The grant proposal submission checklist: 40 checks to run before you press submit

Every experienced grant team has a story about the application that failed on something trivial. A page limit exceeded by two pages. A partner from an ineligible country. A mandatory annex that existed in a folder but not in the submission. These are not the failures of careless people. They happen because in the last week before a deadline, attention goes to the parts of the document that feel important, and the rules that were read once in month one are never read again.

A checklist fixes that, but only if you run it early enough to act on what it finds.

When to run each part

1. Eligibility and administrative rules

Read these against the call text itself, not against your notes from the kick-off meeting.

1. The applicant organisation type is eligible under this specific call.
2. Every partner's country and status is eligible, including any associated or third-country participants.
3. The consortium meets the minimum composition rule, for example the required number of independent entities from different eligible countries.
4. Project duration falls inside the permitted minimum and maximum.
5. Requested budget falls inside the indicative range stated for the topic.
6. The co-financing or own-contribution rate is met and evidenced.
7. The organisation is registered on the required portal with a valid identifier, and the registration data is current.
8. No exclusion condition applies, including any restriction from previous grants.
9. All mandatory annexes named in the call exist, in the required format.
10. Any topic-specific condition, such as a mandatory sector, region, or type of activity, is met and visibly stated.

2. Coverage of the evaluation criteria

This is the single highest-value check on the page. Build a two-column table with every criterion and sub-aspect from the call on the left, and the section and page where you answer it on the right.

11. Every criterion from the call appears in the table with a location in your document.
12. Every sub-aspect within each criterion also has a location, not just the criterion heading.
13. The answer is where an evaluator would look for it, not buried in a different section.
14. Section headings mirror the call's own language, so a reader scanning for a criterion finds it immediately.
15. The weighted criteria receive proportionate space and detail, as explained in our guide to Horizon Europe evaluation criteria.
16. No criterion is answered only by cross-reference to another section.
17. Each criterion has been scored honestly by someone who did not write the proposal.

3. Impact evidence

Impact is where assertion most often replaces evidence, and where the point loss is largest.

18. Every claimed outcome names who is better off, specifically.
19. Each outcome has a baseline, meaning the current situation in numbers.
20. Each outcome has a quantified target with a timeframe.
21. Each target has an indicator and a stated method of measurement.
22. The pathway from your activities to the outcome is written out step by step, not implied.
23. Dissemination, communication and exploitation are distinguished from each other and each has its own plan.
24. Sustainability after the funding period is addressed concretely, with named owners.
25. Any claim of scale or replication is supported by something other than optimism.

4. Work plan and implementation

26. Objectives are stated as results, not as activities.
27. Every objective is delivered by at least one work package.
28. Every work package produces deliverables with dates and responsible partners.
29. Milestones have verifiable completion criteria.
30. Effort in person-months is consistent between the work package tables and the budget.
31. Risks are specific to this project, each with a likelihood, an impact and a named mitigation.
32. The Gantt or timeline matches the work package descriptions.

5. Budget

33. Budget totals match the effort and activities described in the work plan.
34. Cost categories, rates and ceilings follow the call's financial rules.
35. Every significant cost line is justified in the narrative, especially equipment, subcontracting and travel.
36. Partner budgets are proportionate to their role and visible work.

6. Consortium and partners

37. Every partner has a described role and appears in the work plan, not only in the partner table.
38. Capacity claims are evidenced with relevant past projects, staff or facilities.
39. Required letters of support, commitment or endorsement are signed and dated.

7. Document mechanics

40. Page limits, template version, fonts, margins, section numbering, figure legibility, file naming, language consistency and internal cross-references have all been verified against the call instructions on the final document.

Treat that last item as a small checklist of its own. Page limits are enforced mechanically in several programmes, and text beyond the limit is simply not read.

8. The final 24 hours

These are not checks so much as a routine that removes the last category of avoidable loss:

• Submit a complete version early. A submitted proposal can be replaced until the deadline. An unsubmitted perfect one cannot.
• Re-read the call conditions once more. Not the guidance, the conditions.
• Confirm the platform shows the proposal as submitted, and keep the confirmation.
• Archive the exact files you sent, in a folder with the date, so the next application starts from what was actually submitted.

Running this automatically

Roughly half of this list is judgement about your project, and that stays with your team. The other half is mechanical comparison between your document and the call, which is exactly the kind of work that should not depend on a tired person at 2am.

The Evalora Engine reads the official call and extracts the eligibility rules, evaluation criteria and weights, required sections, budget limits and deadlines, then checks your draft against that structure section by section. Unmet requirements are flagged, gaps are listed per criterion, and improvements are ranked by the points they recover. The evaluator simulation then predicts the panel score with reviewer-style comments, and Eva helps you implement each fix in the document.

This is also why a general chat assistant makes a poor checklist tool. A conversation does not maintain a list of forty requirements and hold each one against a hundred-page document. It answers the last thing you asked. The reasoning is set out in what ChatGPT, Claude and Gemini miss on grant proposals.

Frequently asked questions