Advertisement

Scaling the Analytics Team: Developing Key Roles

By on

In an enterprise analytics team, different roles exist to fill different needs, and those needs must be met in order to be successful. Launching an analytics program doesn’t necessarily require a massive influx of personnel before producing usable insights from data, yet it’s important that critical roles are filled, whatever the size of the team. Multiple options exist for starting small and scaling up an analytics program, according to Evan Terry, VP of Operations at CPrime and co-author of Beginning Relational Data Modeling, in his presentation titled Roles in Enterprise Analytics at the DATAVERSITY® Enterprise Analytics Online Conference.

“Art is ‘I,’ Science is ‘We’” Claude Bernard

Data scientists often explore data independently, but the reality is that an entire support team is necessary for this type of exploration, he said. Data Science operates less like a rock climber and more like a baseball team, where all nine individuals with different specialized roles are on the field at the same time working together, all necessary to compete successfully.

Rather than focusing on specific job descriptions, Terry talked in terms of roles and their functions. By separating the notion of a job description from someone filling a role, multiple options exist for staffing, and in scaling up an analytics team as you grow. That doesn’t mean a different person on the team must be dedicated to each specific role, but roles need to be met across the larger team in order to be successful, he said. 

Exploration is Complicated

Terry drew parallels between an analytics team and the Apollo 11 moon mission 50 years ago. Buzz Aldrin could not have been a successful explorer had he not had a massive team behind him. He needed a lunar lander and a command service module, and all the pieces of equipment that required thousands of people to engineer. So much went into the preparation before exploration could happen, requiring the efforts of a massive support team even before the astronaut reached the moon. “I think it’s equally applicable when you start talking about your analytics environments.”

Three Key Categories

Terry divided roles into three broad categories: leadership, strategy and decision-making; engineering and exploration; and foundational environmental support.

Leadership, strategy, and decision-making entails setting objectives, monitoring performance and measuring value. In terms of the Apollo mission, President Kennedy provided leadership, setting the objective with the decree, “We’re going to go to the moon.” Analytics requires leadership, a strategy, and a decision-making body as well, “in order to make those analytics decisions real and valuable to the organization,” Terry said.

Similar to the detailed engineering that went into the Apollo program, focus is required in an analytics environment to explore and engineer solutions to technical problems. Data needs to be acquired, assessed, and cleansed. Depending on its use, challenges may arise relating to storage, management, and security from within the analytics environment. To define and organize data, standardization processes are needed. Appropriate roles must be in place to understand how to use the data appropriately.

Roles that provide foundational environmental support for exploration are necessary for both moon exploration and analytics. With analytics, those roles are largely technical in nature.

Roles in Modern Analytics

Terry grouped roles within those three categories, and outlined basic responsibilities for each, as well as the questions each set of roles should answer.

Environment and Support

What will we use to get things off the ground?

  • Systems and Database Administrators: Responsible for keeping systems running, managing performance, ensuring the lights stay on and systems continue to compute
    • Activities: Administration of data systems, facilities, storage, networks, and other infrastructure needed for exploration
  • Infrastructure Engineer: Responsible for configuration and preparation of the fundamental computing environment
    • Activities: Builds, configures, establishes, maintains facilities, data storage, networks, and other infrastructure needed for data exploration
  • Support Staff: Responsible for tracking down concerns, issues, and performance problems specifically related to analytics (as opposed to a general help desk)
    • Activities: Brought on as scaling up occurs. Assists other players in capturing, identifying, and resolving problems related to infrastructure, computing environments, data storage and integration, and other technical challenges

Data Exploration and Analysis

What does it mean?

  • Data Scientist: Responsible for research into what insights data has to offer and feeding back into an analytics process further down the line
    • Activities: Curious data whisperer, combination of tech and mathematical skills, modeler, understands distributed computing
  • Analyst: Responsible for analysis and problem solving, either technical or business-focused
    • Activities: Data detective, skilled in statistical analysis, data wrangling

How will we do it?

  • Data Architect: Responsible for organization and integrity of data lakes and other data stores
    • Activities: Data designer, talents in Data Management, design patterns, data organization, and storage systems, QA
  • Data Engineer: Responsible for engineering required to obtain data and configuration of that data so that it is accessible and useful
    • Activities: Data blacksmith, engineering data and systems to run together, data acquisition, storage, ETL, data cleansing

Data Governance and Strategy

What is worth exploring?

  • Stakeholder: Responsible for providing insights and identifying important business metrics. Should have some “skin in the game” regarding the outcome of your analytics activities
    • Activities: Not involved in day-to-day, serves on prime strategy-setting body, determines business objectives and definition of success for analytics program
  • Visionary: Responsible for leadership of analytics team, while maintaining a deep understanding of how analytics can move the organization closer to the goals of the business
    • Activities: Translates strategic objectives determined by stakeholders into an action plan
    • Business Strategist: Responsible for using results from Data Science to change necessary business practices and policies so that they have a significant impact on the business
      • Activities: Makes policy and decisions based on qualitative data analysis and understanding of business strategy

Scaling Up – Sequencing the Addition of Roles

Terry displayed a Venn diagram of roles in four overlapping circles representingStrategy, Data, Analysis, and Technology to illustrate the process of ensuring needs are met as the analytics team scales up. He said it’s crucial to stay balanced in all four areas as roles are assigned, even if that entails having multiple roles filled by a single individual. Based on the skills they demonstrate and their interests, Terry often considers how candidates might be used most effectively. “As you’re scaling up, you’ve got to be able to make sure you can meet the needs of all of these roles, but you don’t necessarily have to have individual people tagged to those,” he said.

Terry said that there are multiple ways to add roles and scale the team, but to build a team from the ground up, start with four roles: an infrastructure engineer, a data engineer, an analyst, and a visionary. These roles can be combined, and initially a single technical person dedicated to your analytics organization could do systems and database administration, Data Architecture, and serve as data engineer, for example.

  • First Phase:
    • Infrastructure Engineer — technology
    • Visionary — combines data and strategy
    • Data Engineer — data, technology, and analysis
    • Analyst — analysis and technology

The second phase involves adding some professionalism around how the environment is controlled and how stakeholders interact.

  • Second Phase:
    • Data Architect — data
    • Systems and Database Administrators — data and technology
    • Stakeholder — strategy

The third phase entails “Enhancing the business through the use of more sophisticated analytical techniques,” he said.

  • Third Phase:
    • Data Scientist — combines strategy, analysis, data and technology
    • Business Strategists — analysis and strategy
    • Support Staff — technology

Growth Is a Balancing Act

Terry said to consider the growth process similar to that of adjusting a graphic equalizer for better sound rather than just checking boxes sequentially. “So you’re saying, ‘I need a little bit more capability in the data space. I need a little bit more capability in the analysis space. Now I’m ready for some strategy,’” moving the levers up at slightly different times and in slightly different ways. If, for example, the technology roles are all filled before the next type of role is filled, the organization will become highly siloed. Trend analysis or higher-level functions such as predictive modeling rely on having the strategy role properly scaled out in relationship to other roles. Without it, “you wind up with an organization that is quite talented and that can produce interesting outputs but no one who is willing to listen,” he said.

Want to learn more about DATAVERSITY’s upcoming events? Check out our current lineup of online and face-to-face conferences here.

Here is the video of the Enterprise Analytics Online Presentation:

Image used under license from Shutterstock.com

Leave a Reply