An Opportunity Solution Tree is not a roadmap or a feature list. It is a living map that connects a business outcome to the customer problems that drive it, the solutions you could build, and the experiments you are running to learn what works. Teresa Torres created it and it is still one of the most useful thinking tools in product management.
The structure and why it works
Outcome at the top, opportunities below it, solutions below those, experiments at the bottom. Each layer answers a different question. What are we trying to achieve? What customer problems could we solve to get there? What could we build? How do we know it works? The power is that it separates three thinking modes most PMs conflate: discovery mode (what problems do customers actually have), solution mode (what are different ways to address each), and validation mode (which solution is worth the engineering time). What changed is how fast you can populate, test, and iterate each layer.
Build it in five steps
- Start with your outcome. The root is a specific, measurable business outcome. Not "make users happy." Something like "increase freelancer activation by 50 percent." No AI picks this for you. That is pure strategy and your job.
- Let AI populate the opportunities. If you have the discovery engine running, an agent synthesizes signals across channels. It tells you "42 support tickets this month mention difficulty finding relevant projects, 8 from enterprise accounts." That is an opportunity you did not need 10 interviews to find. But you still judge it: signal strength, segment relevance, business impact, feasibility signal. Pick three to five opportunities, not fifteen.
- Prototype solutions, do not brainstorm them. Instead of listing "recommendations, better search, curated digest" on a whiteboard, build them as prototypes today. Three prototypes, five hours of work, each a testable solution a customer can react to.
- Test with customers this week. Show five customers in your target segment. Watch them use it. Engagement is signal, politeness is noise.
- Update the tree weekly. If it looks the same three weeks running, you are not learning fast enough.
A real tree that changed a quarter
At Smartcat my outcome was to increase freelancer activation by 50 percent. Monday's brief surfaced four patterns. Tuesday I built two prototypes, a simplified profile flow and a match score. Wednesday I showed both to six freelancers; the simplified profile got polite nods, the match score got excitement. One said "if I could see I am an 85 percent match, I would definitely apply. Right now I just guess and give up." Thursday I iterated the match score. Friday I decided: match score moved to validated and ready for engineering, profile simplification parked. Three weeks later it shipped and activation improved 28 percent in the first month. Total time from signal to shipped feature: four weeks. The old model would have been three to four months.
This week, pick one outcome, review your signal sources, build a prototype for the top opportunity, test it with three customers, and update the tree Friday.