Most discovery sessions ask the wrong question
READ TIME - 4 minutes ⏳
Email formatting looks broken? click here to view in browser.
Every report starts with a discovery session which usually lasts an hour, sometimes less, where you sit with whoever asked for the report and try to work out what to build. It is the cheapest, highest-leverage hour in the whole project, and most of it gets spent on the wrong questions.
Here is what usually happens, and I'm sure this will sound familiar to you...The room asks WHAT: what charts do you want, what KPIs, what filters, what should be on the first page. If it is a good session, it might get to HOW. The How should it look, how often does the data refresh, how will people access it. Then time runs out, everyone feels productive, and the building begins.
The trouble is that WHAT and HOW are the two questions that add the least value, and WHY, the one that adds the most, is the one almost nobody asks.
Three questions, very different returns
Picture a discovery session as building a stack of value, where each kind of question adds a different amount to it.
WHAT sits at the top and is the most visible layer while also being the thinnest. A list of charts and metrics feels like progress, but it is really the user or client's guess at a solution, handed to you as if it were a requirement. Build exactly what they asked for and you have built someone's first idea, not their (refined) second thought.
HOW sits under it, a little more substantial. Here we can talk about refresh frequency, layout, access and all of it matters for a report that works in practice. But HOW is still about the shape of the thing, not its purpose.
WHY is the whole base of the stack, and it is where nearly all the value lives. Why do you need this report. Why now. What decision are you trying to make that you cannot make today. What happens if the number comes back bad, and what happens if it comes back good. Who acts on it, and what do they do next.
Those questions feel slower because they look and feel less like a brief and more like a conversation: there are no pre-defined guidelines and "one-template-rule-them-all" type of guide. That is exactly why they get skipped, and exactly why they are worth protecting time for in the early phase.
What changes when you ask it
When you understand the WHY, something quietly useful happens: the WHAT and the HOW start to answer themselves.
If the real purpose of a report is to catch a production line drifting out of tolerance before the end of a shift, you no longer need to ask which charts they want. The decision tells you. It has to be glanceable, it needs a clear threshold, it needs to refresh fast enough to act on, and it does not need the twelve other metrics someone listed just in case.
The WHY did the filtering for you.
Skip the WHY and you get the opposite. You build a faithful, well-made report that answers a question nobody actually had, and a few months later it sits unused while its owner exports the table to Excel and works it out themselves. That is not a failure of skill, it is a discovery session that stopped one layer too high.
The most expensive shortcut
Rushing past WHY feels efficient because you end up with a list, you can start building, everyone can see movement. But it is the most expensive shortcut in the whole project, because every wrong assumption you make in that first hour gets encoded into weeks of work, and the cost of unpicking it only grows.
The fix costs "nothing", in the sense that it isn't a new tool or a longer process. It is spending the first part of your discovery session refusing to talk about charts, and your favorite Power BI feature and staying on why this report needs to exist until the answer is clear enough that the rest becomes obvious.
Same session, same hour, and a completely different project comes out the other end.
So before your next kickoff, one thing to sit with. In your last discovery session, did you find out what they wanted, or why they needed it? And if it was only the first one, how would you ever know whether you built the right thing at all?
Let me know, I read all feedback :) !
See you in two weeks,
Julien
