Product Discovery & Decision-Making

Real discovery starts with frequent, deep conversations with customers before any code is written. Yaniv Yaakubovich (Verve PM) uses Zume Pizza as the counter-example: a startup that raised hundreds of millions of dollars while its product team barely spoke to customers and built according to the CEO's emails. The result was burnt pizza and a product that didn't meet a real need. Without discovery, he says, product teams become "technicians".

Amit Slutzky of Trullion adds a layer: customers don't always say what's really going on, and without trust you only hear what you want to hear. He describes a 7–8 hour session with a New York customer where a slow conversation and lunch exposed a widespread product problem hidden behind workarounds. Fixing it raised adoption by tens of percent.

Roee Froman of Taboola recommends starting from the customer's pain rather than the solution, and asking customers about their alternatives, not just their wants. Amit also suggests bringing engineers and designers into customer conversations early, so they understand the problem first-hand instead of just receiving a spec.

Most product failures don't come from picking the wrong problem but from never reaching its root cause. Eyal Pinko, a VP at Taboola, estimates that 90–95% of failures stem from a weak definition of what causes the problem. His tool is the Five Whys, a method that originated at Toyota: you keep asking "why", roughly five times, until you move from the symptom to the root.

For it to work, Pinko says, you first need to understand and define the problem with data, run the process with a cross-functional group, ask open questions and rely on facts rather than guesses. You stop when you find a systemic problem that will recur, when you get the "eureka" feeling, or when you hit a dead end.

Shuki Aharonovich of Augury uses an Opportunity Tree: a living document that starts from a metric and breaks it into problems and opportunities down to atomic problems, and doubles as an alignment tool with stakeholders who jump to solutions. Roee Froman counts breaking a complex problem into small, digestible pieces among the most important skills a product person has.

Good product decisions combine data with conversations, and when there's no data you need a strong opinion you're willing to change. Chani Lin of Ledge works by Strong Opinions, Loosely Held: a strong view anchored in vision and experience, people around you who poke holes in it, and a commitment to change course the moment there's a signal, to avoid analysis paralysis. Yaniv Yaakubovich notes that a startup without traffic has no statistical significance, so you look for trends and back them with qualitative testing and customer conversations.

When there's plenty of data, the question is how to pull a real signal out of it. Itai Nof of Wix starts from the quantitative, segments and spots signals, and only then talks to users to understand the sentiment behind the numbers. A feature that looked like a failure in testing became a winner after two weeks of daily data digging exposed a small segment where a voucher wasn't showing up. Betty Antebi describes how at Wix, with over 230 million users, an A/B test tells you within two weeks whether a move moved the needle.

Prioritization is mostly the ability to say no. Ariel Ornstein of Ionix warns against inflated scope and last-minute proposals: if something truly moves the needle, it probably didn't pop up at the last minute. Roee Froman uses a prioritization spreadsheet but doesn't let it decide: the weight of each criterion is a business decision that exposes what you're betting on.