Great summary of how to apply lean ideas to data-heavy products. Just one addition that is necessary to actually answer the question you’re asking in the topic. What actually makes an MVP is split up into two questions:
- How must a product be used so that we can call it adopted?
- What’s the minimum adoption rate within the target group so that we can call it relevant?
You’re indirectly answering those questions in your recommendation example stating that enough customers already buy products when only being suggested the most popular ones in a first step for the product. So this simple solution already generates so much value that it can be shipped (at least to certain customers) and function as a first baseline.
It’s important to mention that whether or not a solution actually is an MVP, is not decided by the product team. The target group decides on that by adopting a solution in a high enough rate or not.