Up until recently, I didn't. Infamous Silicon Valley founders & investors that I admire are fanatical about data. Fully submerged in the chaos of young startups, it was always impossible for me to prioritize. I remember thinking, "just wait until we finish this next feature, then we'll be more stable and ready to start collecting data!" I was wrong. Data is a prerequisite to stable growth.
Sometimes data can precede product too. Angel investor John Danner wrote a thread about using lead-gen data to test the market before wasting engineering time on a product that people might not want. He explains that software companies can typically A/B test landing pages with different messaging & value propositions in order to find the most powerful combination before building any product at all.
I say typically because some startups (games for instance), need a minimal product in order to measure engagement with their customers. It's easy to convince yourself that this feature will finally make users love your product. Deciding if a feature is MVP worthy is simple: if you don't need it to measure core user behavior, don't add it.
Regardless of what MVP means for your specific business model, it's important you get there as fast as possible. If MVP is when you begin to leverage data to learn about your customers, pre-MVP is a sort of assumption matrix that leaves you completely disconnected with market perception of your product. It's no wonder the best founders move so quickly.
The secret sauce to moving fast isn't just hard work, it's non-attachment. The biggest reason founders don't stick to the data-driven, MVP philosophy is because of a fixation on specific ideas or methods. When you write code & hire employees around assumptions instead of hard data, you make it hard to pivot when customers point you in a different direction.
Once you've found product-market fit (PMF), the consequences of not studying customer usage data & collecting qualitative feedback only increase. At this stage, your team should optimize it's efforts around the weakest points in your customer journey. For example, if data shows that only 10% of users who sign up to your app take the next step, your efforts should be focused around optimizing that process until conversion rates improve.
By letting data guide product development, you can be certain resources are being spent fighting the biggest fires that have the greatest impact on growth.