Guide clusterGrants & Proposals
The most damaging proposal diagram mistakes are aims without deliverables, outcomes written as outputs, unlabeled connections, feasibility evidence placed away from the aim it supports, and figures that collapse at print size. Each has a specific fix, and most can be corrected in a single editing pass before submission.
Quick facts
- ClusterGrants & Proposals
- Reading time5 min
- FormatReview checklist
- Sources5 listed
Faults in the aims themselves
Aims fail in three recognisable ways. Some name activities rather than results, so the reader cannot tell what the project will leave behind. Some bundle two objectives into one aim, which makes the budget hard to justify and the reporting ambiguous. Others are so broadly worded that no single study could address them, which reads as ambition without a plan.
The repair is the same in each case. Rewrite the aim so it names one deliverable, in a form that could appear in a final report. If the rewrite produces two sentences, split the aim. If it produces none, the aim was an activity and belongs one level down, inside a method description.
| Fault | Why it costs you | Fix |
|---|---|---|
| Aim names an activity | Reader cannot identify a result to evaluate | Rewrite as aim plus named deliverable |
| Two objectives in one aim | Budget and reporting become ambiguous | Split into two aims with separate deliverables |
| Outcome written as output | Reviewer reads the project as overpromising | State who must change for the outcome to be true |
| No measurable success criterion | Panel cannot judge feasibility | Add the comparison or threshold that indicates success |
Faults in the connections
Unlabeled arrows are the most common graphic fault and the easiest to fix. An arrow without a label asks the reader to guess whether a dependency, a sequence or simply a thematic grouping is being shown. Reviewers guess conservatively, which means your strongest design element can be read as decoration.
The second connection fault is the missing dependency. Projects with genuinely sequential work often present aims in parallel, losing the chance to explain why the budget is ordered the way it is. The third is a connection that overstates causality, drawing a direct line between a method and an impact with no intermediate outcome. Inserting the missing intermediate box usually fixes both problems at once.
Label every arrow with the artifact, decision or event that travels along it.
Delete any arrow that cannot be labeled; the two boxes are probably unrelated.
Redraw sequential dependencies so the reader can see why the order is necessary.
Insert missing intermediate boxes where a jump from method to impact is too large.
Check that no line crosses another, which usually indicates a misplaced stage.
Faults in how feasibility is shown
Feasibility evidence loses most of its force when it is pooled in one place. A block of preliminary results at the top of a diagram reads as a general claim of competence. The same results attached to the aim they de-risk read as a specific answer to a specific doubt, which is how reviewers actually weigh them.
The related fault is silent confidence. When a proposal shows no risk anywhere, reviewers assume the team either has not thought about failure or has chosen not to say. A single honest fallback beside the riskiest aim does more for credibility than another paragraph of justification.
- Attach each preliminary result to the aim it de-risks rather than grouping them together.
- Name access agreements, equipment and personnel beside the step that depends on them.
- Mark one or two genuine risks and state what changes if they materialize.
- Avoid listing credentials as a general block; place expertise where it is needed.
Faults that show up only in the submission PDF
Diagrams are often built in a wide format and then squeezed into a page-limited document, where they shrink past legibility. Small gray text, thin hairlines and low-contrast color fills all survive on screen and disappear in print or in a compressed PDF. Panels frequently read on tablets and laptops of varying quality, so the figure has to hold up under the worst viewing conditions rather than the best.
Two habits prevent this. First, decide the final printed width before drawing rather than after. Second, test the exported figure at that width with the smallest label you can find. If a label fails the test, remove content rather than shrinking the type, because a reader who cannot read a box will not credit it.
Design at the final print size, not at full screen.
A pre-submission pass in fifteen minutes
Run this sequence once, in order, and resist the urge to jump to styling. Read the caption aloud, then the aim labels, then the arrow labels, then the outcome box. Each read should make sense without reference to the narrative. Then cover the narrative and ask whether the figure alone explains what will be done and what will result.
Finish by checking consistency between the figure and the text. Numbering, terminology and counts must match exactly. Mismatches are small individually, but they accumulate into a reader who starts auditing instead of evaluating, and an auditing reader is a harder reader to convince.
Worked example
Worked example: three fixes that changed a review outcome
A materials science team submitted a renewal proposal whose figure showed four aims connected by unlabeled arrows. The narrative was strong, but the panel summary used the phrase unclear whether the aims depend on one another, which is a signal that the diagram was being read as a list.
The first fix was labeling. Each arrow received a short phrase naming what it carried: alloy composition data, validated thermal model, fatigue test protocol. The panel could now see that the third aim could not begin until the second produced a calibrated model, which also explained the budget ordering that had previously looked arbitrary.
The second fix was the success criteria. Each aim had previously described what would be done, with no statement of what result would count as success. Adding a threshold or comparison to each aim gave the reviewers something to evaluate rather than merely follow.
The third fix was placement of preliminary results. Two earlier figures had been reproduced in a single corner as evidence of capability. Moving one beside the aim it de-risked and the other beside the validation step made the same evidence count twice, because each placement answered a doubt a reviewer would otherwise have raised in discussion.
Common mistakes
Frequently asked questions
How much of the panel's decision depends on the diagram?
The diagram rarely decides an outcome on its own, but it shapes the first impression that frames everything the reviewer reads afterward. A clear figure buys patience for the narrative. A confusing or missing figure makes the same narrative harder to credit.
Should I include a timeline in the proposal figure?
Usually not in the same figure. Timelines answer scheduling questions, and mixing them with logic makes both harder to read. If the funder expects a schedule, give it its own small chart and keep the logic diagram focused on dependency and evidence.
Will an AI assistant catch these mistakes for me?
It can flag unlabeled arrows and inconsistent terminology reliably, and it may notice aims that read as activities. It cannot tell whether your aims are fundable or whether your fallback is realistic, so use it as an editing aid and make the substantive judgments yourself.
Does a simpler figure always score better?
No. Simplicity helps only when it reflects a simple design. A genuinely multi-stage project needs the stages shown. The goal is not fewer boxes but a figure where every box and arrow earns its place, so a reader can follow the argument without backtracking.
Sources and further reading
- ORCID - persistent identifiers for researchersORCID
- Funding and tenders portal - application requirementsEuropean Commission
- Funding guidance and application requirementsWellcome Trust
- Grants and funding process, including application structureU.S. National Institutes of Health
- Grant application guidance and reporting expectationsBill & Melinda Gates Foundation
These are the organisations whose published guidance the recommendations on this page follow. Listing them does not imply that they endorse Pathform AI, and this page is not affiliated with them.