Photo by Alice Dietrich on Unsplash

A few weeks ago, I was working on performance tuning of the Power BI report for one of my clients. The report page was rendering super slow (15+ seconds). To provide you with a little background: the report uses a live connection to a tabular model hosted in SSAS Tabular 2016.

What if I tell you that I managed to speed up the performance of the report page more than twice, without changing a single line of DAX code behind the calculations?!

Keep reading and you’ll see why very often the devil is in the detail and how thinking outside of the box may help you become a true Power BI champ:)

Image by author

Let me quickly explain the illustration above. There is a Line and clustered column chart visual, where the four lines represent the user choice from the slicers on the left (Reporting threshold and three layers), while the columns are the total sales amount. Data is broken down per year and product. Each of the lines is calculated using DAX (by the way, there is no SELECTEDVALUE function available in SSAS 2016).

I’ve turned on Performance Analyzer, grabbed the DAX query and executed it in the DAX Studio:

Image by author

As you may see, the query takes 13.5 seconds to execute (with cache cleared before run), whereas most of the time was spent within the Formula Engine (76%). This is important, because we’ll compare this result with an improved version of the report page.