You can look at data and use inductive reasoning to formulate a hypothesis, or use deductive reasoning to define your product goals. Note that you must have a hypothesis or goal to test and measure; otherwise, you're performing the equivalent of asking random sea turtles which shoes to make for the world's fastest Olympic runner. Even then, you're eliding a whole bunch of implicit assumptions about what market you're in (shoes or athleticwear, apparently) and who has valuable insight into that market (presumably ocean-going Cheloniidae). You can certainly sally forth and do that with minimal forethought, but it's not really a good use of anyone's time (including the turtles') and is unlikely to yield useful or measurable results.
Capture Assumptions, Form Hypothesis, Validate...Rinse, Lather, Repeat
You Need a Hypothesis
For effective research of any kind, you always have assumptions (consciously or not), and generally need to formulate a hypothesis to be tested. Even if all you're doing is conducting a survey or interviewing people, you still need to have some idea of what you're trying to validate (or invalidate, if you're investigating the null hypothesis). Otherwise, your research will lack direction and rigor, and floundering around aimlessly is unlikely to provide meaningful results.
Develop an MVP
The level of research and rigor required will vary greatly. One popular lean approach is to formulate a hypothesis to be tested using a Minimum Viable Product (MVP). The MVP should be designed to test your hypothesis by creating validated learning about the marketplace, potential customers, or fitness for purpose.
The Process Never Stops
An MVP or other testable hypothesis is the start of your product journey, not the end. You will have to continuously seek feedback and refine the product throughout its entire lifecycle. Markets change, business needs change, and technology changes, which is why you're not still using a Windows 3.1 desktop. Even in non-agile organizations, effective product management requires an iterative or incremental approach to product design and market analysis, so get comfortable with the continuous generation (or at least refinement) of hypotheses.