Forum Discussion
Community ChampionToday is...
Community ChampionAlexisOlson So then you never use stacked visuals?
Super UserStacked column/bar charts? Yes, I definitely use them but not for non-additive measures. Stacked columns/bars aren't really a good choice for non-additive measures.
I tend to have lots of matrix visuals with return measures like ROI or IRR that act more like weighted averages when calculating the total row.
- Greg_Deckler
AlexisOlson The point is, measure totals are "correct" in stacked visuals but not in tables and matrixes. It's mind boggling. Either calculate totals the right way or calculate totals the wrong way everywhere. They are both totals of what is in the visual so...
- AlexisOlson
Whether or not they are "correct" depends on what you need to achieve with the report. Depending on what you need, it's possible that the matrix totals are "correct" and the stacked visuals are not. I think the current situation is definitely unintuitive but perhaps an acceptable compromise since stacked visuals are usually used for additive measures whereas matrix visuals aren't as limited conceptually.
If we force the behavior of the totals to be consistent, then I would strongly prefer stacked visual totals to behave like matrix visuals rather than the other way around. I can always use an iterator to force a total to visually aggregate additively. If matrix totals couldn't properly handle non-additive measures, then I literally could not use Power BI for my work, at least not without huge changes to nearly all of my reports in order to work around that severe limitation.
It doesn't have to be just one or the other though. Instead of breaking all non-additive measure totals, add an option on the matrix visual to calculate subtotals by aggregating its components and everyone can have what they need. Such a feature would be nice to have but is more of an enhancement than fixing something "broken".
- Greg_Deckler
AlexisOlson So go vote for the idea then, because that's exactly what it says, provide options for how totals are returned. I have to believe you are an outlier on this. I literally answered 4 measure totals questions today. Every Excel user expects totals to work the right way. It also basically destroys any concept of self-service visualizations because if you have additive measures, you can never guarantee that they will be correct in every situation.
I simply cannot comprehend the attitude "Oh yeah that makes sense for stacked visuals" and then turn around and somehow it does not make sense for table and matrix visuals when I would bet you that 99% of the people on the planet expect a total row in a table to be the sum of the individual rows.