Datamart a new self-service developer tool in Power BI

No-code database development

A data mart is a small to midsize relational database usually dedicated to a single subject in which users can access and explore data related to that subject to build reports and dashboards and gain insights.

Typically, however, constructing data marts requires knowledge of coding languages, and they are built by developers in IT departments. In addition, any changes to the data mart that business users might want or need -- for example, ingesting a new source of data -- have to be made by those dedicated developers in centralized IT departments.

With the introduction of Datamart in Power BI, however, Microsoft is attempting to remove the burden of development from IT departments by giving business users the capabilities to build their own relational databases.

With Datamart, Power BI users will be able to extract, transform and load data from any data source without needing to write code to create a relational database up to 0.5 terabytes in size.

Subsequently -- also without requiring code -- they will be able to build semantic models to categorize data and discover relationships between data points, manage and govern the data within the data set to keep it secure, and make the data available for reporting and embedding.

Once generally available, Datamart will be one of the most significant additions to Power BI in years, according to Arun Ulag, corporate vice president of the Intelligence Platform at Microsoft.

"We have millions of developers taking advantage of [Power BI] and building amazing analytics solutions, but today, we're making a huge upgrade," Ulag said during a session at Build. "This is probably the biggest new capability we've delivered in Power BI in the last three years."

An analyst's view

While self-service data mart development is new to Power BI, Microsoft is not the first vendor to offer such capabilities, according to Doug Henschen, an analyst at Constellation Research.

For example, cloud data vendor Snowflake already provides self-service data warehousing.

This is probably the biggest new capability we've delivered in Power BI in the last three years.

"Snowflake is the poster-child example that has demonstrated the appeal and demand for self-service warehousing and mart-like, single-node instances," Henschen said.

In addition, BI vendors have enabled users to curate data sets for specific applications with built-in semantic modeling and governance capabilities to provide oversight, he continued.

Nevertheless, Henschen called the addition of Datamart to Power BI significant.

"It's a significant announcement as it brings self-service the next level down from BI, promising self-service data marts," he said. "The upside is giving analysts and data-savvy power users the no-code flexibility to build their own marts so they no longer have to wait for IT to build a new mart or adapt the existing warehouse to include new sources."

But there is a possible downside, Henschen cautioned.

As with any self-service capability, there's the potential for data to get isolated within a small group of users' workflows and for different users working to build data marts with the same data to come up with different versions of the truth. Therefore, governance capabilities will be critical, according to Henschen.

"What has yet to be seen with this capability is how well the touted 'integrated governance' capabilities will provide a bit of top-down control over how data is defined, accessed and used," he said. "As the saying goes, 'With great freedom comes great responsibility.' Self-service mart and warehouse capabilities could be hugely beneficial, but the guardrails must be effective and must be adopted and used."

Real-world example

One of the Power BI customers using Datamart during its beta testing period was Bayer, a German multinational pharmaceutical and life sciences company with about 100,000 employees around the world.

Using Datamart, the company built a financial dashboard to help evaluate and manage risks associated with COVID-19 to its business and employees throughout the ASEAN region, a group of 10 countries in Southeast Asia that includes Indonesia, the Philippines, Singapore and Thailand.

Bayer needed to ensure business continuity while protecting its employees throughout the region. To do so, it consolidated data sets from different sources within a relational database developed with Datamart and subsequently built the dashboard that has enabled the company to act and react quickly during changing economic and public health conditions.

Manually building a similar data mart would have taken significant time, which has not been a luxury during the pandemic, and Datamart enabled Bayer to quickly consolidate key data, join related data points and build a data asset that has been the basis for key decisions.

"Having to manually pull all this information together and reconciling between sides, countries and the region would have been very time-consuming, and when you're dealing with a changing COVID environment you don't have the value of time," Ernst Coppens, president and CFO of the ASEAN region for Bayer, said during Build.

So to save time, Bayer consulted with Microsoft and used Datamart in Power BI to develop its COVID-19 data mart and financial dashboard.

"Our COVID data mart provides real transparency with real-time data for crisis management meetings as well as ad hoc analysis requests," said Matthias Eisenack, portfolio manager of financial performance at Bayer. "Getting started with Datamart in Power BI was easy and quick … and we have all relevant information at hand to interact live with data and base our decisions on real-time analysis."

Eisenack added that Bayer plans to extend its use of Datamart to build similar relational databases for other applications.

More capabilities

In addition to the introduction of Datamart in preview, Microsoft made generally available Metrics -- previously named Goals when it was first revealed in May 2021 -- for Power BI and the preview of data storytelling in PowerPoint.

Metrics is essentially a performance tracker that enables Power BI users to create key business indicators and business objectives, assemble them in a single location on top of Power BI, and use data curated in Power BI to measure progress against those business metrics and business objectives.

Using the tool, customers can create performance scorecards, share progress updates and take proactive measures when key metrics aren't being met. In addition, customers can assign hierarchies to scorecards.

"I was fascinated by the name change from Goals to Metrics to 'better align with how our customers are using it,'" Henschen said, quoting a blog post from Microsoft. "This suggests to me that the BI-centric users of Power BI have a monitoring mentality, whereas 'goals' suggests the setting of business objectives and targets -- something these users may not have the authority to do."

The name change also suggests a less ambitious product roadmap for the tool, reducing expectations that Microsoft will add scenario planning and predictive modeling capabilities, Henschen said.

Data storytelling in PowerPoint, meanwhile, is an integration between Power BI and PowerPoint that will enable users to embed live, interactive Power BI reports directly in PowerPoint presentations. Currently, users often take static screenshots of reports to pair with their presentations, according to Microsoft.

And though referred to as data storytelling in PowerPoint, the integration is not a data storytelling capability in the sense that it automatically generates narratives about data. Instead, it is the integration of an analytics platform with a platform for developing narrative presentations.

Meanwhile, by making Power BI available in PowerPoint, Henschen said Microsoft is hoping to display its analytics platform to a new set of potential users.

"By integrating Power BI all over the Microsoft 365 portfolio, Microsoft is hoping to seed even more Power BI demand," he said. "Rivals can integrate with Microsoft 365, too, so it's on Microsoft to do it first and do it best."