Three years ago we stopped writing PRDs. Thirty-page documents nobody reads, outdated by the time they are done. We replaced them with a one-pager. One page. Maybe 1.5 if you count an AI prompt at the bottom. That change made our execution 40% faster, because everyone reads one page instead of skimming the first three of a long doc.
The six sections
Copy this, fill it in, done.
Problem statement. Two or three sentences on what the customer problem is and why it matters. Make someone who has never used your product understand it in 30 seconds.
Evidence. Two or three direct customer quotes with names, plus data. "47 customers surveyed, 60% mentioned this" beats "customers have been asking." Name the competitor if they already have it.
Proposed solution. Two or three sentences on what you are building and how it works, with a link to a prototype. Include an out-of-scope line so engineers can push back on creep later.
Success metrics. One to three measurable outcomes. One primary, one or two secondary, one adoption metric if it is new functionality. "Reduce time switching apps by 80%," not "improve engagement."
Risks and mitigations. Every feature has risk. Pair each one with a concrete step, not "we will be careful." Load test with 10x volume, legal review before launch, feature flag to turn it off.
The ask. What is blocking you. Design sign-off, feasibility assessment, legal review, a specific person's buy-in.
Why one page works
The typical PRD is 20 to 30 pages. Engineers read the first two and ask questions anyway. Designers jump to the wireframes. QA reads nothing. Executives read page one. Meanwhile you spent 15 hours writing it, and a week later the assumptions changed but the PRD did not. Now it is a source of confusion: but it says on page 17.
The one-pager wins for four reasons. It fits on a page, so you cannot hide a bad decision in section 4.2.3. Everyone reads it because it takes five minutes. It is a conversation starter, not a command, so engineers and designers add to it. And it stays relevant mid-project. PRDs are fossils by day five.
Start on your next feature
Do not retrofit the template to old PRDs. Use it going forward. Share it with the team, make "a feature needs a one-pager before engineering starts" the norm, and accept that the adjustment period runs about two sprints while people ask for more detail. Answer those in comments, and do not let the page grow into 20. Write your next brief as one page: problem, evidence, solution, metrics, risks, ask.