Why Scores Fall on the Horizon Europe Impact Criterion

A Horizon Europe proposal can describe excellent research, name credible partners and present a competent work plan, then lose its competitive position on Impact. This is not because the evaluators failed to appreciate the science. It is because the horizon europe impact criterion asks a different question: whether the proposed work can plausibly create the changes the call seeks, at a meaningful scale, through a route the consortium can actually deliver.

That distinction matters when the draft is close to submission. You only get one chance to submit. A late Impact section full of enlarged claims, uncosted communication activity and generic references to policy relevance may read positively to the team that wrote it. To an evaluator working from the award criterion and its sub-criteria, it often reads as assertion without an evidential chain.

What the Horizon Europe Impact Criterion Actually Tests

Impact is an award criterion. It is separate from admissibility and eligibility checks, which determine whether the proposal may proceed to evaluation at all. A compliant proposal is not therefore a competitive proposal.

For most Horizon Europe actions, evaluators score Excellence, Impact, and Quality and Efficiency of the Implementation on the official 0-5 scale, using the descriptors in the applicable evaluation form. The usual criterion threshold is 3 out of 5, with an overall threshold commonly set at 10 out of 15. However, applicants should not treat those figures as universal. The call conditions, type of action and published evaluation form govern thresholds, weighting and any call-specific departures.

Impact is frequently decisive because it tests the proposal at the point where technical ambition meets market, policy, public-sector or societal reality. Depending on the action type and form version, evaluators will scrutinise the credibility of the pathways towards expected outcomes and impacts, the scale and significance of the contribution, and the quality of plans for dissemination, exploitation and communication. The wording is not decorative. Each limb creates a distinct route to lost marks.

A proposal may have a convincing expected result, for example a validated demonstrator, a data space, a protocol or a training method. That is not yet an outcome. The proposal must establish who adopts it, what changes as a result, what conditions are required, and why the consortium is positioned to influence those conditions. It must then show how that outcome contributes to the destination-level impacts stated in the call.

The Evaluator’s Reading Is a Chain, Not a Vision Statement

A strong Impact section permits an evaluator to follow a continuous chain from activity to result, outcome and longer-term impact. The chain need not claim control over every external condition. In fact, it should not. Evaluators know that regulation, procurement cycles, uptake incentives, standards bodies and market behaviour sit outside a consortium’s direct control.

The better approach is to distinguish contribution from attribution. State what the project will produce, identify the actors who need to use or act on it, describe the mechanism by which they will do so, and address material dependencies. If national authorities must recognise a method, say so. If a manufacturer must integrate a component after the project ends, identify the commercial or technical condition for that decision. If the pathway depends on interoperability, procurement or a forthcoming delegated act, put that dependency in the pathway rather than hiding it in a risk register.

This is where many drafts become vulnerable. They describe outputs in precise detail, then jump directly to a European-scale benefit. The missing middle is the evaluator’s problem to solve. It should be yours.

Expected impacts are not a menu of aspirations

Calls may identify expected outcomes and impacts in language that teams reproduce almost verbatim. Repeating that language demonstrates awareness of the destination, but it does not demonstrate contribution.

For each relevant expected outcome, the proposal should make clear which work package, result, partner capability and uptake route supports it. There should be proportion between the project’s funded scope and its claimed effect. A €4 million pilot cannot credibly promise to transform an entire European system during its lifetime. It may credibly generate evidence, validated methods, interoperable assets, implementation capacity or adoption commitments that contribute to a wider transition.

Overclaiming is not harmless optimism. It can weaken credibility across the criterion. A more bounded claim, supported by baseline evidence and a realistic post-project route, is usually easier to score well.

Scale and significance need a reference point

Statements such as “significant reduction”, “wide uptake” or “improved competitiveness” are difficult to score because they lack a comparator. Evaluators need to know significant compared with what, for whom, by when, and on what evidence.

Where the evidence exists, use a baseline, target population, adoption assumption and time horizon. A health innovation might distinguish trial sites from procuring health systems. A climate proposal might separate direct project effects from modelled wider replication. A digital project might distinguish registered users from active organisational deployment. These are not semantic refinements. They determine whether KPIs measure real progress towards the claimed outcome or merely record project activity.

Where Impact Drafts Commonly Lose Marks

The most expensive weaknesses are often distributed across the proposal rather than confined to the Impact section. Evaluators will test internal consistency. A pathway promising rapid market uptake is undermined by an implementation plan with no route to validation, certification, customer engagement or IP decision-making. A promise to influence public policy is undermined when no partner has access to the relevant policy forum and no engagement activity appears in the work plan.

Four patterns recur:

• Results are mistaken for outcomes. Deliverables, publications, prototypes and datasets are outputs. They only become impact-relevant when a specified user adopts or applies them.
• The target group is too broad. “Industry”, “citizens” and “policy-makers” are categories, not engagement strategies. A credible plan identifies the segment, the decision-maker and the reason to act.
• Exploitation is treated as an end-stage task. A final-month exploitation deliverable cannot resolve ownership, access rights, business model choices or regulatory constraints that should shape the technical work from the start.
• Communication is confused with dissemination. Press activity and social posts may support visibility. They rarely constitute dissemination to users who can validate, adopt, standardise, procure or replicate the results.

Not every action needs a commercialisation plan. A Research and Innovation Action, Innovation Action and Coordination and Support Action may have different expected pathways, beneficiaries and sub-criterion emphasis. A public-good outcome may require institutional uptake rather than sales. The test is not whether the proposal uses commercial language. It is whether its route to use is credible for that action and call.

Test the Evidence Behind Each Claim

Before submission, read the Impact material as an evaluator who has no access to the consortium’s internal knowledge. For each major claim, locate the supporting passage and ask three questions.

First, is the claim specific enough to assess? “Support the green transition” is a policy aspiration. “Enable 30 municipalities to procure an interoperable monitoring method, with training, implementation guidance and a maintained repository” is assessable, even if the final uptake number remains conditional.

Second, does the proposal identify the actor and mechanism? If a standard is the route to scale, which standards body, working group, contribution and partner relationship make that plausible? If exploitation relies on a spin-out, licence or product integration, who owns the foreground, who can make the decision and when will the route be tested?

Third, is the claim reflected elsewhere? The budget, work packages, milestones, governance arrangements and partner roles should all show the means to do what the Impact section promises. If they do not, the proposal has an unsupported claim, not a pathway.

This process also exposes a common consortium problem: each partner supplies credible local activity, but no one owns the European-level uptake logic. Assigning an exploitation or dissemination lead is useful only when that partner has authority, resource and a defined role in decisions that affect uptake. Otherwise, responsibility exists only in the organisation chart.

Score Against the Published Form, Not Internal Confidence

Internal reviews often ask whether the impact section is “strong”. That is too vague to guide revision. The relevant question is how the text performs against the actual criterion wording, score-band descriptors and call-specific evaluator brief.

A score below threshold indicates shortcomings that evaluators regard as serious. A score above threshold is not necessarily competitive, particularly where proposals cluster in the upper score bands. The margin between a 3.5 and a 4.5 can be a handful of unproven assumptions: an uptake actor not evidenced, a KPI without baseline, a post-project maintenance obligation without an owner, or a claimed policy contribution with no engagement route.

A disciplined pre-submission review should preserve disagreement rather than conceal it. If one reader accepts a pathway while another sees a material gap, averaging their impressions produces false comfort. The useful question is what passage generated the different readings, and whether a real evaluator could reasonably reach the less favourable interpretation.

BidShark applies separate readings against the call’s stored evaluation form, then flags supported findings and adjudicates material scoring disagreement. The automated score is not a human verdict. Where a team needs judgement on a strategic revision, the Expert Q&A package is the point at which a practising EU proposal evaluator reads the report and answers the team’s written questions.

The practical objective is not to make Impact sound larger. It is to make every consequential claim traceable to evidence, delivery capacity and a realistic route to use - before the real evaluators decide whether they can believe it.