Register for "A Guide to Data Products: Everything You Need to Understand, Plan, and Implement" - Friday, May 31,1:00 p.m. ET

A Framework for Blending Generalists and Specialists in a Federated Model

Eckerson Group advocates a federated center of excellence because it blends the best of both centralized and decentralized organizational models without the downsides of either. One reason the federated model works is because it stitches together technical generalists and specialists.

T-Shaped People. These days, management recruiters, consultancies, and agile development teams actively seek “T-shaped” people. As a symbol, the letter “T” has a vertical line that represents depth or concentration and a horizontal line that represents “breadth” or versatility. Thus, an ideal executive, consultant, or scrum team member has both deep skills in one area plus experience in other areas. In popular parlance, an I-shaped person is a specialist, while a T-shaped person is the “jack of all trades and master of (at least) one”— someone who is both broad and deep, skilled and versatile.

T-Shaped Teams. Of course, most people aren’t versatile in either experience, training, or temperament. Consequently, most organizations need to figure out ways to blend two or more people together to create T-shaped teams. In the world of data analytics, this means creating a federated organizational model that blends technical generalists in departments and technical specialists at corporate. A federated model is key to succeeding with self-service analytics.

Federated Models 

Phase 1: Silos. In many companies, technical generalists (i.e., data analysts) reside in departments and technical specialists (i.e., data engineers, data architects, software developers, data scientists, data administrators, and so on) reside on the corporate data analytics team. They occupy separate orbits and work on separate projects. As Rudyard Kipling wrote, “Oh East is East and West is West, and never the twain shall meet.” (See figure 1.)

Figure 1 – Phases of Integration

In this type of siloed environment, the organization gets minimal benefits. Technical generalists deliver a multiplicity of ad hoc projects but rarely build the infrastructure needed to address 80% of departmental questions. In addition, they don’t have the skills, time, or resources to build a cross-functional application that delivers game-changing business insights.

Phase 2: Shared Service. In the next phase, department heads solicit the services of the corporate data analytics (CDA) team to deliver a project. The CDA head creates a project team from a pool of available technical specialists. The project team builds the solution, but interacts little with the embedded data analysts except to gather requirements and perform user testing. Although technical specialists on the CDA team can build complex applications, they lack the domain knowledge of technical generalists. Consequently, their applications take longer, cost more, and frequently fail to meet user expectations.

Phase 3: SWAT Teams. In the next phase, the CDA configures its specialists as multi-disciplinary agile or SWAT teams. Each specialist on the team is responsible for covering the work of other team members if they are out or absent. That means a data engineer needs to learn how to build data models, and BI developer needs to learn how to build data pipelines, and so on. Over time, these teams become more versatile and agile and individual developers gain more confidence and skills, giving them more career options. In addition, agile teams incorporate a “product owner”—usually a data analyst who participates on the team and prioritizes stories, approves output, and recommends process improvements. Here, the corporate and departmental teams fuse together to create a capital I-shaped team—a multi-disciplinary team linked to a domain specialist, delivering huge benefits.

Phase 4: Spanner. Multidisciplinary teams are a great leap forward, but they are not as effective as a multi-disciplinary individual. An individual who learns all the skills of a multi-disciplinary, including domain knowledge from the data analyst, becomes a “spanner”- someone who “spans the stack” of skills that typically are apportioned to data analysts and all technical specialists. A single spanner can build better applications, faster, and at lower cost than a multi-disciplinary team.

Few individuals are capable of being spanners, nor want to be. And some applications are simply too large or too complex for a single person to build. But for a majority of solutions, a spanner can be an effective method for delivering “faster, better, cheaper” solutions. It takes longer for an individual to build the skills of a multi-disciplinary team, but once they do, they significantly outdistance the performance of those teams. Many of the world’s preeminent musicians, athletes, and inventors are spanners—people who started slow, sampling a range of activities and genres, and eventually pulled away from the pack because they could make unique and innovative connections where others could not.

Conclusion – From T to I-Shaped Teams

A federated model stitches together corporate and departmental resources as well as technical generalists and specialists. It then grafts deep domain knowledge onto a multi-disciplinary team to turbo-charge the delivery of data analytics solutions.

So, rather than search for T-shaped people or teams, organizations should create a capital I-shaped teams that blend multi-disciplinary teams of technical specialists with technical generalists (i.e., data analysts) embedded in the business.

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