Learnings from largescale Power BI deployment for a Fortune 1K Company
Microsoft Power BI is a modern, industry-leading, cloud-native BI platform and one of most cost-effective ones out there in the market.
Ever more enterprises are adopting Power BI. The adoption scenarios vary from Power BI co-existing with other BI tools that might be in use already, but lately the more commonplace scenario is Power BI altogether replacing existing solutions such as MicroStrategy, Qlik, SAP Business Objects, Tableau etc. It’s becoming common for enterprises to migrate their reports from other BI platforms to Power BI.
Here are few learnings we have had as we migrated a large enterprise from their existing BI tools/solutions to Power BI:
Source Data Remodeling
The various reports have their data sources and getting source data in the right format for a particular BI solution is a non-trivial task. So, in migrations to Power BI, the enterprises look for least disruption and tend to put Power BI on top of the data sources that had been created for the existing BI solution. However, by design the various BI solutions define the data in their own formats. For example, many BI solutions place all the data together in one big table, sometimes even with data at different granularities being in the same table. Power BI’s source data is modelled as a star schema. So, if enterprises don’t model the source data as a star schema, they’ll run into issues such as incorrect values in the reports and poor performance.
Importance of Semantic Layer
Even if you have a star schema, you also have to understand that how Power BI works with data may be different to the way your previous BI tools worked. Power BI wants to you invest time up-front to create a semantic layer and once you have done this, you’ll find that building reports and calculations is much quicker and easier. If you’re wondering why you can’t just write a SQL query to get the data for that chart you need to build, you’ve made this mistake.
Recreating exact functionality as your old BI platform
If you’re building a Power BI report to replace an existing report on a legacy platform, and you ask your users what they want the report to look like, the most common reply is “Just like the old one”. This is a warning sign and a surefire way to make the Power BI report hard to build and slow to run is to try to make it do something it wasn’t designed to do. An example of this is the way some users ask for reports that recreate the huge tables they are used to seeing in their old Excel reports. These tables may be slow to render in Power BI, plus they don’t make the data easy to understand. Best approach to avoid this issue is to understand the business problem the report is trying to solve and implement a solution using the functionality Power BI has built in.
Analyze as in Excel
If your users want to be able to explore their data by changing the measures and fields used in a visual, they are probably thinking of how they use PivotTables and PivotCharts in Excel. Power BI’s Analyze in Excel feature doesn’t give you something that looks like an Excel PivotTable, it gives you an actual PivotTable connected to data stored in Power BI. It’s the functionality your users are comfortable with none of the downsides of traditional Excel reports such as the tedious, error-prone, manual data refreshes. You can use Excel cube functions for more complex report layouts, such as those needed by financial reports.
If you want to create reports sourced from hand-written SQL queries with big tables that can be printed easily, you’ll find that Power BI Paginated Reports work much better than regular Power BI reports. Closely related to SQL Server Reporting Services, paginated reports have a different set of strengths compared to regular Power BI reports.
Making the right changes
Migrating to Power BI is a chance to make a break from the old ways of doing things, and that includes who does what. Typically, two approaches to BI are seen in enterprises:
1. Corp BI, where all the work is done by the IT department that quickly becomes the bottleneck, unresponsive to business needs and unable to understand them properly.
2. Excel Hell, where the business builds everything itself and ends up with a lot of duplicated effort, multiple versions of the truth, and reports that break when the person who built them leaves for a new job.
Power BI enables you to chart an optimal course between these two that fits the needs and culture of an enterprise. It’s easy to use and affordable enough for you to deploy it to a larger number of users in your organization than any other BI tool except Excel. At the same time, it gives you the tools you need to avoid the problems of Excel with centralized data, automated refresh, security, monitoring and a lot more.
If you’re thinking about migrating from your existing BI tools & solutions to Power BI and embracing Microsoft Azure in the process, it can seem like a big step. With the expertise, support, and insights of Dynapt.ai’s team of certified Power BI and Azure experts you don’t have to go it alone.