RICE vs ICE vs MoSCoW for solo founders: pick a score, keep the evidence
Published August 10, 2026 · Hub: Product prioritization hub
You pick a framework. You fill a sheet. Reach is a guess, Confidence is optimism, and by Friday the ranking matches whoever argued hardest. That is not a RICE problem. That is an evidence problem wearing a formula.
Solo founders and small-team PMs do not need a better calculator. They need scores that stay attached to what customers actually said—and a decision you can hand to Cursor without re-litigating the numbers.
Deep dive on the formula: RICE prioritization framework. Weekly process around scoring: how to prioritize features. Hub: product prioritization.
Why framework debates usually fail
Common failure modes:
- Scoring random feature requests before any captures exist
- Inventing Reach or Impact so a favorite idea wins
- Using MoSCoW as a ranking system (it is not)
- Switching frameworks every quarter instead of fixing evidence hygiene
- Running RICE in a spreadsheet while the quotes live in Slack, Intercom, and your head
Frameworks structure conversation. They do not create judgment. If inputs are vibes, outputs are vibes with decimals—and coding agents will happily ship the top row.
Quick definitions (then pick with evidence)
RICE
Popularized at Intercom:
(Reach × Impact × Confidence) ÷ Effort
Best when you have many comparable candidates and at least rough reach estimates tied to real funnel or segment data.
ICE
Impact, Confidence, and Ease (often 1–10). Best for fast triage when reach data is thin. Weaker when you need comparable numeric ranking across very different bets.
MoSCoW
Must / Should / Could / Won’t (this cycle). Best for stakeholder alignment and scope negotiation. It does not tell you which Must beats which Must.
RICE vs ICE vs MoSCoW comparison
| RICE | ICE | MoSCoW | |
|---|---|---|---|
| Primary job | Numeric ranking | Fast triage | Stakeholder categories |
| Needs reach data? | Yes (even rough) | No | No |
| Precision | Higher (if honest) | Coarse | None (labels only) |
| Common failure | Fake Reach/Confidence | Inflated Impact | Everything becomes Must |
| Best for solo / small teams | Shortlist with metrics | Early triage | Scope conversations |
Decision tree: Need agreement on what is in scope this cycle? → MoSCoW (or now/next/later). Comparing a few similar bets with metrics? → RICE. Thin data, need speed? → ICE. Confidence mostly opinion? → Do not score yet—capture and theme first.
How you run the frameworks inside Caret
By the time you open a score sheet, the work should already live in product memory—not in a blank row titled “CSV export.”
In Caret, the loop is:
- Captures — feedback, call notes, churn reasons, and research land as project memory with segment and source.
- Insights — themes form from those captures so you score jobs, not orphan tickets.
- Opportunities — promote a shortlist against the named constraint (activation, retention, monetization, reliability, differentiation).
- Score with evidence attached — when you set Reach, Impact, or Confidence, the factor points at the captures and themes that justify it. Ask against project context if you need “what evidence do we have for Impact on activation?”
- Decision brief — the winner becomes a what/why/evidence artifact you hand to a coding agent—not a lonely number in a cell.
The framework is a lens on opportunities you already earned—not a substitute for the trail. See how Caret works and AI product brain for why session chat cannot replace that trail.
What each factor should point at
| Factor | Invented (bad) | Evidence-backed (good) |
|---|---|---|
| Reach | “Lots of users” | Trials/accounts hitting the step in a named window, from captures or metrics |
| Impact | Preference | Severity + constraint link across ICP themes |
| Confidence | Optimism | Pattern strength: channels, segments, recency |
| Effort / Ease | Dream scope | Smallest useful version of the opportunity |
If you cannot open the captures behind Confidence, you are not ready for RICE—you are ready for more intake or discovery.
Worked example: same three bets, scored from memory
Constraint: Activation—trials never reach first insight.
In Caret (captures already in the project):
- Six support notes + two churn quotes on empty-state abandon
- One enterprise asking twice for CSV export
- Twelve community mentions of dark mode
Insights: “Empty-state / first-run dead end,” “Finance handoff / export,” “UI preference (dark mode).”
MoSCoW pass (scope): Dark mode → Won’t this cycle. Export and onboarding skip → Should. Must is empty until something blocks shipping entirely.
Promote to opportunities: Skip optional onboarding integrations; CSV export slice.
ICE (fast triage in Caret): Onboarding skip scores higher—Impact on activation is clearer from the theme, Ease is better, Confidence is backed by multi-channel captures.
RICE (if funnel numbers exist): Onboarding skip with ~500 trials/month at the step beats niche export with tiny Reach—even though one account asked twice.
Decision brief (handed to Cursor): Job = first successful insight in session one; evidence = capture links; scope = skip optional integrations + guided first action; non-goals = export suite, dark mode; success = trial-to-first-insight in 14 days.
The framework did not invent the answer. Product memory did. The score made the trade-off inspectable. Full brief shape: product decision briefs.
Mistakes that recreate scoring theatre
- Scoring before captures exist
- Confidence at 90% with zero linked quotes
- Effort estimated on the full vision, not the smallest useful bet
- Re-scoring every week without new evidence—just new opinions
- Pasting the winning feature name into Cursor without the brief
Where the score still drifts
You can keep RICE in a sheet for a while. Then Reach gets renegotiated, Confidence floats up, and nobody can open the quotes that justified last month’s ranking.
Caret keeps the captures and themes next to each factor so you score from evidence you can reopen—not from whoever argued hardest in the room.
Start a free 7-day trial—no credit card—and attach this week’s feedback before you score again. Or see what Caret is if you want the product overview first.
FAQ
What is the difference between RICE and ICE?
RICE scores Reach, Impact, Confidence, and Effort. ICE scores Impact, Confidence, and Ease. Solo founders usually start with ICE or constraint gates, then RICE a shortlist once captures justify the numbers.
When should I use MoSCoW instead of RICE?
Use MoSCoW (or now/next/later) to decide what is in scope this cycle. Use RICE only to rank a few comparable bets that already have evidence in product memory.
How does Caret help with RICE or ICE scoring?
Caret stores the captures and themes behind each factor so Confidence and Impact are inspectable—not invented in a meeting. Scores stay attached to the decision brief you hand to shipping.