5 Ways Product Discovery Breaks Down (part 1) - Itamar Gilad

Companies love the idea of evidence-guided product development, but few actually make it work. The problem isn't lack of understanding - it's that leaders create escape hatches. They build two-track systems where some ideas go through discovery while others skip straight to delivery as "must-haves." Given the option, every manager will position their pet feature as table stakes. The discovery track gets starved of resources while the company ships unvalidated ideas using the SCORE method: Sparse data, Consensus, Opinions, Rank, and Ego.

The second killer is weak goals. Most product teams lack clear, measurable outcomes. The company has revenue targets and departmental OKRs, but teams don't know what they're actually supposed to achieve. So they default to following the roadmap. Without clear outcomes, you can't evaluate ideas properly - everything seems important to someone. You can't define success in experiments - any result can be spun as "supporting evidence." And teams optimize for shipping features because that's what they're measured on, not achieving outcomes.

The fix starts with showing business value. Track how many "must-have" features actually contributed anything. Quantify the waste in dev weeks and product bloat. Express it in language managers understand: dollars, efficiency, competitive losses. Then build trust by sharing interim results - how you pivoted based on customer interviews, how you picked the winning idea in a study, how you improved conversions through A/B tests. Keep stakeholders in the loop so they feel control without micromanaging.

For goals, start with a dual metrics tree - your top business metric and your north star metric measuring user value. Build supporting metrics underneath. This gives you measurable outcomes at every level. Express them as proper OKRs that align top-down and bottom-up. The product org usually has to do this work first, then propagate it across the company. But even partial adoption beats output roadmaps and feature factory thinking.