Agree with everything @Greg and @Heather have said so far, and would add in one additional critical lesson learned. Power BI is relatively new to our organization, and a lot of the managers/clients don’t have much or even any experience with it. Thus, HOW you ask the “what” question is incredibly important. The frame of reference for the vast majority of our clients is Excel, so if you ask what they are looking for in a report, there’s a tendency for them to filter their response through the lens of a) what they are familiar with; and b) what they think is feasible. Thus, their responses tip heavily towards analyses and visuals that mirror Excel Pivot tables.
The analogy I use for this is asking someone from the 1920s how they want their private jet configured? They’ll respond something like this = “well, I don’t really know what that is, but I guess I want it to look like this”:
To avoid this, we now do two things differently:
- we frame the “what” question to the effect of:
“Imagine you had a box that could answer ANY question about your data. What are the X most important questions you would ask? Ignore any thoughts as to whether you think the box could feasibly answer those questions or not - what would you most want to know?”
- deemphasize the “how” question - we typically don’t ask the client about visuals (at least initially). We use our judgment as to the best way to present the answers to 1), and then provide a working draft to them. At that point, they now know what a private jet looks like and the “how” conversation is much more fruitful.