The one-page product plan: the seven answers before anyone builds anything
Seven answers before anyone builds: customer, pain, promise, price, proof, channels, budget. The template and a filled example for a bookkeeping client portal.
A one-page product plan answers seven questions before any money is spent: who the customer is, what pain they pay to remove, what you promise, what it costs, what proof you can show, which channels reach the customer, and what the budget and the stop rule are. If any of the seven is blank, building is early.
The page takes an evening, and it is the document a builder, a freelancer or a done-for-you team works from.
What are the seven answers?
The template below is filled in for a client portal for bookkeeping firms, a product that keeps documents, deadlines and invoices in one place for each client.
| # | Question | Example answer, bookkeeping client portal | Why it matters |
|---|---|---|---|
| 1 | Customer: who pays, by name and number | Owners of bookkeeping firms with 5 to 50 clients in the US and UK, on a list you can count before the build starts | 43% of venture-backed companies that shut down since 2023 cited poor product-market fit (CB Insights). A named group can be counted and reached |
| 2 | Pain: what costs them time or money today | Chasing clients for documents by email every month, missed filing deadlines, invoices sent from a third tool | Pain that already costs money is paid to remove. Convenience is not |
| 3 | Promise: one sentence, in their words | "Every client's documents, deadlines and invoices in one place, before the deadline" | The promise becomes the headline, the ad and the first email. One idea per sentence |
| 4 | Price: a number on the page | USD 90 a month per firm, shown on the page | Hiding the price makes visitors leave and worsens the eventual sticker shock (Nielsen Norman Group) |
| 5 | Proof: what a stranger can check | A working portal they log into on a 20-minute call, the written process, a refund line | A stranger can check a live product and a written process. They cannot check claims |
| 6 | Channels: where the customer already is | Named firm owners messaged one by one; search ads with a daily cap; the owner's own clients | Cold email replies average 3.43% (Instantly) and 5.8% on lists under 50 contacts (Belkins via Woodpecker); a business-services search lead costs USD 93.69 (LocaliQ) |
| 7 | Budget and stop rule | USD 1,000 for 30 days: the build, the domain and USD 300 of search ads. It works at 10 requests, 3 calls and 1 paying firm by day 30 (the verdict numbers); otherwise change the pitch once, then stop | 70% of the failed companies in the CB Insights set ran out of capital. A stop rule written before day 1 is the cheapest insurance there is |
Why does each answer matter?
Customer before product. The product is the last thing you decide, because every other line changes it. "Small businesses" is not a customer; "bookkeeping firm owners with 5 to 50 clients" can be listed, counted and messaged.
Pain before promise. If the pain does not already cost money or hours, no promise fixes the sale. Write it in the words you heard, not in yours.
Price on the page. Nielsen Norman Group's advice for business buyers is to show the price, or at least a starting price or a range, because hiding it adds effort, sends people to competitors and makes the company look as if it is hiding something.
Proof you can give. Before the first customer you still have a working product, a written process and a guarantee. Those can be checked. Claims cannot.
Channels you can count. Pick two that produce a number every day. Lovable's survey of more than 400 builders found that a third did not know where to find their customers and another third had only posted on social media and asked their network; a plan with "social media" as the channel has no channel. Unbounce's 2024 data puts the median landing page in professional services at 6.1%, so 300 visitors should give about 18 requests, and 10 is a fair floor.
Budget with a stop rule. It is the line most plans leave out and the one that saves you. Decide now what a miss looks like and what you do then.
How do you fill it in one evening?
- Write the customer and the pain from three real conversations, not from memory.
- Write the promise in one sentence of fewer than 12 words. Read it aloud.
- Set the price by what the pain costs them each month, then check it against two things they pay for today.
- List the proof you can produce this week.
- Name two channels and the number each should produce in 30 days.
- Set the budget, the bar and the date of the verdict.
- Give the page to whoever builds, and change nothing for 30 days except what the numbers tell you.
What does a weak plan look like?
Customer: "anyone who needs bookkeeping". Pain: "it is hard". Promise: a paragraph. Price: "we will see". Proof: "coming soon". Channel: "social media". Budget: none, so the project ends when the money or the energy does. Every blank on the page becomes a month of building in the wrong direction.
Sources
- cbinsights.com/research/report/startup-failure-reasons-top/
- nngroup.com/articles/b2b-trust-from-b2c/
- instantly.ai/cold-email-benchmark-report-2026
- woodpecker.co/blog/cold-email-statistics/
- localiq.com/blog/search-advertising-benchmarks/
- lovable.dev/blog/introducing-growth-tools
- unbounce.com/conversion-benchmark-report/professional-services-conversion-rate/
- cofounderdesk.com/
Cofounder Desk builds your product as real software, puts it live on your own domain and sells it in your name, and its pay comes out of your sales: from USD 2,000 once for the build, then a share of what we sell. If you do not sell, we do not get paid.