I usually know by page two. Not because the rest of a proposal does not matter, but because the first two pages tell me whether the team understood the call, or wrote a general proposal and changed the title.
I have spent 25 years in international socio-economic development, across Egypt, Jordan, Lebanon, Yemen and Turkey, working on MSME finance, financial inclusion and market systems. For the last twelve of those years I have sat on both sides of the proposal. I have led technical proposals since 2014, and I have sat on a bilateral donor fund’s selection committee every year since 2014 and on more than 15 sub-grant selection committees under USAID programs.
What the Reviewer Is Checking
From the reviewer’s chair, losing proposals look alike. They are rarely badly written. They describe the sector well, list many activities and promise an experienced team. What they do not do, in the first two pages, is answer the questions the reviewer is scoring.

Seven Checks for the First Two Pages
- The problem in this call’s words. Name the problem the call names, in the place it names, with one number. “MSMEs face many challenges” could open any proposal in the region.
- One sentence on what changes. Who will do what differently by the end of the project, and how you will know. Activities come later.
- A named lead. One person, one-line specialism, one comparable result. “A team of highly experienced experts” gives the reviewer nothing to score.
- The approach in one picture or one paragraph. If the method needs three pages before it makes sense, the reviewer has already formed a view.
- Partners on the ground, named. Who is already there and what each one will do. “Local stakeholders” is not a partner.
- The biggest risk, named first. If you name it before the reviewer finds it, with one line on how you will manage it, you get the credit for seeing it.
- What the budget buys. One line that ties the money to the activities and the change. A reviewer who cannot see the link assumes there is none.
Two Patterns That Fail the First Check
Two newer patterns lose proposals before the reviewer reaches the second check. Both produce text that could belong to any call.
AI-Generated Proposals
They read fluently and say what every other proposal says: a holistic approach, a robust MEL framework, synergies leveraged. AI can draft a structure and tidy the language. It cannot tell the reviewer what you saw in that governorate, with those partners, last year. That knowledge is exactly what checks 1, 3 and 5 score, and a reviewer who reads a dozen proposals in a week notices quickly when it is missing.
Stagnant Templates
The background section from the last call, the same team paragraph, the same results language, with the donor’s name changed. Templates save time on structure, but the first two pages are where the reviewer looks for proof that you wrote for this call. A quick test: swap the country and the donor name in your first two pages. If the text still reads true, the reviewer is looking at a template.
A Before-and-After Opening
The example below is illustrative, not taken from a real proposal.
- Before: “MSMEs in the region face many challenges, including limited access to finance and weak skills. Our organization has extensive experience in this sector.”
- After: “Women-led food processors in two governorates lose most of their working capital to a three-month payment delay from wholesale buyers. Over 18 months, we will work with two local business associations and one microfinance institution so that these firms can access invoice-based credit and negotiate shorter payment terms.”
The second version answers checks 1, 2 and 5 in two sentences. The reviewer knows the problem, the change and who will carry it before reaching the second paragraph.
Reviewers rarely reject a proposal. They just cannot find what they are scoring.
Work With Me on This
I lead technical proposals and take expert roles on bids for consulting firms, NGOs and consortia, writing to what the reviewer is scoring.
For teams that bid often, my Opportunity Management System keeps calls, go or no-go decisions and submissions in one place, so the first two pages start from the call and not from the last template.
Common Questions
Whether the proposal names the call’s own problem in its place with evidence, states what will change, and shows a named lead, named partners, the main risk and what the budget buys, all in the first two pages.
Because they describe the sector and list activities instead of answering the questions the reviewer is scoring. A reviewer who cannot find the answer cannot give the points.
Name a specific group, place and problem with one number, then state in one sentence what will change, over what period and with which partners.
AI can help with structure and language, but reviewers score what is specific to the call: the local problem with evidence, named partners and a named lead with a comparable result. Text that could fit any call scores poorly, whoever wrote it.