We’ve recently talked about layering, and how you can use Tooltips and Drillthrough pages within your Power BI reports to make your content easier to explore for your customer base. These are not the only layers; however, you have to be prepared to consider and understand the difference between Power BI Dashboards and Reports within the Power BI Service. None of this is helped by the simple fact that Dashboards and Shared datasets are not immediately clear when you start within Power BI Desktop.
The major benefits that you get from using the Power BI Service only come to the fore when you stop – as a colleague so eloquently put it – “Being a short-order cook” and start to think about the bigger picture. Centralising and standardising your data is a game-changer in so many ways. What I see so often is people who think you can implement a Data Strategy. The truth is it cannot. Start simple, capture the business rules that apply across the various data sets already in use. Go to the source of current reports, reverse-engineer them if necessary to determine exactly what is being done if the documentation/schemas are missing (or more likely if the intern who set them up 15 years ago is no longer contactable). Only when you know what is currently being reported upon and how can you map any changes. However, you may still want to wait. Starting simple with Power BI and bringing in a basic spreadsheet as a source for a first report is not bad; a daily spreadsheet can become a rich dataset when placed in a folder, and all the copies you have are read together. That can then get the necessary buy-in to move to a direct database connection, now knowing all the business rules and having got people used to the idea of using Power BI content.
People’s biggest mistake is thinking, “If we invest heavily in a data platform that will deliver success”, that is often the case in the short term. Still, unless you can embed the skills, roles and responsibilities in managing it consistently, it will falter as soon as minor tweaks are requested. Like all development processes, it is important to remember that “Adding is easy, but removal is nearly impossible”, so look to start with basic data and add to it and its complexity, rather than trying to pull all business data in with no plans to be able to use it.