Power BI was ranked very well by Garner in its annual Magic Quadrant for Business Intelligence and Analytics report. Most people never read the report because the only thing that really matters to them is the picture above. However, if you do bother to read it you will find a few statements about Power BI that are not factually accurate. This post is not meant to criticize the report but rather to highlight these inaccuracies.
-
“Power BI offers data preparation, data discovery and interactive dashboards via a single design tool” – there are actually two design tools in Power BI: Power BI Desktop and Power BI Service (http://app.powerbi.com). Power BI Desktop is the only Power BI component today that “offers data preparation”. Reports can be built in Power BI Desktop and Power BI Service and Dashboards are only available in Power BI Service
-
“Microsoft is positioned in the Leaders quadrant again this year, with continued strong uptake of Power BI, accelerated customer interest and adoption, and a clear and visionary product roadmap that includes vertical industry content” (highlighting is mine) – I am guessing the vertical industry content refers to this link https://powerbi.microsoft.com/en-us/industries/ which would be somewhat flattering albeit a little perplexing at the same time
-
“customers should be aware that additional data scale-out options incur additional costs when leveraging Microsoft SQL Azure or HDInsight in the cloud — once they reach the 10GB per user limit in the standard Power BI Pro price” – this is wildly inaccurate probably due to misunderstanding of Power BI features. 10Gb limit is for “My Workspace”, however, a user can create a new Group which will get another 10GB allocated to that workspace and when that gets filled up, another group can be created to get another 10Gb… The true limitation of Power BI today is a) single model cannot be more that 1Gb in size in PowerBI.com and b) no more than 10Gb of Power BI datasets can be hosted in any one Workspace (or Group)
-
“absent or weak functionality and an inability to handle required data volumes” – this comment in the Cautions section was probably the most unexpected for me… Having done many bakeoffs against most of the players in this field, handling large dataset with great performance is the easiest challenge to overcome. Analysis Services engine that is bundled with Power BI Desktop is very robust and highly performant compared to most if not all other tools on the market
-
Power BI is “missing basic functionality such as the ability to display data in a pivot table or to create subtotals within a tabular display. Microsoft’s current work-around is to use Excel to create the pivot table” – this is misleading because Excel is NOT a workaround for this glaring hole in Power BI functionality. I am eagerly awaiting the next release of Power BI Desktop praying to see the Pivot Table visual there
-
“While Microsoft offers scale-up options, the path is not clear and is further complicated by differing strategies for Analysis Services on-premises versus in the cloud” – what’s not clear is what’s Gartner is trying to say here. A direct query based model can happily access on premise data while the import based models store their data in the cloud. If “scale-up options” refer to addressing the 1Gb limit for import based models in Power BI, then the answer is very clear as well, you have to move the model to Analysis Services (on premise or Azure), at least for right now.
- “Microsoft’s scores from its reference clients place it in the bottom quartile for breadth of use (as with last year). Breadth of use looks at the percentage of users who use the product for a range of BI styles, including viewing reports, creating personalized dashboards and doing simple ad hoc analysis, to performing complex queries, data preparation and using predictive models. Microsoft Power BI is mainly being used for parameterized reports and dashboards” – this statement also seems nonsensical to me and quite a bit counterintuitive. I would agree with predictive modeling use-case (although you can definitely do predictive modeling using R in Power BI) but all other categories are quite prevalent with all customers that I am working with (I also have to agree that this point is not very factual and is mostly subjective as it is based on my personal experience)
business intelligist 