Problem strength
How painful, frequent and valuable is the problem? A clever solution is not enough if the customer has little reason to change behaviour.
Evidence of demand
Look for behaviour, not compliments. Interviews, waitlists, manual pilots, deposits, paid trials and repeat usage provide stronger evidence than people saying they like the idea.
Commercial logic
Estimate who pays, how much they may pay, what it costs to acquire and serve them, and how long the business can fund learning before it needs stronger traction.
Delivery feasibility
Identify technical, operational, regulatory and data dependencies before committing to a large build. The right first test is often much simpler than the final product.
Keep the goal simple: spend enough to learn what matters, then use that evidence to decide what deserves the next investment.