| As someone else who interviews a lot of PM candidates, I have to admit that my process is quite different. - Understand the product, target audience, competition I don’t really care if they understand the product during the interview process. The purpose of the interview, in my opinion, is to see if they have good product intuition, a well reasoned problem solving framework, strategic insight, leadership qualities, mental horsepower, etc. If these things exist, they can learn about my product, target audience, and competition. In fact, I actively prefer to not talk about the products I work on because it leads to a biased and loaded conversation. It’s a lot harder to impress me when it comes to something I think about every single day, especially when you’ve only had 3 days to think about it. However, if we both talk about Medium, for example, the playing field is leveled and the candidate will feel more comfortable. - Metrics Metrics absolutely matter. But if you’re letting them lead the conversation based on projects they’ve worked on and just talking about metrics they measured, your results will obviously not be calibrated. I much prefer to ask a calibrated question that allows me to compare candidates to each other. Ideal questions walk through defining metrics that would dictate success for a certain product, outlining how to run experiments, and then walking through a fictitious metric drop to get an understanding of how they would decompose the problem. - Don’t bullshit Agreed. But I would prefer “I don’t know, <insert follow up questions to get to an answer>” when possible. - Prioritization Agreed, they should have a good framework for prioritization. I like to couple this with a product intuition question where we’re riffing off of each other to come up with solutions to a problem. - Engineering Divide Sure. I think it depends on the role, and if this is ideal for you then that’s reasonable. - Brainteasers Brainteasers are easy to game and don’t mean anything. Google famously published that brainteasers are not effective at predicting outcomes. You can test their ability to make assumptions, tackle a problem, etc without giving brainteasers. |