Self Service BI is here to stay

Just when Microsoft announced the upcoming composite Power BI model some friends of mine were wondering out loud: "what's the deal, will they focus on power to end users or on IT enterprise BI?". Just to be clear: the upcoming feature Microsoft announced is the ability to extend Power BI and Azure Analysis Services Models with custom sources such as Excel files. Read about it here: https://powerbi.microsoft.com/en-us/blog/driving-a-data-culture-from-insight-to-action/

My friends were worried that this would lead to more "Excel hell" situations. I on the other hand am excited. Sure; an Azure Analysis Services data model is often on top of a integrated data warehouse or data mart. Meaning that the integration should have been done before the data model. To just add an Excel file to a nicely integrated data model can go against everything modern data architects stand for. The reason for that is that it opens the door for more data silo's. Think about it: you have a central data model in Analysis Services and 10 reports on top of that. If each of those reports starts to add excel files you would end up with 10 different data models, creating 10 different inconsistent solutions.

But, this is what defines self service BI: POWER to the end users! That is why it is called Power BI I guess. The business is in charge. Sure, BI consultants have the job to protect the end user for creating a messy data landscape with tons of incompatible and different solutions. "What is a customer" might be a question that has 100 answers, based on 100 different reports with different data models. And that is perfectly fine. The more power the end user gets, the more they will understand data and the underlying problems and the need for consistency.

Do not get me wrong, I do prefer to integrate the data before it ends up in a data model, and data silo's are not a good thing. But we do not need to protect end users by limiting their options.

It is like self service in the real world: you can hire a plumber, or you can go to the shop and buy all the tools to do it yourself. Experience will lead you to an insight on when to do it yourself, and when to hire a professional. With data it is just the same. Self service is here to stay, and that is a good thing!

Principal BI consultant at Rubicon

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Next ArticleStore streaming data to Azure Synapse Pool (Part 3)