I think the author of this article makes a compelling argument for why building a recommendation engine is not an efficient strategy in creating an MVP. I've been working on an e-commerce startup for over a year now (eep!) and after spending countless hours arguing over this exact prioritization with my co-founder, I had to finally call it quits, because we just never got anywhere. Recommendation engines sound sexier when pitching ideas, but at the end of the day, how do you pitch a product that doesn't exist to customers that only exist in theory?