I've built RICE spreadsheets. I've sat in rooms where a roadmap decision came down to whose reach estimate looked more credible in a shared doc. And I've watched the same pattern every time: the framework doesn't settle the argument, it just moves the argument into the inputs. Someone decides "impact" is a 3 instead of a 2, the total score flips, and the meeting ends feeling resolved when nothing about the actual disagreement (should we go deeper on one market or wider across three) was ever discussed out loud.
What scoring models are actually good for
To be fair to RICE and its cousins: they're excellent at ranking ideas that already sit inside the same strategy. If every option on the table is "a way to improve activation for our existing ICP," scoring them against each other is a reasonable, fast way to sequence a backlog. The failure mode is using the same spreadsheet to compare options that don't share a strategy at all, a retention feature, a new-market bet, and an infrastructure migration, as if "impact × confidence ÷ effort" means the same thing across all three. It doesn't. You're not comparing effort anymore, you're comparing world-views, and the math hides that.
The decision a scorecard couldn't have made for me
When we scoped Ovela, the tempting move (the one a scoring model would have rewarded) was breadth: list every specialty a medical-travel patient might want, maximize theoretical reach. Instead we picked one: dental. Lower "reach" on paper. The strategic bet wasn't about the market size of dental care, it was about what one specialty, done with real trust-building (JCI-accredited partners, a concierge who actually answers), could prove that ten specialties done thinly couldn't. No scorecard produces "go narrower to go deeper" as an output. That's a strategic stance, argued for, not calculated.
A scoring model answers "which of these ideas is biggest." It can't answer "which kind of company are we trying to be," and that second question is the one that actually decides the roadmap.
What I use instead
Not nothing. Structure still matters, you just put it earlier, before scoring:
- Name the strategic bet first, in one sentence, before any option is scored against another. For Ovela: "win on trust, not price, by going deep on one specialty."
- Sort options into "serves the bet" and "doesn't," before ranking. Most scoring debates are actually two different bets arguing through the same spreadsheet.
- Only score within a bucket. Comparing effort and impact among ideas that already serve the same strategy is legitimate; comparing across buckets with one number is where the theatre starts.
- Write down what the bet predicts, so six months later you're arguing about results, not re-litigating the original scores.
The honest objection to this
"Just argue about strategy" sounds like it replaces rigor with vibes, and in a room without a strong facilitator, it can. The fix isn't to retreat to the spreadsheet for safety. It's to write the bet down in one falsifiable sentence and hold the team to defending it out loud. A bad strategic argument, made explicitly, is easier to kill than a bad strategic argument smuggled in through someone's "confidence" score.