Product and engineering · Answered by Golden Section from more than 400 B2B software companies observed
Rank work by what the plurality of your best-fit customers need to reach the outcome they bought, not by who asked loudest. Give reliability its own swim lane so it is never traded away for a feature, and hold a quarterly roadmap meeting (monthly while early) where sales, support, engineering and the executive team argue their case and the product manager versions the result. Customer requests are an essential input, but one customer's request is a data point until several core customers share it. A feature drives revenue when it raises the monetized return for a segment you sell to, or removes a reason you lose deals in that segment. Start by tagging every open request with the customer, their segment and the outcome it serves, then cut whatever serves only one account.
The roadmap belongs to the product manager and the decisions are made in a standing meeting, not in the hallway. Build what serves the plurality of core customers, protect reliability with its own lane, and treat every single-customer request as unproven until the segment confirms it.
| Metric | Value | What it means | Source |
|---|---|---|---|
| Roadmap meeting cadence | at least quarterly; monthly at an early stage | the meeting where competing priorities are argued and the roadmap is versionedthe product manager owns the file and the version after each meeting | Golden Section playbookProduct Management Process |
| Roadmap resolution | no more than 3 sprint cycles per unit | time span covered by one column of the roadmapthree swim lanes and quarterly columns is, in our words, too loose for an investment plan | Golden Section playbookProduct Management Process |
| Engineering and product allocation in the AI era | about 45% core, 45% AI-native features, 10% skunkworks | share of engineering and product resourcesGolden Section's suggested split for incumbent software companies, March 2026 | Golden Section, publishedThe Enhancement Doctrine |
| Customer ROI that signals underpricing | above 10x | monetized customer savings less price, divided by pricepast this point another feature rarely raises revenue; the price does | Golden Section operating viewGolden Section operating view |
A roadmap is the place every part of the company fights for its view of the product. Sales pushes for the feature that closes this quarter's deal, support and operations push for reliability, and the executive team pushes for competitive position. The product management play describes the two ditches on either side. When the squeaky wheel wins, engineering works on urgent rather than important things and the product stays in catch-up. When engineers win, the product fills with interesting technology customers did not ask for.
The filter that resolves most of it is the one in The Balanced Path: does this serve a true problem for the plurality of the customers you have chosen to serve? That requires knowing which customers those are, which is the work of customer segmentation, and knowing what each problem is worth to them, which is the work of the customer ROI play. A request that improves monetized return for a core segment earns a place. A request from one large account outside the core is usually a services engagement or a no. Writing the reason down when you say no is what stops the same request returning every quarter.
The figures are invented. A company carries 46 open feature requests going into its quarterly roadmap meeting. Tagging them shows 19 came from two large customers outside its core segment, while 11 are variants of the same reporting problem raised by a dozen core customers. Reliability has no lane, and production defects have been rising for two quarters. The product manager proposes the reporting work and a reliability lane with fixed capacity, defers everything requested by a single account with a written reason, and offers the two large customers a paid services option. Sales objects that one of them is renewing. The founder holds the decision and asks sales to bring the renewal risk to next month's meeting with numbers.
A contractual commitment already signed has to be honored, and a regulatory change in the vertical can jump the queue because every customer needs it. Before the company has chosen its segments, roadmap prioritization is really segmentation work in disguise.
From the Golden Section mistakes list, each paired with the play that prevents it.
Sales controlling the roadmap is the most common way prioritization collapses into whoever closed the last deal.
A low-resolution roadmap cannot show the trade-offs, so every argument is won on volume instead of evidence.
Promising roadmap dates to customers freezes one version of a document that has to keep changing.
Saying yes to every request broadens the product until it is hard to sell to anyone.
In the order we would run them. Each is on its own page, most with a free Excel template.
Names the two or three customer cohorts whose problems get priority.
Puts a dollar value on each problem so features can be ranked by what they are worth to customers.
Sets the roadmap meeting, the swim lanes and the owner who versions the result.
Keeps the roadmap themed, current and published company-wide.
Shows whether what you shipped is used, which is the evidence for the next round of priorities.
Development plays Product management, roadmap and engineering process plays that turn priorities into delivered work.
Most of it should trace to customer problems, but not to individual requests. Weight a request by how many core customers share the underlying problem, not by the size or volume of the customer asking. Reliability and the work that protects the core need their own capacity, or customer requests will consume all of it.
Decline anything that serves one account outside your core segments, anything you cannot tie to a monetized outcome, and anything that broadens the product toward a market you have not chosen. Record the reason. A product that tries to serve every request becomes too broad to sell.
Ask what it does to customer ROI in a segment you sell to, and whether lost deals in that segment cite its absence. If the feature only matters to one prospect, it is a sales tactic. And if customer ROI is already above 10x, the revenue is more likely to come from price than from another feature.
No. Sales is an essential input and should argue its case in the roadmap meeting with deal evidence, but the product manager owns the roadmap and the founder arbitrates disagreements. A roadmap run by sales produces early wins and an unreliable product that the installed base comes to resent.
Reviewed by Dougal Cameron, CEO & Co-Founder on 2026-10-22. Golden Section observations are labeled separately from external benchmarks and illustrative arithmetic.