
Use credible baselines and defined indicators. A target such as “reach 500 organisations” is weak if the proposal does not identify the channel, audience, conversion assumption, responsible partner and means of measurement. A more defensible formulation explains that a named partner has access to a network of a stated size, will run a specified engagement activity, and will measure uptake through agreed indicators at defined points.
Do not confuse a communication activity with impact. A webinar, website or policy brief is an output. Its contribution to adoption, behaviour, investment or policy change needs evidence of audience relevance and a realistic pathway.
Implementation: evidence that the work can be controlled
Implementation scores fall when a work plan looks complete but cannot be managed. A Gantt chart with work packages and milestones is not sufficient evidence of feasibility.
The evaluator looks for dependencies, decision points, resources, risk ownership and the relationship between tasks, deliverables and objectives. Where a critical result depends on one partner recruiting users, obtaining approvals or supplying data, identify the dependency and the contingency. A risk register that says “delay” and “mitigation: monitor closely” does not permit confidence in the response.
Budgets and person-months should also tell the same story as the technical narrative. If a work package claims to conduct large-scale pilots across several countries but assigns limited staff time, evaluators may reasonably question delivery. The issue is not always the amount requested. It is whether the resource logic is visible.
Consortium capacity should be evidenced at the point of need, not confined to generic partner profiles. Show which partner has performed the relevant task, with what infrastructure, dataset, network, IP position or operational authority. Past projects help only where their relevance to the proposed responsibility is made explicit.
Test evidence from six evaluator seats
A final review should not ask whether the team likes the draft. It should ask whether separate evaluator readings reach the same conclusion for the right reasons.
Read the proposal at least through the lenses of Excellence, Impact, Implementation and consortium capacity. Add a red-team reading whose sole purpose is to locate unsupported claims, missing comparators, unstated assumptions and numbers without sources. Where appropriate, ask a sector specialist to challenge whether the stated baseline, route to adoption or regulatory position is credible.
Disagreement is useful evidence in itself. If one reviewer reads a passage as a convincing validation plan and another reads it as an aspiration, the proposal has not made the case clearly enough. Do not average that disagreement away. Return to the criterion, identify what each reading can and cannot quote, and revise the text so that the intended interpretation is the only reasonable one.
This is the logic behind BidShark’s independent readings and adjudication: the automated assessment is tied to the relevant published marking form, while disagreement and quoted passages remain visible for the applicant to inspect. It is a pre-submission control, not a prediction of funding and not a substitute for the call documentation.
Revise the evidence, not just the wording
When time is short, teams often polish language because it is faster than resolving an evidential gap. That can improve readability, but it cannot turn an unvalidated assumption into proof.
Prioritise revisions that change what an evaluator can substantiate: insert the baseline, name the responsible organisation, quantify the resource, state the decision rule, show the dependency, or narrow an over-ambitious claim to what the project can actually demonstrate. Then make the evidence easy to locate with direct signposting between objectives, methods, work packages, KPIs and impact pathways.
The final question before submission is not whether the proposal sounds convincing to the people who wrote it. It is whether a sceptical evaluator, reading against the scoring form, can find enough evidence to justify every point you are asking them to award.