How Grant Writing Actually Works: From Call for Proposals to Submission

Most advice on “grant writing” starts with the narrative: how to hook a reviewer, how to tell a compelling research story. That advice isn’t wrong, but it skips the step that actually decides whether your narrative gets read at all. Grant writing is not one skill — it’s four, done in sequence: reading a funding call closely enough to see what it will actually be scored against, building a narrative and budget that answer that scoring exactly, and running a realistic timeline that gets a genuinely finished draft in front of internal reviewers before the deadline panics everyone. Miss the first step and the other three are wasted effort, however well-written the prose is.

This guide walks that real lifecycle in order: from the call for proposals to the moment you hit submit.

What a Funding Call’s Evaluation Criteria Actually Tell You

Every major funder scores applications against a small, fixed, publicly stated set of criteria — and reviewers are instructed to score against those criteria specifically, not against their general impression of “good research.” Reading the call means finding that list and treating it as the actual grading rubric, because that’s what it is.

• NIH scores five criteria for most research grant mechanisms: Significance, Investigator(s), Innovation, Approach, and Environment. (The Fellowship application form uses a simplified, mostly-different framework — check which form series applies to your mechanism before assuming the five-criteria model.) Each gets its own score from every assigned reviewer, and reviewers are trained to look for language that maps to each one — a strong Approach section does not compensate for an Investigator section that never explains why this team, specifically, can do this work.
• NSF uses two agency-wide criteria applied to every proposal: Intellectual Merit (potential to advance knowledge) and Broader Impacts (potential to benefit society). Both are evaluated against the same five review elements — potential to advance the goal, creativity/originality, a well-organized plan with a way to assess success, team qualifications, and adequate resources. A proposal that treats Broader Impacts as an afterthought paragraph is being scored on a criterion it never really addressed.
• Horizon Europe (including MSCA actions) scores three criteria: Excellence, Impact, and Implementation — each typically scored out of 5, with the weighting varying by call type. See CASRAI’s Excellence/Impact/Implementation breakdown for MSCA proposals for how that structure maps onto Part B specifically.

The practical consequence: before you write a word of narrative, pull the actual review criteria for your specific mechanism and call — not a generic summary of “what funders look for,” the real, current, program-specific language, because funders revise the weighting and framework more often than applicants assume (NIH’s own Fellowship review framework and NSF’s Broader Impacts guidance have both been substantively revised within the past two years). Then draft an outline where every major section has a one-line note: which named criterion is this section primarily building evidence for? If a section doesn’t map cleanly to any criterion, it’s either padding or it belongs somewhere else.

Aligning Your Narrative to the Reviewer’s Rubric, Not Just Your Own Excitement

The single most common gap between a strong researcher and a strong grant writer is this: the researcher writes about the work the way they’d explain it to a colleague at a conference — with the details that convinced them it mattered. A reviewer scoring twelve applications in one sitting is not looking for what convinced you. They’re looking for the specific evidence their scoring sheet asks them to find, and they are actively hunting for the sentence that lets them write “strong” next to a specific criterion in their written critique.

Concretely, that means:

• Echo the funder’s own vocabulary. If NSF’s Broader Impacts guidance describes “a mechanism to assess success,” your plan should contain a section that plainly is one — not a vague promise to “disseminate results,” but a named evaluation method with a target and a way of knowing whether you hit it.
• Front-load the answer, then support it. A reviewer who has to infer your significance argument from three paragraphs of background is a reviewer who might infer the wrong one. State the gap and why it matters in the first sentences of the section that criterion maps to, then spend the rest of the section proving it.
• Make the Investigator/Environment case explicitly, not implicitly. Researchers routinely under-write these sections because the qualifications feel self-evident to them. They are not self-evident to a reviewer who has never met you and is explicitly instructed to score this as its own criterion.
• Don’t let enthusiasm substitute for a plan. “This could be transformative” is not evidence for Innovation or Excellence; a specific description of what existing approaches can’t do, and why yours can, is.

Tip: try code CASRAI at checkout for 15% off, if the offer is currently active for this program — codes vary by vendor and aren’t guaranteed.

This is also usually the point where AI-assisted drafting tools earn a mention, because rubric-alignment is exactly the kind of structural, iterative work they’re actually good at: restructuring a section so its topic sentence answers the named criterion, tightening a paragraph that buries its own point, or generating a first-pass draft of a section you already know the content of but haven’t found the words for yet. Jenni.ai is built around that kind of assisted drafting and citation-aware writing rather than one-click generation, which is the right shape of tool for this stage — you still need to know what the section has to prove; the tool helps you say it faster and revise it more times before the deadline.

Try Jenni.ai for your next draft →

Budget Justification: What Program Officers Actually Check

A budget justification is not paperwork attached to the science — it’s a second place your proposal gets tested for internal consistency, and it’s often reviewed by someone (a program officer or grants management specialist) whose specific job is to check whether the numbers actually match the narrative you just wrote.

On NIH modular budgets (the $250,000/year-or-less direct-cost format used by many R-series mechanisms): the modular format removes line-item budget detail, but it does not remove justification. Still required: a personnel justification listing every person contributing effort — PI, co-investigators, other key personnel, consultants, including “to be named” positions — by role and percent effort or person-months (without individual salary dollar figures), and, for any subaward, an itemized statement of the combined direct and F&A costs of each participating organization (typically rounded to the nearest $1,000) with a letter of intent or commitment attached. Any change in the number of modules requested between budget years has to be explained in the narrative — cost-allowability rules under the Uniform Guidance (2 CFR 200) apply in full regardless of the modular format’s reduced disclosure detail.

On NSF and non-modular budgets, the check is closer to a line-by-line audit: does the requested equipment match something the Approach section actually describes needing? Does the personnel effort requested match the roles described in the narrative? Does travel align with a dissemination or fieldwork plan that’s actually in the proposal, not a generic conference-travel line? A reviewer or program officer who spots a budget line with no narrative counterpart reads that as a proposal that wasn’t proofread end-to-end — which undercuts confidence in everything else in it.

See CASRAI’s budget justification narrative comparison for NIH vs. NSF, a worked budget justification example, and, for the allowability side specifically, writing a budget justification that survives allowability review under 2 CFR 200.

How Long Does Writing a Competitive Grant Proposal Really Take

Applicants who have never gone through a full cycle consistently underestimate this, and it’s the single most common reason a genuinely strong research idea produces a weak application: not enough real drafting-and-revision time before the deadline.

NIH’s own guidance to applicants — a reasonable proxy for the realistic minimum across funders, since NIH mechanisms are among the most heavily reviewed — suggests starting 9 to 12 months before a target deadline for a new application, especially one targeting an institute for the first time. Within that runway, NIH advises budgeting at least two months of dedicated preparation time even for a comparatively simple R01 (no vertebrate animals or human subjects), a further two months or more for the actual writing, and at least two weeks just to finalize the application — on top of whatever lead time your institution’s sponsored-programs or grants office needs to review, sign, and submit it, which is often measured in weeks, not days. The full process from initial planning to an award decision, resubmission included, can run close to two years.

The practical planning implication is not “start writing a year early” in the literal sense — it’s that the timeline has three genuinely separate phases that each need their own real time, not a shared few weeks at the end: (1) reading the call and outlining against its scoring criteria, (2) drafting and substantively revising the narrative and budget together, and (3) internal review and institutional submission processing. Compressing all three into the final month before a deadline is the single most reliable way to submit a proposal that reads, to an experienced reviewer, like it was written in the final month before a deadline.

Common Reasons Grant Applications Get Triaged Without Full Review

Many funders don’t discuss every submitted application at the review meeting. NIH’s long-standing practice is to score every application on all criteria before the meeting, then discuss only the applications projected to fall in the upper range for that meeting — the rest are marked “Not Discussed” and receive no overall impact score, only the individual criterion scores each reviewer submitted beforehand. (NIH modified this specifically in 2026, temporarily reducing the discussion rate and adding a middle “competitive but not discussed” category between “discussed” and “not competitive” — check the current policy on your specific study section’s cycle, since this is an active, still-evolving process change rather than a settled long-term rule.)

Being triaged or not-discussed isn’t usually about the science being wrong — it’s usually one of a small set of avoidable, structural problems:

• The proposal doesn’t answer the criteria it’s actually scored against. A scientifically sound Approach attached to a Significance section that never establishes why the gap matters, or an Investigator section that doesn’t address why this team specifically, reads as weak on criteria the reviewer is required to score independently.
• Non-responsiveness to the specific call. Good work aimed at the wrong funding opportunity — the wrong mechanism, the wrong program priorities, a scope the call doesn’t actually cover — gets marked non-responsive regardless of quality.
• Budget/narrative mismatch. See the section above — a budget a reviewer can’t reconcile with the plan reads as inattention, and inattention is contagious to how the rest of the application gets read.
• Feasibility gaps the narrative doesn’t address. Ambitious aims with no discussion of what happens if the primary approach doesn’t work — no alternative strategies, no acknowledgment of risk — score poorly on Approach even when the underlying idea is strong.
• Generic or templated narrative. A section that reads like it could be dropped into any proposal on the topic, rather than one that engages specifically with this funder’s stated priorities, is exactly the kind of writing an experienced reviewer has seen many times before and reads as under-invested.

From First Draft to Submission: A Realistic Sequence

Putting the pieces above together, a defensible sequence looks like this:

1. Read the call, not a summary of it. Extract the exact scoring criteria and any program-specific priorities or restrictions.
2. Outline against the criteria, noting which section builds evidence for which named criterion, before drafting full prose.
3. Draft the narrative and the budget together, not sequentially — a budget written after the narrative is “done” is where mismatches get introduced.
4. Get a real internal read from a colleague or your sponsored-programs office who will read for the reviewer’s rubric, not just for typos.
5. Revise for gaps, not just polish — the previous section’s list of common triage reasons is a useful final checklist.
6. Submit with real institutional lead time built in, not at the literal deadline.

Step 3 and step 5 are where a drafting assistant like Jenni.ai is genuinely useful rather than a shortcut: producing a fast first-pass draft of a section so you’re revising prose instead of staring at a blank page, or restructuring an existing paragraph so its argument matches the criterion it’s meant to address. It is not a substitute for steps 1, 2, or 4 — no drafting tool has visibility into your specific program officer’s private feedback from a prior submission, an unwritten institute priority a study section has been informally signaling, or the domain-specific judgment call about which risk to flag in your Approach section. A rushed AI-assisted draft that skips real input from the research team reads as generic to an experienced reviewer for exactly the reason described in the triage section above — genuinely under-invested, criterion-by-criterion, is a pattern reviewers are trained to notice, not a subtle one.

Readers working through a current draft can use the code WT10 at checkout on Jenni.ai.

Start drafting with Jenni.ai →

Go Deeper: Funder-Specific Guides

This page covers the shared lifecycle across funders. For mechanism- and funder-specific mechanics, CASRAI maintains a deep bench of dedicated guides, including:

• How to Write a Grant Proposal: A Research Administrator’s Guide — narrative components and drafting sequence in more depth
• How to Apply for an NSF Grant: A Step-by-Step Process
• NIH Specific Aims Page: Examples and How to Write One
• Horizon Europe and EU Research Funding: An Overview
• NIH Priority Score — what the number on your summary statement actually means
• NIH Program Officer vs. Scientific Review Officer vs. Grants Management Specialist — who to actually ask when you have a question about a call

Frequently Asked Questions

What a funding call’s evaluation criteria actually tell you

They tell you the exact, fixed rubric your reviewer is instructed to score against — not a general impression of quality. NIH’s five criteria (Significance, Investigators, Innovation, Approach, Environment), NSF’s two (Intellectual Merit, Broader Impacts), and Horizon Europe’s three (Excellence, Impact, Implementation) each require their own explicit evidence in your narrative; a proposal that’s strong on the science but silent on one named criterion will still score poorly on that criterion specifically.

Budget justification: what program officers actually check

Whether every requested cost has a clear counterpart in the narrative — personnel effort matching described roles, equipment matching a described need, subaward costs itemized with a letter of commitment — and, for NIH modular budgets specifically, the required personnel and consortium justifications that survive even though line-item detail is waived. See the dedicated budget justification guides linked above for worked examples.

How long does writing a competitive grant proposal really take

Realistically, months, not weeks: commonly-cited NIH guidance suggests starting 9-12 months ahead of a new-application deadline, with at least two months of dedicated preparation and two more of actual writing, plus institutional lead time before the deadline. Compressing that into the final few weeks is one of the most reliable ways to produce a proposal that reads as rushed.

Common reasons grant applications get triaged without full review

Most often: the narrative doesn’t answer the specific criteria it’s scored against, the proposal is non-responsive to the specific call’s scope or mechanism, the budget doesn’t reconcile with the narrative, the plan doesn’t address feasibility risk, or the writing reads as generic rather than engaged with this funder’s actual priorities. It is rarely simply “the science wasn’t good enough.”