Register for "Diverse, Integrated, and Real Time: Delivering the Right Data for AI/ML Success" - Tuesday, June 18,1:00 pm EST

Organizing for Success, Part I: How to Organize a BI Team in the Age of Self Service

business intelligence team management

The goal of self-service analytics is to empower business people to build their own reports, dashboards, and predictive models. If that happens, does your company still need a corporate business intelligence (BI) team? 

Although some business leaders, burned by BI backlogs and high costs, may gladly jettison their corporate BI teams, most would not. That’s because they are still highly dependent on the corporate BI team, as grossly underfunded and undermanned as most are. In fact, self-service analytics doesn’t work well if the BI team is not a strong and coequal partner with the business. For self-service to succeed, companies need to invest more, not less, in their BI teams. 

Self-service analytics doesn’t work well if the BI team is not a strong and coequal partner with the business.

 For self-service to succeed, companies need to invest more, not less, in their BI teams.

The Role of the BI Team

To be clear, by BI team, I mean a corporate group that primarily designs and builds data and analytics applications, governs and manages data resources, and builds and manages the data environment and underlying infrastructure. (Some companies now call these data analytics teams.) In some organizations, however, the BI team has a lighter footprint: it serves primarily as an intermediary between business units and an information technology (IT) group whose mission is primarily to gather and prioritize requirements. In either case, self-service analytics fundamentally transforms the role of the BI team. 

Facilitate Not Dictate. The primary shift in roles is that the BI team no longer designs and builds everything. As I mentioned in a 2017 video blog, the goal of modern BI teams is to “facilitate, not dictate.” That is, a BI team’s primary job today is to nurture the data and analytics capabilities in each department so it can service its own analytical needs. 

Community of Practice. One way it can do this is by fostering a community of practice (CoP) among distributed data analysts and data scientists. A CoP can take many forms: it can consist of regular meetups or lunch-and-learn sessions; monthly webcasts where practitioners share best practices; case studies in email newsletters; certification programs and continuous education; and annual contests where data analysts engage in friendly competitions (i.e., vizathons or hackathons) to demonstrate and sharpen their skills. 

First Line of Support. But a more important way that BI teams can facilitate self-service is to serve  as the first line of support for embedded data analysts who build reports and dashboards. Data analysts are the prototypical spreadsheet jockeys (now Tableau or PowerBI jockeys) with strong knowledge of a specific business domain (e.g., finance, marketing, sales). When they face technical challenges, they should be able to turn to data and analytics experts with questions. Consequently, the BI team should have data, BI, and data science experts on hand who can provide guidance and assistance to embedded analysts when needed. At some companies, BI and data experts post regular “office hours” to support data analysts. 

Data Infrastructure. More importantly, the BI team is responsible for building and managing the data infrastructure that undergirds all self-service activities. This means creating a “data refinery” that collects, integrates, and manages data assets that data analysts can query. It might also mean creating subject-specific data marts to support each department or specific sub-domains. It also means curating a data catalog that data analysts can search and annotate when looking for data assets and maintaining a collaborative data preparation environment where users can share and reuse transformation code when building data sets for analysis. 

Self-Service Reality 

The reality of self-service analytics, however, is that few departments have a full set of data analysts—typically, just finance, marketing, and sales—who can build reports and dashboards. Most departments, including operations, human resources, and the executive team,  still require the BI team to build all their reports and dashboards. So, besides designing and managing the data infrastructure, the BI team still must excel at data and application development. 

Even departments with lots of data analysts need the BI team to build certain types of applications for them. That’s because data analysts and even data scientists don’t get paid to design, test, build, and manage complex, high-performing, multi-user applications, nor do most have the skills or time to do so. 

In addition, who is going to design and build enterprisewide applications, such as corporate scorecards and enterprise reports? And who is going to build and manage the data infrastructure that supports these applications as well as all self-service activities? Since no department funds these types of activities and infrastructure, that responsibility falls to the corporate BI team. 

So, even in an organization with a flourishing self-service analytics program, the corporate BI team still spends a lot of time building custom applications. That means it needs to have a full roster of data and analytics experts on its team, both to support self-service analytics and build custom analytic applications. To support these activities, the BI team also owns or shares responsibility for myriad  other tasks, including data architecture, data governance, data infrastructure, and data innovation. 

Conclusion—A Strong BI Team Fosters Self Service

Thus, a strong BI team accelerates self-service: it builds complex departmental and enterprise applications that the business can’t; it creates a data refinery that delivers quality data the business can trust; it fosters a culture of analytics through communities of practice, office hours, and gamification; and it improves data analyst productivity by managing centralized data catalogs and data preparation tools that capture tribal knowledge and foster greater reuse and standardization. 

The next article in this series will discuss the composition of a modern BI team that facilitates self-service, spearheads the development of data-driven applications, and builds, manages, and governs the enterprise data environment. Future articles in the series will discuss business engagement models, the data science organization, and the BI Council or oversight board.

Read - Organizing for Success, Part II: How to Organize a Data Analytics Program

Wayne Eckerson

Wayne Eckerson is an internationally recognized thought leader in the business intelligence and analytics field. He is a sought-after consultant and noted speaker who thinks critically, writes clearly and presents...

More About Wayne Eckerson