The Impact section of a Horizon Europe proposal, demystified

This is about the Impact section of a collaborative Horizon Europe proposal, a Research and Innovation Action or an Innovation Action (Pillar 2). Impact is one of the three scored criteria, and for an IA it carries a ranking weight of 1.5 at the panel stage, so on close calls it is the criterion that decides funding. Evaluators read your Excellence section to decide whether you can do the work. They read your Impact section to decide whether the work is worth funding at all.

Here is the fact that reframes the whole section: most of what you promise here lands after the project is over. Your outcomes begin during the project and mature in the two or three years after the final review; your impacts are long-term, five to ten years out or more. Nobody will be there to check them. So the Impact section is not judged on change you will show, it is judged on the credibility of the pathway you draw: does this project stand a real chance of producing the change it declares, on the scale the call's expected impacts announce, in line with the EU policies underneath them, across the four kinds of impact Horizon Europe recognises, scientific, economic, environmental and societal?

This is the biggest shift from Horizon 2020, and Horizon Europe puts it up front: a RIA or an IA now has to show it is not research and innovation without impact. Strong science with a weak or generic pathway is exactly the profile that fails today. And with applications surging and the bar rising call after call, the Impact section is where fundable proposals are most often lost, not because the science is weak, but because the chain between "we will build a prototype" and "Europe will be better off" does not hold.

This page walks that chain, link by link, the confusions that break it most often, and how to put numbers on it that survive scrutiny.

The impact pathway, from needs to impacts

The Impact section is a causal chain, and a specific one. It does not run through your work packages in general: it runs through the people you set out to change and the activities that reach them. Five links, in order:

1. The problem and needs of your target groups. Quantified as far as you can, because you cannot measure a change you never sized. And the target groups must be specific: not "stakeholders", but "regional transport authorities in eight EU countries".
2. Your key results. What the project produces (a prototype, a dataset, a method, a policy recommendation), and here the question is: for which target group. In the Impact view these are sometimes called your "outputs", which is not the same as the outputs of your individual tasks; here it means the substantive results that will be taken up.
3. Your dissemination, exploitation and communication activities. How those key results actually reach the target groups and get used. All three drive adoption, and communication earns its place in the pathway: a website, a demonstration event, a workshop that shows the result working all make uptake more likely, they are not decoration. This is the engine of the pathway, the reason the next link can exist at all, and it is developed in full in the dissemination and exploitation plan.
4. Your outcomes. The mid-term effects on the target groups that those dissemination, exploitation and communication activities are expected to produce: the uptake, diffusion, deployment or use of your results by the direct target groups, beginning during the project and continuing shortly after it.
5. Your impacts. The long-term, systemic change that accumulates from those outcomes, five to ten years out or more.

Every claim in your Impact section should sit on one of these links, and each link should connect to the next through an explicit mechanism, not an assertion. "Our results will lead to widespread adoption" names no target group, no dissemination activity and no timeframe: it is the empty middle of the chain. "Airport operations teams will adopt the forecasting system because the demonstration cuts their delay costs, and the exploitation work package onboards nine airports before the project ends" is a pathway.

Two confusions break this chain more often than anything else.

Confusion one: key result versus deliverable

A key result is what the project generates as substance: an algorithm, a proof of feasibility, a prototype, a demonstrator, a dataset, a business model, a policy recommendation, a set of guidelines, know-how, a trained researcher, a network. It is the thing whose use changes something in the world.

A deliverable is a report sent to the Commission for monitoring purposes: an activity report, the data management plan, an ethics report. It is an administrative artefact. It proves you did the work; it does not change anyone's behaviour.

The common mistake is listing deliverables as if they were key project outputs. "D3.2 Report on stakeholder engagement" is not a result anyone outside the consortium will ever use. If your pathway starts from a report instead of the substantive thing the report describes, the chain is broken at its first link, and the rest of the section inherits the weakness. This mistake shows up constantly in our review of typical mistakes in Horizon Europe proposals, and it is one of the easiest to fix: for every work package, name the substantive thing it generates, then build the pathway from that.

Confusion two: output and outcome

This is where our own review of proposals sees the most points lost, and it starts with a word.

The trap is "outcome" itself, and it is squeezed from two sides. On one side sits "output": in your Excellence section the substantive output of the work is called a key result, while in your Implementation section "output" gets used loosely for whatever a task produces, so the same word already points at two different things. On the other side, "outcome" literally means "a result", so it gets written as a synonym for a key result, which is exactly what it is not.

Here is the clean line. A key result (your "output", in the impact sense) is what your project produces: a prototype, a dataset, a method, a policy recommendation. An outcome is what a target group does with it: the mid-term uptake, diffusion, deployment or use of that key result, driven by your dissemination, exploitation and communication, beginning during the project and continuing after it. A key result is yours; an outcome is theirs.

A good outcome has a fixed shape, five parts: a target group, an adoption verb, the key result they take up, a quantified scale, and a timeframe. For example: "nine European airports adopt the forecasting system demonstrated during the project." Leave any of the five vague and the outcome stops being scorable.

An impact is then the systemic long-term change that accumulates from those outcomes, five years and beyond: scientific, economic, environmental or societal, measured as a shift in a macro indicator.

The subject is the tell. First the project produces, then the target group adopts, then the system shifts:

• A prototype produced is a key result.
• Nine airports using it routinely is an outcome.
• A measurable EU-wide fall in flight delays years later is an impact.

If a sentence in your outcomes reads like something the project does, it is not an outcome, it is a key result or an activity in disguise. That single slip is the most common on this criterion.

Keep one more pair apart: the call's expected outcomes and impacts are the destination the Commission has already set, your input. Your project's outcomes and impacts are what you commit to, and your job is to map the second onto the first, so an evaluator can tick each expected outcome and see which of yours serves it. Leave one expected outcome unmapped and you have handed the panel a ready-made weakness.

Where pathways fail: uptake written as aspiration

Even proposals that get the vocabulary right often fail on the mechanism. Here is a real, anonymised case.

Example: the seamless-uptake pathway

Minor shortcoming (this proposal still scored 3.5 on Impact, but the comment points at the structural axis evaluators care about most)

What the proposal wrote: "Enhanced decision-making for our target groups. Stakeholders will be empowered by a holistic, multi-actor system that establishes a seamless analytical link from the individual unit to the entire wider system."

What the evaluator saw: "The impact pathway is described with a clear logic from outputs to outcomes and societal impacts. However, the pathway assumes seamless uptake by users without sufficiently addressing structural barriers such as digital literacy, access to technology, regulatory incentives or economic constraints. The multi-actor approach is asserted but not demonstrated with concrete, step-by-step mechanisms for uptake."

The fix: an outcome is a change in someone's behaviour. So trace it per target group, with the mechanism spelled out: who does what differently, why they would (what is in it for them), and what removes the barrier standing in their way. "Operators will be empowered" is an aspiration. "Maintenance teams will integrate the tool into their planning routines because it replaces a paid service, and WP4 trains their staff to remove the digital-literacy barrier" is a pathway. The activities that carry uptake, training, piloting, standardisation, engagement with regulators, belong in the dissemination and exploitation plan, and the two sections must tell the same story.

Horizons, dimensions, areas: the measurement frame

Once the chain holds qualitatively, evaluators expect you to measure it. Three frames to apply.

Time horizons. Short term is roughly one to five years, mid term roughly three to seven, long term roughly ten. Long term differs from mid term mainly by scale, not by kind: the same change, spread across the system.

Three dimensions of measurement:

• Scale: how many people, organisations or sites you affect.
• Scale of potential: how many important stakeholders take up the results and push them further, multiplying your reach beyond the project.
• Significance: how much actually changes for those affected, often expressed as a percentage.

A 2026 change worth knowing: from Work Programme 2026 and 2027 onwards, evaluators no longer score the scale and significance of your contributions. Impact is now judged on the credibility of the pathway and the suitability of your dissemination, exploitation and communication measures. Scale and significance still help you set honest targets, but chasing a bigger number to win a mark is chasing a mark that no longer exists. Credibility is the whole game.

Three impact areas. Horizon Europe expects you to consider scientific, economic or technological, and societal impact. You do not have to claim all three, but you need indicators for every area you do claim, and a credible proposal covers at least two.

KPIs: baseline, target, deadline

KPIs are a subset of the pathway, not the point of it. A pathway with no numbers is an essay; numbers with no pathway are decoration. A usable KPI has exactly three parts: a baseline, a target, and a deadline. "A reduction in X" scores as vague. "A reduction of 15% by 2030 from a baseline of Y" is a KPI an evaluator can weigh.

The failure pattern here is asymmetric quantification, and it is remarkably common. One anonymised case makes the point.

Example: precise technology, vague society

Minor shortcoming, and it lands on the criterion where it costs the most

The proposal gave precise numbers for its technical outputs: model-error targets, alert lead times, numbers of adopters. But its societal outcomes were expressed as "improved efficiency" and "better conditions", with no baseline, no target, no deadline.

What the evaluator saw: the societal outcomes lack verifiable, time-bound indicators.

The fix: put the baseline, target and deadline on the societal rungs of the chain, because that is exactly where the Impact criterion is scored. Quantifying what your engineers control is easy; the section is judged on whether you quantify what society gains.

Where do you find these indicators? Here is the uncomfortable truth that even experienced trainers admit openly: there is no ready-made public database of impact indicators. Grant offices tell us the same thing, that for outcome and impact KPIs they have nothing to start from, so every proposal rebuilds the list from scratch. Two anchors help: align your indicators with the SDGs and relevant EU policies, and reuse mission-level and programme-level indicators wherever they exist, since evaluators already trust them.

Long-term impact: credibility is the only currency

Nobody will ever verify a claim about 2036. Evaluators know it, and you know it. That is precisely why the long-term rungs of your chain are scored purely on credibility: a plausible mechanism, sober numbers, barriers named and addressed, and a pathway where each step follows from the one before. An inflated impact claim does not impress; it signals that you do not understand how the chain is read.

And the stakes are concrete. Under how Horizon Europe proposals are scored, the Impact criterion carries a ranking weight of 1.5 at the panel stage for Innovation Actions. On close calls, where several proposals clear the thresholds, Impact is the criterion that decides funded versus rejected.

So before you submit, read your Impact section the way a panel will: link by link, asking at each one, "who uses this, what changes, and would I believe the number?"

→ Run a pre-submission evaluation in GrantForge: it scores your Impact section the way real evaluators do, in the same language, before the panel does it for you.

Part of our guide on how to write a winning Horizon Europe proposal.