Opportunity Sizing
Estimating how much a problem is worth before committing to solving it.
The point of sizing isn't a precise number. It's to catch the ideas that can't possibly be worth it before they eat a quarter.
A rough number that changes a decision
Opportunity sizing is estimating the value at stake in a problem — how many users it touches, how often, and what each instance is worth — enough to decide whether it deserves real investment. It's deliberately rough; the goal is the right order of magnitude, not two decimal places.
A useful size is one that would change a decision. "This affects maybe 2% of users who rarely convert" and "this blocks half of new signups on their first day" both point to the same feature area, but they lead to completely different priorities.
Core insight
Where sizing goes wrong
The common failure is a top-down number so large it's meaningless — "the market is $40B, so 1% is $400M." It sounds like rigor but assumes the hard part (getting that 1%) away. Bottom-up sizing, built from your actual users and conversion, is slower to fake and far more honest.
The other failure is false precision: a model with eleven assumptions, each a guess, multiplied together. Every added assumption widens the error. A defensible size states its two or three load-bearing assumptions plainly so others can push on them.
Size bottom-up, expose your key assumptions, and stop at the precision that changes the decision. Anything more exact is theater.