Writing an EU proposal that scores: 12 practical rules

Write for the evaluator, not for the reviewer you imagine

• Mirror the topic language. If the topic says 'expected outcome: reduced time-to-diagnosis', use that phrase and show the number.
• Answer the template questions in the order they are asked. Creative restructuring costs you points because evaluators score section by section.
• Make every claim locatable. Bold the key sentence of each subsection so a skim-reader still finds your answer.

Excellence: ambition with evidence

Excellence is where you state the objectives, the state of the art, and what is genuinely beyond it. Weak proposals describe activity; strong proposals describe a measurable advance over a named baseline.

• Name the current state of the art with citations, then state your target in the same units.
• Declare your starting and ending TRL explicitly, and justify both.
• Be honest about risk. An acknowledged risk with a mitigation reads as competence; an unmentioned obvious risk reads as naivety.

Impact: the section most teams underestimate

Impact is not a marketing section. It is a causal chain: results → outcomes → wider impacts, each with a mechanism and a measurable indicator. Evaluators look for whether the chain would survive contact with reality.

• Distinguish results (what the project produces), outcomes (what changes for users during and shortly after), and impacts (longer-term societal or economic effects).
• Attach indicators, baselines, and target values. 'Significant improvement' scores nothing.
• Name your target users and explain why they will adopt. Include the barriers — regulatory, cost, behavioural — and how you address them.
• Treat dissemination, exploitation and communication as three different activities with different audiences, and resource them in the budget.

Implementation: make the plan auditable

• Keep work packages balanced. A WP with 60% of the effort signals poor decomposition.
• Every deliverable should have an owner, a month, a type, and a reason to exist. Delete deliverables that exist only to look busy.
• Milestones need verifiable criteria, not restated deliverable titles.
• Justify each partner's role and person-months. A partner with unclear tasks invites the question of why they are in the consortium.
• Include a Gantt chart, a PERT diagram, and an effort table. Evaluators check them for internal consistency.

The mistakes that cost points most often

• Partial topic coverage — addressing four of six expected outcomes and hoping nobody counts.
• Copy-pasted impact text that could belong to any project.
• Overloaded consortium with passenger partners added for geographic coverage.
• No data management or open science plan when the call requires one.
• Ethics and gender dimension treated as boilerplate rather than a real assessment.
• Submitting in the final hour. The portal validates late; small blocking errors surface at the worst moment.

Before you submit

• Run a mock evaluation with someone outside the consortium using the official criteria and scoring scale.
• Check the page limit — text beyond it is not read.
• Verify every figure is legible in greyscale at print size.
• Confirm consistency between the budget table, the effort table, and the narrative. Mismatches are the fastest credibility loss available.