[shalmanese]

> A lot of times people want to reduce product management to "visit customers, ask what they want, make a spreadsheet", but I think it's more than that. To do it well, you have to know the possibilities the existing stack could grow into, and also what it can't easily grow into, and set a vision that is well beyond what the masses point at.

Customers tell you what their problems are but they're not trained in telling you the solution. The mistake that's often made is that customers tell you their problem in the form of a solution. The role of the PM is to then take the solution, reverse engineer the problem out of it and then work with the team to figure out the true solution.

Where the technical/non-technical split lies between the PM and the engineers can vary. Ideally, the PM is technical enough to come up with new possibilities on their own and understand the technical implications and tradeoffs of each possibility but not every PM has the background or inclination towards that. The next best thing is that the PM works closely with the engineers to communicate the user goals and trusts engineering to find clever solutions to those goals. But the worst is a PM who thinks they're technical enough to come up with features and then steamrolls over engineering objections over the implementation.