Frequency vs. Revenue: A Framework for Ranking Feature Requests
Every product prioritization debate collapses into the same argument. Half the room says "this one is asked constantly." The other half says "yes, but the one asking for this other thing pays us five times more." Both are right. Neither is a decision.
The framework below turns that argument into a two-minute placement exercise.
Why does a single "impact score" fail?
Blending frequency and revenue into one number produces a ranked list that looks scientific and is nearly useless. The reason: the same score can come from very different problems.
A theme with a score of 100 could mean "20 accounts at $50K each," which is a scale problem worth shipping fast. Or it could mean "one account at $1M," which is a custom-work conversation, not a roadmap slot. Same score, opposite decisions.
The teams that produce good roadmaps report frequency and revenue as separate columns, always. The blended score is fine as a sorting default, but nobody makes a decision without seeing the two numbers behind it.
What are the four decision quadrants?
Plot every theme on frequency (x) and ARR at stake (y). You get four regions.
| Quadrant | Frequency | ARR at stake | Decision |
|---|---|---|---|
| Ship this quarter | High | High | Core roadmap. Priority slot |
| Enterprise commit | Low | High | Custom work tied to contract |
| Activation problem | High | Low | Growth or activation squad |
| Defer | Low | Low | Watchlist, revisit quarterly |
The point is not the labels. It is that four different quadrants deserve four different treatments, and treating them all as "roadmap items" is why quarterly plans slip.
When does frequency beat revenue?
Frequency wins the debate in three situations.
- Activation and onboarding gaps. If the same friction shows up in 40% of new-account tickets in the first 30 days, ship it, even if the accounts are small. This is not a feature problem, it is a retention lever.
- Reputational risk. A bug that appears in 20 G2 reviews is a marketing problem before it is a product problem. Frequency signals reputation, and reputation compounds against your inbound.
- Product-led growth motion. If your growth strategy depends on self-serve adoption, the loudest voices in the corpus are your future revenue. Rank on frequency and treat ARR as directional.
The pattern: frequency wins when the theme predicts a future revenue outcome, not the current one.
When does revenue beat frequency?
Revenue wins in the situations that make most CFOs nervous.
- Renewal-gating gaps. A single Fortune 500 raising SSO or SOC 2 concerns in a QBR should outrank 200 mentions from free-tier users. The ARR at stake is measurable, the free-tier mentions are speculative.
- Contract-blocking requests. A prospect deal above 3x your ACV blocked on a specific gap is worth more than 20 low-ACV customers asking for something else, up to the point where you can only build one thing.
- Vertical enterprise motion. If your GTM is enterprise-vertical, deep asks from named accounts are your product. Frequency is a proxy for horizontal usefulness, which you do not want.
The pattern: revenue wins when the theme is directly tied to a specific dollar decision on the table.
How do you handle the ambiguous middle?
Most themes will not land cleanly in a corner. Two thresholds keep the middle honest.
- 5% of unique accounts. Below this, the theme is in the "long tail." Above, it is a system problem worth naming.
- 10% of book ARR. If the theme touches more than 10% of your ARR, it graduates to executive review regardless of frequency.
Themes that sit in the middle on both dimensions belong on a watchlist, not the roadmap. Quarterly, you re-check them. Most stay in the middle. A few move up. Almost none should move down, because if a theme is not decaying but stays flat, it is a stable pattern, not a passing complaint.
What does trend do to the framework?
Trend is a modifier, not a fifth quadrant. Apply it after the two-by-two.
- Growing (+20% QoQ or more). Bump up one quadrant. A "defer" trending up becomes "watchlist." A "watchlist" becomes "candidate."
- Flat (+/- 10%). Keep placement. Flat is not neutral, it is stable, which for a mature theme is a strong signal it is real.
- Shrinking (-20% QoQ or more). Downgrade one quadrant. A theme decaying might be resolving on its own, or your customer base is shifting away from the problem.
The most common trend mistake is treating a growing low-frequency theme as unimportant. If it is up 40% quarter over quarter from a small base, it is on its way to becoming the top theme two quarters from now.
How do you turn placement into a shipping decision?
Placement is not a shipping decision. It is a shipping candidacy. The shipping decision requires two more inputs:
- Cost to build. Rough T-shirt sizing. If the top-quadrant theme is a nine-month build and your quarter is 90 days, it needs a discovery slot, not a build slot.
- Cost of not building. For high-ARR themes, this is quantifiable, and it is often larger than the build cost. A $2M ARR theme that ships in 90 days at a cost of $200K is a 10x investment, and if you would not fund that from other budgets, your framework is not the problem, your capital allocation is.
The prioritization meeting exists to reconcile placement, cost to build, and cost of not building. Not to re-litigate whether the theme is real.
The mistake to avoid
The trap is running the prioritization meeting without the numbers on the wall. When someone says "customers keep asking for X," the right response is not "how many customers?" It is "which quadrant is X in and what is its trend?" If your team cannot answer that in 30 seconds for the top 20 themes, no framework will save the meeting. Frequency and revenue are the two axes. Trend is the modifier. Everything else is opinion, and opinions do not ship a defensible roadmap.
Frequently asked questions
Should we ever ship a feature requested by just one large customer?
Sometimes, but frame it as enterprise custom work with a paid contract commitment, not general roadmap. The rule of thumb: if the one customer's ARR is above 5% of your book and they will sign a contract expansion or renewal contingent on shipping, it belongs on the roadmap. Anything below that is professional services or a partnership, not core product.
How do we handle requests from prospects vs. customers?
Track them separately. Prospect requests are pipeline weight, not ARR weight, and they are more likely to be pitch objections than actual product gaps. A useful rule: a prospect request only qualifies for roadmap weight if it is a deal-blocker on a deal above 2x your average ACV, and the request is corroborated by at least three existing customers.
What frequency threshold is 'high' vs 'low'?
Relative to your corpus size. As a rule, a theme with mentions in more than 5% of unique accounts is high frequency, and a theme with fewer than 1% is low frequency. In absolute terms, for a 300-customer B2B SaaS, that is roughly 15 accounts or more for high, 3 or fewer for low. Use the same threshold across quarters or the trend line breaks.
Does trend matter more than frequency?
It depends on the stage. In a mature product, trend is more predictive because absolute frequencies are stable and change signals real product shift. In an early-stage product with rapid feature launches, frequency is more predictive because trend lines are noisy. Track both, but weight trend higher after year three.
How does this framework change for horizontal vs. vertical SaaS?
Horizontal SaaS should skew toward frequency because most customers share the same core workflow. Vertical SaaS should skew toward revenue because vertical customers pay more for narrow, deep asks that horizontal customers do not need. If you are vertical and ranking primarily on frequency, you are building for the loudest small customer, not the paying enterprise.
Price every roadmap debate in ARR
Palarel clusters every ticket, call, review, and survey into ranked themes with the accounts and revenue behind each one, then files the evidence in your roadmap tool.
Request early access