How to get started with Power BI in your organisation - 5 recent lessons
We’ve done a couple of Power BI projects recently and have some insights about implementation that are worth sharing. Both client organisations are in the education and training world but the lessons could apply to organisations in other sectors.
- Start by identifying real business problems that could be solved with better reporting. Focus on some quick wins and then it’s easier to sell the solution to colleagues. Leave the more advanced reports that uncover new insights until later. Management reports that take a lot of time to produce every month are good candidate. Use them to inform the design of the data warehouse. Share the reports within the Power BI online service and set them to update automatically.
- Invest in a central reporting database or data warehouse. It may cost more upfront but over time you’ll see a big return on your investment. A central reporting database gives you a ‘single source of truth’ and it can be designed to make report authoring much easier. Create a set of denormalised views in the reporting database that can be used as the data sources for your Power BI reports. They will definitely save authoring time and they can simplify issues around data security and report sharing - you create different views for different user groups.
- Consider building the reporting database in Microsoft SQL Server in the Azure cloud. Scheduling data source refreshes in the Power BI online service from SQL Server in Azure is very straightforward - you keep all your users on the latest version of the data. MS SQL Server in Azure also means no need to purchase reporting server hardware or install SQL Server locally.
Develop some simple reports in Power BI Desktop. Author the reports in the desktop application and then upload them to the online service. Our experience is that the simplest workflow is:
- Develop reports in Power BI Desktop.
- Upload to the Power BI Service.
- Schedule the update frequency - all reports are updated directly from your SQL Server db in Azure.
- Create a dashboard in the Power BI online service as a ‘front end’ for your report. You can then share that dashboard and the underlying report with colleagues.
- Revisit data security. Have you got distinct Power BI user groups/audiences within your organisation and are you happy that they all see reports built on the same underlying data source? One way around Power BI’s fairly open data sharing model is to create different database views for different user groups - restrict access to confidential data by using the right view as the datasource for your report.
Power BI gives you the tools you need to author reports and then share them across your organisation. With the right approach to implementation it has the potential to transform the way you find and share hidden value in the data you collect about your customers and the work you do with them.