In the Power BI Desktop August Feature Summary (https://powerbi.microsoft.com/en-us/blog/power-bi-desktop-august-2017-feature-summary/) Microsoft announced the “live connect” Power BI Service connector is generally available. Now everyone can start using this feature to connect to any of their already built Datasets in their Power BI service. This makes it possible to efficiently create a high-quality Dataset model once and reuse it in many reports with ease.
So what problems does this new feature solve, or what opportunities does it create? Three immediately come to mind:
- One model developer, many report authors.
- The skill set required to build a Power BI data model, understanding relationships and directionality among tables, building lookup tables, using complex DAX to build new measures, implementing Row Level Security (RLS), etc. is quite different from creating visualizations that are meaningful to a specific set of users who can use that information to make better decisions, take different action and create value for their organization. It may be that only 1 or 2 individuals are tasked with “model development” responsibility, and likely live in IT.
- Report authors can be individuals with sufficient technical skills to build visualizations, but just as importantly be domain experts in Finance, HR, IT, Operations, etc. with a deep understanding of what questions the decision makers in their domain need answered. It may be that many individuals can be tasked with “report authoring” responsibility, and likely live inside the teams and departments who need to take different and smarter action.
- Information drawn from a single, controlled and managed Dataset, integrating siloed data created by multiple systems, for consistent use across teams, departments and processes.
- As changes are made to the shared Dataset, they are available to all your report authors who use that Dataset. Did you recently define a new KPI for your entire organization? Build it once in the shared Dataset and use it in multiple reports in consistent ways by various report authors.
- Do you already have multiple reports built by various authors using different data sources and data models, wondering which calculation for “Gross Profit” or “Net New Customers” are the right ones for everyone to be using? Or why a great KPI in one report is not available in another? With a single shared Dataset, everyone sees the same thing.
- The ability to quickly and inexpensively create a “lightweight data mart” that can prove value before consideration of further investment for integration into or development of a formal enterprise data warehouse infrastructure.
- Power BI is great for rapid development of dashboards and reports that answer questions through multiple refinements and iterations. But not all answers are created equal. It may be that some answers are worthy of promotion to an enterprise data management infrastructure, while others are not, given the upfront and ongoing costs to do so.
- If you make the decision to invest in your enterprise data warehouse based on actual value captured in Power BI vs anticipated value when deployed to your enterprise data warehouse, you’ll spend less using Power BI to learn the difference and invest big only when you have a proof to do so.
Additionally, as described in the August Feature Summary, you can switch from one Dataset to another while authoring your report in Power BI Desktop. This is a potentially useful feature if you want to use a Dataset in a staging or development Workspace when building your reports, then switch your reports to point to the production Workspace Dataset right before publishing and sharing. You can switch out what Dataset you are connecting to using the Data source settings option under Edit Queries in the Home ribbon.
So what does it look like for your report authors to create reports when connected to a Power BI Service Dataset?
Open Power BI Desktop
Get Data and choose the “Power BI service”
When you connect, if you are not already logged into to your Power BI service, you will be prompted to do so:
Then select the Power BI Dataset that you want to use to build new reports. (You must have access to that Dataset in the Power BI service, as if you had logged in to powerbi.com )
Now notice something different in the left-hand navigation inside Power BI Desktop…
You see only the Report View icon; normally you’d see the Data View and Relationships View icons, too.
And although you cannot create a “New Column” or “New Table”, you can still create a “New Measure”
Even though a report author won’t be able to change the underlying data model (i.e. tables, columns, relationships, measures) themselves with a connection to a Power BI Service Dataset, they *can* still create custom measures in their reports that will be of interest to their audiences, expressing complex performance ratios or calculations that are specific to users of their reports and dashboards. Over time, some of those measures might be worthy of incorporation into the shared Dataset for all to use, if that made sense to do for your organization.
If you missed this update, or haven’t yet considered it for your own organization, I’d recommend you take a day and explore the potential value.