You need to understand exactly what problem you are trying to solve. In that sense, analysis is really important. But analysis for months and attempting to come up with a full set of requirements and a complete design of the final product, will be detrimental to the project. Making a minimum viable product and iterating over it is a lot more important. However, understanding the context and making informed decisions about the architecture is important. But the view about the entire project needs to be really high level. Abstract thinking. No details. Focus on details should be only when a specific problem is actually solved, or it will eat so much time, you end up never starting to do anything, but instead thinking about doing the job.
Detailed analysis is an obstacle for progress. And I say this from experience. Sadly, the way I work is waterfall, and all project needs to be defined, so it can be estimated, budgeted, outsourced... The trouble is that after 6 months, we are just finishing the analysis. If we would have started with developing features we knew we would need and iterate on that, we would have been almost finished by this time. However, in order to go for an agile approach, there has to be a culture for that in the company. If politics prevent this from happening, analysis paralysis will inevitably follow.