General guide

How to Write a Grant Proposal That a Reviewer Can Score

Most grant proposals fail on structure, not on the business behind them. Reviewers score applications against a published list, under time pressure. This guide shows you how to build a proposal around that list: the problem with proof, the plan in steps, outcomes in numbers, a budget that adds up, and a reporting promise a funder can trust.

Last updated: November 2026

The short answer

A grant proposal is not an essay about why your business deserves money. It is a set of answers a reviewer can score against a published list. Reviewers work through many applications, often with a scoring sheet beside them. Every section of your proposal should make one scored point easy to find, easy to believe, and easy to check. Write the problem with evidence, say exactly what you will do, put a price on each part, and say how you will prove it worked. If a reviewer has to hunt for any of that, your score falls, even when the project is good.

Start with the funder's scoring sheet, not a blank page

Most programs publish how applications are judged: the criteria, and often how much each one counts. That list is your outline. Before you write a word, turn each criterion into a heading and note what evidence it asks for.

This one step fixes the most common failure. Applicants write what they want to say. Reviewers score what they were told to look for. When those two lists differ, a strong business loses to a weaker one that simply answered the questions in front of it.

If the program offers a template or a set of application questions, use its order and its wording for your headings. Do not invent a cleverer structure. A reviewer moving between applications should find your answer to question four where question four belongs.

The six parts almost every proposal needs

  • The problem, with proof it is real. Name who is affected, where, and how you know. One local figure and one concrete example from your own work beat three pages of general claims. A reviewer can score "214 of our customers waited more than a week for this repair last year." Nobody can score "there is a huge unmet need."
  • What you will do. Describe the project as a short list of steps, in order. Say what happens first, what it produces, and who does the work. If the reader cannot picture the project running, it is not described yet.
  • What changes, in numbers. Pick outcomes you can count: people served, units made, hours saved, revenue or jobs created. Give each one a target and a date. "Train 40 people by March" can be checked. "Help lots of people" cannot.
  • The budget, line by line. Every cost gets its own line, a real basis (a quote, a salary, a market rate), and a short note on why the project needs it. A single round total with no workings reads as a guess, and reviewers treat it as one.
  • Why your business can deliver it. Show past work that is close to this project: the same kind of customer, the same kind of output, finished on time. Track record in the same work counts for more than awards in different work.
  • How you will report back. Say what you will measure, when, and what you will send the funder. Funders lend their money to the public story of what it achieved. A clear reporting plan tells them that story will exist.

A worked example, in miniature

Weak: "We will use the grant to grow our bakery and serve more customers in the community. This will create jobs and have a big impact."

Stronger: "The grant will pay for a second deck oven (£18,400, quote attached) and a part-time baker for six months (£9,600). The oven doubles our morning bake, from 180 to 360 loaves a day. We sell out by 10 a.m. on four days a week, so the extra loaves have buyers waiting. By the end of month six we expect two full-time jobs, up from one part-time role, and we will send the funder our monthly sales and payroll figures."

The second version is not better written. It is better evidenced. Every claim in it can be checked, priced, or counted, which is what scoring means.

Note: the currency and figures above are an illustration of the shape, not a real business. Use your own real figures, from your own records.

Write so a tired reviewer can follow you

Reviewers are people reading their twentieth application. Help them and you help yourself.

  • Put the scored point in the first sentence of each section. Add detail after it, not before it.
  • Use plain words. If a term needs explaining, explain it once, briefly. Jargon does not make a project sound serious; it makes it hard to score.
  • Keep one idea per paragraph, and keep paragraphs short.
  • Use the funder's own words for its goals, then show how your project meets them. Do not rename its priorities.
  • Cut anything that does not answer a criterion, prove a claim, or price a cost. Length is not effort in a reviewer's eyes. Relevance is.

The budget mistakes that cost the most

Budgets are where good proposals quietly lose. The recurring problems:

  • Costs that appear in the budget but nowhere in the project description, so the reviewer cannot tell what they are for.
  • Round numbers everywhere, which signal estimates where quotes should exist.
  • Staff time priced at rates the business does not actually pay.
  • No match funding shown where the program requires it, or match claimed with no source named.
  • Asking for the maximum because it is the maximum, rather than the amount the project needs. Reviewers fund plans, not ceilings.

Before you submit: the one-hour check

  • Read the eligibility rules once more and confirm, out loud, that you meet each one.
  • Check every number in the text against the budget and your records. They must agree.
  • Check that every attachment the program asks for is attached, in the format it asks for.
  • Read the proposal as the reviewer: can you find, in under a minute each, the problem, the plan, the outcomes, and the budget?
  • Submit early enough that a portal problem is an inconvenience, not the end of the application.

Frequently asked questions

How long should a grant proposal be?

As long as the program allows, and no longer than its questions need. If the program sets a word limit, treat it as part of the test: it is checking whether you can be clear in the space given. Most weak proposals are not too short. They are too vague for their length.

What is the difference between a grant proposal and a grant application?

Often nothing in practice: the application is the form, and the proposal is the written case you put in it. Some programs ask for a short letter of intent first, then invite a full proposal from a shortlist. Follow the stages the program sets out, and do not send a full proposal where a letter was asked for.

Can I reuse one proposal for many grants?

You can reuse your facts: your history, your figures, your project steps. You cannot reuse the case unchanged, because each funder scores different criteria and cares about different outcomes. Rewrite the framing for each program, keep the facts identical everywhere, and never let two versions disagree on a number.

Should I pay someone to write my grant proposal?

You can pay for help with structure and editing, and many businesses do. Be careful with anyone who charges a percentage of the grant, promises success, or wants to submit in your name without you reading it. You sign the application and you answer for every claim in it, whoever typed it.

What do reviewers mean by "measurable outcomes"?

Results you can count, with a date attached. People trained, products shipped, hours saved, money earned. If you cannot say how you would know the project worked, the reviewer cannot score that section, and unscored sections score zero.

Find grants you can apply for

Browse verified grant listings, and read each one's eligibility before you write a word.

Browse verified grants →