Skip to content

Federal Grant Proposal Guide: Sections & Scoring

Updated 16 official sources

Quick Answer: In 2026, a federal grant proposal requires a project narrative, SF-424 application forms, budget with justification, organizational capability statement, and supporting documents like letters of commitment.

Full answer and how we checked it

Most competitive proposals take 100+ hours to prepare and range from 20 pages to 100+ pages depending on the agency. Reviewers score against published criteria, so structure every section to match the scoring rubric point by point.

Data note: Federal proposal guidance was checked on June 2, 2026 against Grants.gov forms and Workspace guidance, NIH simplified peer review, NSF PAPPG proposal preparation rules, 2 CFR 200 performance measurement, agency program data, USASpending.gov, and Grants.gov. Review-criteria point values for eight SAMHSA, DOJ, ED, HUD, HRSA, DOL, USDA and EPA NOFOs were checked on October 17, 2026. Amounts may shift with annual appropriations.

In this guide

If you're a nonprofit leader, researcher, or grant writer preparing your first (or fifth) federal proposal, this guide covers each section with the specifics reviewers actually score.

Learning how to write a federal grant proposal is not a matter of good writing alone. It's a matter of understanding the structure reviewers expect, the criteria they score against, and the unspoken rules that separate funded applications from the pile that don't make it.

Federal proposals range from 20 pages for a small community grant to well over 100 pages for a complex NIH application with multiple subcontracts. But the underlying architecture is consistent. Every section exists for a reason, and reviewers know what they're looking for in each one.

According to Grants.gov applicant guidance, federal applicants need the correct workspace package, registration, standard forms, and agency-specific attachments before submission. According to NIH's FY2025 extramural data, R01-equivalent success rates fell to 13.0%, so a complete application is just the entry point. The proposal still has to make the reviewer's scoring job easy.

Our analysis is that federal proposal quality is easiest to judge before a sentence is polished. For example, one applicant can have a strong program model and still lose points if the outline ignores Section E, while another can win with plain language because every scoring criterion is answered in the order reviewers expect.

What official NOFO and Grants.gov pages don't tell you is whether the proposal is score-ready before drafting. Before deciding to write, skip, or redesign, compare eligibility, award history, reviewer criteria, page limits, budget pattern, required attachments, partner commitments, local need data, and submission-system risk. Apply when the project can answer every scored criterion in the order reviewers will score it. Skip when the applicant is eligible but not competitive on evidence, budget fit, or partner commitments. The better route may be a smaller pilot, a state opportunity, a different federal mechanism, or a delayed cycle after data improves.

For example, an applicant with a strong program but no evaluation data would be a bad fit for a high-evidence federal NOFO that weighted evaluation heavily. The stronger route would be to use a smaller grant to produce outcome data, then return to the federal competition with proof reviewers could score.

This guide covers agency-specific federal proposal requirements. For the broader guide to grant proposals across all funder types, including foundations and state agencies, our complete guide to writing a grant proposal covers the universal structure every proposal needs. Last updated: October 17, 2026.

Government Grant Proposal: 10-Step Writing Sequence

Use this order for any government grant proposal. It keeps the proposal aligned with the review process instead of becoming a generic narrative.

StepOutputCommon failure mode
1Confirm eligibility, deadline, and submission systemWriting before SAM.gov, Grants.gov, or agency registration is ready
2Copy the scoring criteria into your outlineUsing a recycled structure that does not match the NOFO
3Draft a one-page fit memoChasing a program where your organization rarely wins
4Build the budget range from past awardsRequesting far above the actual award pattern
5Write the needs statement with local dataUsing national statistics without proving local need
6Map activities to outcomesListing services without a causal chain
7Draft the evaluation plan before the narrative is finalAdding weak measures after the work plan is locked
8Collect partner letters with specific commitmentsSubmitting template support letters
9Reconcile every form, attachment, and budget totalFinal-week assembly errors
10Submit early enough to fix validation errorsUploading at the deadline and losing the correction window

According to Grants.gov Workspace guidance, organizations use Workspace to complete required application forms and attach required elements. According to Grants.gov forms guidance, the SF-424 family is the government-wide form base for application packages. That means proposal writing is only half the job. Package assembly is part of the score protection plan.

Looking for your next grant?

About once a week, we email a short list of open grant calls with real deadlines from different funders. It lists open calls only.

No spam. Unsubscribe anytime.

Source Check: Grant Proposal Guidance

Checked June 2026.

Five official rules protect a proposal from administrative rejection and point loss:

Proposal controlCurrent official signalHow to use it
Forms and package assemblyGrants.gov says the SF-424 form families are the government-wide standard forms for application packagesComplete the SF-424 first, then reconcile every attachment and budget total against it
NIH research scoringNIH's simplified framework applies to most RPG applications with receipt dates on or after January 25, 2025Write Importance and Rigor as the main score drivers, and make Expertise and Resources easy to mark sufficient
NSF proposal structureNSF PAPPG says Project Description must answer what, why, how, success measure, and benefits, with a separate Broader Impacts headingPut Intellectual Merit and Broader Impacts where reviewers expect them, not in a generic narrative
Federal performance measurement2 CFR 200.301 says agencies must measure recipient performance against program goals and expected outcomesBuild outputs, outcomes, baseline data, and reporting responsibilities before drafting the evaluation section
Submission roleGrants.gov applicant guidance says only an Authorized Organization Representative can submitConfirm roles before the final week, not after the narrative is done

Grantsights' June 2026 proposal rule: do not start with prose. Start with the scoring table, SF-424 package, outcome measures, budget basis, and submission roles. Then write the narrative in the order the reviewer will score it. Use the grant proposal outline builder, budget justification guide, evaluation plan guide, SAM.gov registration guide, and grant database before committing 100+ hours to the wrong route.

The Universal Sections: What Every Federal Proposal Needs

These sections appear in virtually every federal grant application regardless of agency, program, or funding amount.

The SF-424: Your Administrative Cover Sheet

The SF-424 is the standard cover form required for all Grants.gov (opens in a new tab) submissions. Reviewers don't score it. But errors on the SF-424 can get your application rejected before a single reviewer ever reads your narrative.

The Division of Receipt and Referral at NIH, and equivalent intake offices at other agencies, screen submissions before they reach review panels. Applications with administrative errors on the SF-424 are returned without review.

The fields that cause the most problems:

Applicant legal name. This must match your organization's registration in SAM.gov exactly. Not your program name, not your DBA, not your common name. The legal registered name as it appears in SAM.gov. A single word difference can trigger rejection. Before you submit, pull up your SAM.gov entity record and copy the name character by character. Learn how to register on SAM.gov before you start an application.

CFDA/ALN number. The Assistance Listing Number identifies the specific program you're applying to. Don't guess it - pull it directly from the Notice of Funding Opportunity (NOFA) or Funding Opportunity Announcement (FOA).

Proposed start and end dates. These must match the project period specified in the FOA. Don't extend the project period beyond what the FOA allows.

Total federal funds requested. This must match your detailed budget exactly. A $1 discrepancy between the SF-424 and your budget spreadsheet is a red flag that reviewers and grants management staff will notice.

According to Grants.gov's standard forms library, SF-424 family forms are still the administrative base for most federal applications. Treat them like a compliance screen, not like paperwork you can delegate at the end.

The Abstract / Project Summary

Most applicants write the abstract last and treat it as a formality. That's a mistake.

The abstract is often used for administrative routing before your application reaches reviewers - it determines which program officer and which review panel your application is assigned to. If the abstract doesn't clearly signal your project's content, your application may land with reviewers who aren't equipped to evaluate it.

At NIH, every member of the study section can read the abstract. Your assigned reviewers (typically 2 or 3) will read your full proposal, but the other members may form initial impressions from the abstract alone. Those impressions influence the discussion during the review meeting.

At community-based agencies, the abstract sometimes travels to agency leadership who make final funding decisions without reading the full application.

Write the abstract last, but don't write it carelessly. A tight abstract states four things: the problem you're addressing, your approach, who benefits, and the expected outcomes. Keep it to the page limit. Cut everything that doesn't carry weight.

The Project Narrative

This is the substantive core of your proposal. Different agencies call it different things - Research Strategy at NIH, Project Narrative at SAMHSA and DOJ, Project Description at NSF. The name changes, but the function doesn't: this is where you explain what you're doing and why it will work.

According to NSF proposal-preparation guidance, a standard research proposal's project description is typically up to 15 pages and must explain what you plan to do, why, how, how you will know whether it works, and what benefits could result. That same reviewer logic applies even when another agency uses a different attachment name.

The project narrative typically includes:

  • Statement of need or problem statement
  • Description of the intervention or research approach
  • Evaluation plan
  • Organizational capacity
  • Work plan and timeline

Some agencies separate these into distinct attachments with individual page limits. Others want them folded into a single narrative document. Read the FOA carefully - page limit violations can result in rejection before review.

The Budget and Budget Narrative

The budget is a line-item spreadsheet, typically using form SF-424A for non-construction projects. The budget narrative (also called the budget justification) is the written explanation for every line item.

Reviewers score the budget narrative, not just the spreadsheet. A number without an explanation is a problem. Every cost needs to be shown as allowable (permitted under federal regulations), allocable (directly benefiting this project), and reasonable (what a prudent organization would pay).

For a full guide to writing the budget justification, see how to write a grant budget justification.

Key Personnel and Biographical Sketches

Most federal agencies require abbreviated CVs (called biosketches or biographical sketches) for key personnel. NIH has its own biosketch format, and it's not interchangeable with NSF's or with a standard resume.

Key personnel includes the principal investigator (PI) and co-investigators. Staff listed as key personnel can't be changed after award without prior agency approval. That's worth thinking about before you designate someone as key.

Organizational Capacity

This section documents that your organization has the infrastructure, experience, and administrative systems to manage a federal award. Reviewers want to see relevant prior experience (ideally with similar federal grants), your financial management systems, your compliance history, and your organizational chart.

Don't confuse this with the needs statement. Organizational capacity is about YOU. The needs statement is about the PROBLEM.

Evaluation Plan

You'll need to describe how you'll measure whether your program worked. This means specifying outcomes, data collection methods, instruments, timeline for data collection, and who's responsible. Many agencies require both process evaluation (did you do what you said?) and outcome evaluation (did it work?).

For a full breakdown, see how to write a grant evaluation plan.

According to 2 CFR 200.301, federal agencies should establish program goals and objectives and measure recipient performance against them. In practice, that means your evaluation section should map every major activity to a measurable output or outcome, not just promise a final report.

Assurances and Certifications

These are the legal attestations required by federal law - certifications regarding lobbying, debarment, drug-free workplace, civil rights compliance, and others. Most are submitted through Grants.gov checkboxes or standard forms. They're not scored, but missing ones will stop your application.


Program-Specific Sections: What Some Agencies Require

These sections are required by specific agencies or program types. Not every proposal needs all of them, but you'll need to know which apply to your application.

NIH Specific Aims Page

The Specific Aims page is the most important single page in any NIH application. It's the only page every member of the study section reads. The assigned reviewers (2-3 people) read your full Research Strategy, but the other 20-plus members of the study section form their opinions of your proposal largely from this page.

The standard structure:

Paragraph 1: The gap. What is the unresolved problem in the field? What do we not yet know, and why does it matter? This paragraph establishes significance without the word "significant."

Paragraph 2: Your approach and central hypothesis. State your hypothesis directly. Reviewers shouldn't have to infer it. Your long-term goal, your central hypothesis, and the rationale for your approach all go here.

The aims list. Three aims is standard for an R01. Each aim should be a specific, testable question. Write them as active, hypothesis-driven statements: "To determine whether X" rather than the passive "To characterize X." The aims should be related but not completely dependent - if Aim 1 fails, Aims 2 and 3 should still be viable.

Closing paragraph: payoff and impact. What does the field look like if your aims are achieved? What does success mean for patients, for the scientific community, for public health?

If you can find someone who has served on an NIH study section to give feedback on your Specific Aims page, get it. They understand how reviewers read it in a way that even experienced grant writers sometimes don't.

Logic Model

Some NOFOs score the logic model directly. ED's FY2026 Innovative Approaches to Literacy notice gives 10 of its 50 design points to "the quality of the logic model or other conceptual framework." The other seven NOFOs in the scoring table below, including SAMHSA's and DOL's, don't ask for one, so check your NOFO rather than assuming. NIH and NSF don't require them.

A logic model is a visual framework - usually a table or flowchart - that shows how your inputs and activities produce outputs and lead to outcomes. For a complete guide on building one, see what is a logic model for grants.

Dissemination Plan

NSF doesn't score a separate dissemination plan. Reviewers weigh how you'll share results under the two merit review criteria, Intellectual Merit and Broader Impacts, and NSF requires a Data Management and Sharing Plan. If you're applying for an NSF grant, don't treat dissemination as an afterthought. Describe specifically how you'll share results: journals, conferences, public databases, community reports, and timeline.

NIH Data Management and Sharing Plan

Since January 2023, NIH requires a Data Management and Sharing Plan (DMSP) for all research generating scientific data. The plan describes how you'll manage, preserve, and share your data. It's reviewed as an attachment - not scored by reviewers, but reviewed by NIH program staff and is a condition of award.

Human Subjects / IRB

If your project involves research with human subjects, you need to address it in your application. For NIH, this means completing the Human Subjects sections including protection of human subjects, inclusion of women, minorities, and children, and IRB status. Review-ready applications have at least a pending IRB determination - reviewers notice when it's missing.

Sustainability Plan

Many community-facing grant programs want to know what happens to the program when the federal funding ends. HRSA's FY2026 Delta States Rural Development Network NOFO, for example, scores the sustainability plan inside its 20-point Impact criterion. A sustainability plan describes your strategy for continuing services after the grant period - through other funding, institutional absorption, fee revenue, or state/local funding.

Be concrete. "We will seek additional funding" isn't a plan. Identify specific sources, describe your development strategy, and show that you've already taken steps toward sustainability (conversations with foundations, letters of intent from local government, etc.).

Letters of Support

Letters of support from community partners, referral agencies, and local organizations signal to reviewers that your project has buy-in from the people it needs to succeed. A letter that says "we support this project" is nearly worthless. A letter that commits specific resources - staff time, facility space, client referrals, data access - adds real weight.


Proposal SectionWeight in the 8 point-scored NOFOs belowTime to Allocate
Specific Aims / Project SummarySets tone; not a scored criterion in any of the 830-40% of writing time
Approach / project design25% (HRSA, USDA) to 50% (ED) of the score; no separate factor in HUD's30% of writing time
Organizational Capacity12.5% (EPA, 25 of 200) to 60 of 102 points (HUD)10% of writing time
Budget and JustificationNot scored in 3 of 8 (SAMHSA, ED, HUD); 4% (DOL) to 15% (OJJDP) in the rest10% of writing time
Letters of SupportNo separate points, but EPA's 40 partnership points and DOL's 20 partnership points are judged on partner commitment documents5% of writing time
Evaluation PlanNo separate criterion in 3 of 8 (OJJDP, ED, HUD); 7.5% (EPA, 15 of 200) to 20% (HRSA, USDA) in the rest10% of writing time

How Reviewers Score Proposals

Understanding the scoring system for your specific agency is non-negotiable. Here's how the major agencies score applications.

NIH Scoring (Effective January 25, 2025)

NIH implemented a new simplified scoring framework per Notice NOT-OD-24-010, effective for applications submitted on or after January 25, 2025.

According to NIH notice NOT-OD-24-010, the simplified framework applies to many research project grant activity codes for receipt deadlines on or after January 25, 2025. That is why old R01 templates that still lead with separate Significance, Investigator, and Environment scoring language should be rewritten before reuse.

Factor 1 - Importance of the Research (scored 1-9): This combines what was previously Significance and Innovation. Does the project address an important problem? Does it offer a novel approach or advance scientific knowledge?

Factor 2 - Rigor and Feasibility (scored 1-9): This is the former Approach criterion. Is the research design scientifically rigorous? Are the methods appropriate? Can the team realistically execute the work in the proposed timeframe?

Factor 3 - Expertise and Resources (binary - Sufficient / Not Sufficient): This replaces the separately scored Investigators and Environment criteria. The key change: investigator reputation and institutional prestige no longer get a numerical score. The intent is to reduce bias that historically boosted scores for famous PIs at elite institutions.

The overall impact score is a separate judgment from 1 to 9 that synthesizes all factors. With an R01-equivalent success rate of 13.0% in FY2025, per NIH success rate data (opens in a new tab), only the strongest percentiles are realistically fundable, and NIH stopped using paylines in January 2026, so there's no published cutoff to check. Ask your program officer how your score compares with recent funding decisions.

NSF Scoring

NSF uses two criteria, weighted equally:

Intellectual Merit: The potential to advance knowledge within and across fields.

Broader Impacts: The potential to benefit society and contribute to important national goals.

Every section of your NSF proposal should be written with these two criteria in mind. Reviewers are explicitly asked to evaluate both.

SAMHSA, DOJ, ED and Other Point-Scored NOFOs

Community-program agencies publish their review criteria and point values in each NOFO, and the weights differ more than most templates assume. Grantsights checked the review criteria in each NOFO below on October 17, 2026. Each cell gives the NOFO's own criterion name and its maximum points; where one criterion spans two columns, the NOFO's sub-criterion points are used.

AgencyProgram and NOFONeed / problemApproach / designEvaluation / performanceOrg capacity / personnelBudgetOtherTotal
SAMHSAMental Health Awareness Training, SM-26-030 (FY2026) (opens in a new tab)Population of focus and need: 25Proposed implementation approach: 35Data collection and performance measurement: 10Organizational experience and staffing: 30Not scoredNone100
DOJ (OJJDP)Mentoring for Youth Impacted by Opioids and Other Substance Use, O-OJJDP-2026-172743 (FY2026) (opens in a new tab)Description of the need: 5%Project design and implementation: 35%No separate criterionCapabilities and competencies: 30%Budget: 15%Project goals and objectives: 15%100%
EDInnovative Approaches to Literacy, 84.215G, DOL-OESE-33831 (FY2026) (opens in a new tab)Need for the project: 10Quality of the project design: 50No separate criterion (10 design points cover measurable targets and data)Quality of the management plan: 40Not scoredCompetitive preference priorities: up to 15 bonus100 + 15 bonus
HUDLead Hazard Reduction, LHC-2600-DC-0013 (FY2026) (opens in a new tab)Justification of applicant need: 40No separate factor (target area and work plan are 10 of the capacity points)No separate factorCapacity of the applicant and relevant organizational experience: 60Not scoredOpportunity Zone preference: 2102
HRSADelta States Rural Development Network Program, HRSA-26-045 (FY2026) (opens in a new tab)Need: 15Response: 25Performance management: 20Resources and capabilities: 15Support requested: 5Impact, including sustainability: 20100
DOL (ETA)Workforce Opportunity for Rural Communities Round 7, FOA-ETA-26-23 (FY2025 funds, posted June 2026) (opens in a new tab)Statement of need: 20Project design: 30Expected outputs and outcomes: 10Organizational capacity: 16Budget and budget narrative: 4Sector partnership and industry engagement: 20100
USDA (AMS)Specialty Crop Multi-State Program, USDA-AMS-TM-SCMP-G-26-0020 (FY2026) (opens in a new tab)Alignment and intent: 25Technical merit: 25Achievability (outcomes and indicators): 20Expertise and partners: 20Fiscal plan and resources: 10None100
EPAFY27 Brownfields Job Training Grants, EPA-OLEM-OBLR-26-01 (opens in a new tab)Community need: 40Training program description: 25; program structure (recruitment, job placement, continuity): 35Outputs and outcomes: 15Programmatic capability: 25Budget: 15Partnerships: 40; leveraging: 5200

Notes on the table: OJJDP states its weights as a percentage of the application score rather than points. DOL's 16 capacity points and 4 budget points are the two parts of its 20-point "Organizational, Administrative, and Fiscal Capacity" criterion. EPA's 35 program-structure points and 15 outputs-and-outcomes points are the parts of its 50-point "Program Structure, Anticipated Outputs and Outcomes" criterion. The ED notice was posted by the Department of Labor on ED's behalf, as the notice's cover page states.

Across these eight notices, approach or design carries the most points in six, ties for the most in USDA's, and HUD's Lead Hazard Reduction NOFO is the outlier, putting 60 of its 102 points on organizational capacity and experience. Organizational capacity outscores evaluation in six of the eight, and three (OJJDP, ED and HUD) don't score evaluation as its own criterion at all; only HRSA's Delta States NOFO gives performance management (20) more points than capacity (15). These are patterns in eight notices, not a rule for every program, so copy the criteria table from your own NOFO into your outline before you write a word.


Page Limits Are Not Suggestions

This deserves its own section because it's the fastest way to get an application rejected before review.

NIH applications exceeding page limits are rejected by the Division of Receipt and Referral before they reach the study section. There's no appeal. The application simply doesn't get reviewed.

NIH R01: 1 page for Specific Aims, 12 pages for Research Strategy (covering Significance, Innovation, and Approach combined as of the 2025 framework).

NSF standard research proposal: 15 pages for the Project Description.

Community-program NOFOs: Limits vary widely, even within one agency's portfolio. SAMHSA's FY2026 Mental Health Awareness Training NOFO caps the Project Narrative at 10 pages. DOL's WORC Round 7 announcement limits it to 20 double-spaced pages. ED's FY2026 Innovative Approaches to Literacy notice recommends no more than 25 pages. HRSA's Delta States NOFO caps the application at 50 pages of attachments (the forms themselves don't count), and OJJDP's FY2026 mentoring NOFO swaps the narrative for a Proposal Questionnaire answered in 1-2 paragraphs per question. Don't assume. Read the page limit instructions in the NOFO before you start outlining.

Page limits apply to text, figures, tables, and graphs - not just prose. Many applicants underestimate how much space figures take. Build your narrative in your final font size and margins from day one so you're not reformatting at midnight before the deadline.


The 80/20 Rule in Federal Grant Writing

A useful rule of thumb: most of your score is determined by a small share of your content.

The three sections that determine most of your score:

The problem statement: Does the need exist? Is it documented with credible data? Is it specific to the population you serve? Reviewers who aren't convinced by your needs statement don't give your intervention logic a fair read. See how to write a grant needs statement for the full breakdown.

The intervention logic: Does your proposed approach actually address the problem you documented? Is there evidence it works? Is the causal chain clear?

The evaluation plan: Can you prove whether it worked? Do your outcomes connect to your needs statement? Are your measures specific and your data collection methods described?

Reviewers read dozens of proposals in compressed time. Applications that make reviewers work to extract the key logic - burying the central hypothesis on page 6, presenting the intervention before the need is established, leaving the evaluation plan vague - lose points not because reviewers disagree, but because reviewers can't follow.

Write for a smart non-expert. Your assigned reviewer may be in your broad field but not your specific niche. If a well-informed person who doesn't know your subfield can't follow your logic, your proposal has a problem.


Writing for the Reviewer's Experience

Use headers and subheaders so reviewers can go to the scoring criteria they're responsible for. Use bullet points for lists of activities, personnel, or data sources. Don't make reviewers hunt through dense paragraphs to find the answer to a specific scoring criterion.

State your central hypothesis or core program logic in the first page of the narrative. Not page 4. Not after three pages of background. Reviewers who don't understand your core idea in the first page may form negative impressions that carry through the rest of their read.

Don't bury your strongest argument. If you have strong local data showing a gap that your program uniquely addresses, lead with it. If your PI has published the foundational studies in this area, make that clear early.


Should You Contact the Program Officer Before Submitting?

Yes, and most applicants don't do it enough.

Program officers (POs) are federal staff who manage grant programs. They're not reviewers - they don't score applications - but they're an invaluable source of information before you submit. They can tell you:

  • Whether your project fits the program's priorities
  • Whether a specific approach is likely to be competitive
  • What reviewers have flagged as weak in past cycles
  • Whether the funding level you're considering is appropriate

Contact the PO listed in the FOA at least 4 to 6 weeks before the deadline. A brief email describing your project and asking whether it fits the program's intent is appropriate. Don't send a draft proposal. Don't ask for a commitment. Just confirm fit.

Most POs respond. Those who don't respond before the deadline will often take a call if you're persistent. It's worth the effort.

This is also the point where mechanism confusion gets fixed. A short call can tell you whether you should be drafting against an NSF CAREER route instead of a broader NSF program, an NIH R15 instead of a larger NIH mechanism, Head Start instead of a generic ACF search, or Continuum of Care instead of a generic HUD funding hunt. Education applicants should also separate direct competitions from formula flows by checking Title I Grants to Local Educational Agencies and Special Education Grants to States. On the NIH side, that same conversation can clarify whether a project belongs under an R21, K99/R00, K22, R03, or K18 mechanism. Use the federal grant database to compare these pages before you outline the wrong proposal.


Why Federal Proposals Fail in the Final Week Before Submission

A common failure pattern isn't a weak needs statement or a vague evaluation plan. It's a proposal that was written well but assembled wrong in the final 72 hours. For example, an SF-424 budget total that doesn't match the budget detail spreadsheet, even by a few hundred dollars, can get an application flagged at intake.

The FY2025 NIH data make the competitive pressure concrete: R01-equivalent success rates dropped to 13.0% in FY2025 (5,471 awards from 42,022 applications), down from 18.7% in FY2024, per NIH success rate data (opens in a new tab). When success rates were higher, a tightly written application with strong preliminary data had a realistic shot. In the current environment, a tightly written application with strong preliminary data may still not be in the top 13%.

What those statistics mean practically is that reviewers can't afford to give marginal proposals the benefit of the doubt. Under NIH's Simplified Peer Review Framework effective January 25, 2025, Expertise and Resources is now a binary (Sufficient / Not Sufficient) rather than a scored criterion. That change was designed to reduce bias toward famous PIs at elite institutions, but it also means your Importance and Rigor scores carry even more weight than they used to. A PI can be clearly capable but fail to demonstrate it in the writing itself. Reviewers won't assume competence from a CV entry they don't immediately see in the proposal narrative.

The program officer contact is an often-skipped step. Program officers don't score your proposal, but they do know what the study section has been looking for and what's been missing across the cycle. A program officer can tell you what a review panel is particularly interested in that cycle, and the conversation costs little more than a short phone call.

For community-facing agencies like SAMHSA, DOJ, and ED, 2 CFR 200.301 (opens in a new tab) says the federal agency must measure recipient performance to show achievement of program goals and objectives, and should spell out in the award the expected outcomes, indicators, targets, and baseline data the recipient must measure and report, plus any required participation in federally funded evaluations. Expect those terms to show up in NOFO evaluation requirements. If your program uses an evidence-based approach, cite where it's documented, such as the What Works Clearinghouse or the SAMHSA Evidence-Based Practices Resource Center, and name the outcomes, indicators, and baseline data you'll report.

What Federal Peer Review Panels Don't Tell You

Reviewers read your specific aims page first and sometimes only that. At NIH, if the specific aims page doesn't convince the reviewer, the rest of the application gets a surface-level read at best. Spend 40% of your writing time on that single page.

Budget justification is scored even when the NOFO doesn't list it as a criterion. Reviewers use the budget to gauge whether you actually understand the project scope. An unrealistically low budget raises as many concerns as a padded one.

Letters of support that use identical language across multiple partners signal a template job. Reviewers notice when three community partners submit the same letter with only the letterhead changed. Each letter should reference specific deliverables that partner will contribute to.

Sources


If you're building a federal proposal from scratch, the budget justification is the section most applicants underestimate. See how to write a grant budget justification for the full breakdown of every budget category and the "reasonable, allowable, allocable" standard under 2 CFR 200.

For the needs statement specifically - which determines a significant portion of your score at most community-based agencies - see how to write a grant needs statement.

For program evaluation structure, including SAMHSA National Outcome Measures and DOL performance reporting requirements, see how to write a grant evaluation plan.

Federal grant writing is learnable. It takes time to understand the conventions of each agency, but the core logic - document the need, explain the approach, prove you can measure the results, show you can manage the money - holds across programs. Start with the FOA. Know your scoring criteria. Write for the reviewer.

Next steps

Where to go next

These next pages help you move from research into a specific grant choice.

Frequently Asked Questions

How many hours does it realistically take to write a competitive federal grant proposal?

A competitive federal proposal takes 100 to 250 hours of total staff time, depending on agency and complexity. NIH R01 applications average 150 to 200 hours because the 12-page Research Strategy requires extensive literature review and preliminary data analysis. SAMHSA and DOJ community grants run 80 to 150 hours for a shorter narrative (SAMHSA's FY2026 Mental Health Awareness Training NOFO caps it at 10 pages) plus budget, timeline, and supporting documents. NSF proposals take 80 to 120 hours for the 15-page Project Description. These estimates include drafting, internal review cycles (plan for at least 3 rounds of feedback), budget development, and partner coordination. First-time applicants should add 30-50% more time. Start 4 to 6 months before the deadline, not 4 to 6 weeks.

What percentage of federal grant proposals actually get funded?

It varies widely by agency and program. NIH's R01-equivalent success rate was 13.0% in FY2025, down from 18.7% in FY2024. NSF awards 25-28% of proposals across all directorates. SAMHSA, DOJ, and Department of Education rates vary by program and year, so check the number of expected awards in the specific NOFO. But here's what most guides won't tell you: some applications are eliminated before serious review due to administrative errors, missing documents, or failure to address required criteria. A complete, responsive application avoids that avoidable loss.

What's the single most common mistake that causes federal grant proposals to be rejected?

Failing to match your narrative structure to the scoring rubric. Most community-program NOFOs publish review criteria with point values. SAMHSA's FY2026 Mental Health Awareness Training NOFO (SM-26-030), for example, gives 25 points to need, 35 to the implementation approach, 30 to organizational experience and staffing, and 10 to data collection and performance measurement. NIH and NSF score differently, without point tables. Applications lose points when they bury answers to scoring criteria inside unrelated sections or skip criteria entirely. Use the exact headings from the NOFO as your section headers. If the approach is worth 35 points and data collection is worth 10, give the approach about 3.5 times the page space. Administrative rejections (SF-424 errors, expired SAM.gov, exceeded page limits) are a separate issue that can eliminate a submission before review.

Should I hire a grant writer or write the federal proposal myself?

If you're applying for your first federal grant and the award is under $250,000, writing it yourself is reasonable if you budget 120+ hours and get feedback from someone who's won federal grants before. For awards over $500,000 or agencies with complex requirements (NIH, SAMHSA), professional help pays for itself. Fees for experienced federal grant writers vary widely, with NIH applications at the high end due to scientific complexity. The ROI math is straightforward: if a writer increases your odds from 15% to 30% on a $500,000 grant, the expected value gain is $75,000. Many nonprofits hire full-time grant staff (salary $55,000 to $85,000) once they're pursuing 3 or more federal applications per year.

How different are NIH, NSF, and SAMHSA proposals from each other?

Very different, and using the same template across agencies is a common mistake. NIH R01 proposals center on a 1-page Specific Aims document and 12-page Research Strategy scored on Importance, Rigor/Feasibility, and Expertise. NSF proposals are 15-page Project Descriptions scored equally on Intellectual Merit and Broader Impacts. SAMHSA uses point-based scoring: its FY2026 Mental Health Awareness Training NOFO scores a 10-page Project Narrative on 100 points across four sections (need 25, implementation approach 35, organizational experience and staffing 30, data collection and performance measurement 10) and scores a timeline attachment. Other required sections, such as a logic model or sustainability plan, vary by NOFO. Budget rules differ too: NIH uses modular budgets under $250,000 per year, NSF caps PI salary at 2 months across all NSF awards, and SAMHSA requires line-item justification for every cost. Read the FOA for your specific program before outlining anything.

Last updated October 17, 2026. Corrected October 17, 2026: what changed. Published by Grantsights from public sources. See our editorial policy, how we build and check guides, corrections and how we make money.

Guides in This Series

How to Write a Grant Proposal: Complete 2026 Guide

How to write a grant proposal for foundations, community foundations and corporate funders: 8 sections, budget tips, and what 10 funders require.

How to Write a Grant Cover Letter (With Examples)

Many federal programs and most foundations expect a grant cover letter. Learn the 1-page, 4-paragraph format and who should sign it, with 3 examples.

How to Write a Grant Evaluation Plan (With Examples)

Only 3 of 37 open federal NOFOs scored a separate evaluation plan. Write SMART outcomes, baseline plans, and evaluation budgets that reviewers score.

How to Write a Grant Needs Statement That Scores Well

The needs statement is one of the scored criteria in most NOFOs. This guide covers the 4-part structure, data sourcing, and what reviewers look for.

How to Write a Data Management Plan for Grants

NIH, NSF, DOE and NASA each want a data management and sharing plan. Compare 5 agencies' rules, page limits and review, plus 2 templates.

How to Write a Grant Abstract (With Template)

Grant abstracts must fit strict limits: NIH 30 lines, NSF 1 page, Grants.gov 4,000 characters. Get the exact structure reviewers want.

How to Write a Grant Budget (Line-by-Line Guide)

Federal grant budgets have 8 required categories under 2 CFR 200. Walk through each line item with a complete $379,870 sample budget table.

How to Write a Grant Dissemination Plan (With Examples)

Write a dissemination plan that scores well with federal reviewers. Covers 6 required components, audience targeting, and budget for 5+ methods.

How to Write a Grant Letter of Inquiry (LOI)

A grant LOI is 1-3 pages that determines whether you get invited to submit a full proposal. Here's the exact structure with a section-by-section template.