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

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

The previous article in this series examined how to organize an enterprise data analytics program that aligns corporate and departmental teams through standards, a Data Analytics Council, a robust business engagement model (e.g., agile), and communities of practice. This article drills down to examine the teams and roles that comprise a data analytics program.

Teams and Roles. The resources required to support a data analytics program are immense. Many executives think they can become a data-driven organization with a dozen or so corporate resources and others sprinkled throughout the organization. But truly data-driven companies invest significantly in both the people and technology required to harness the full power of data and analytics. 

But truly data-driven companies invest significantly in both the people and technology required to harness the full power of data and analytics.

Eckerson Group’s organizational framework defines 9 teams and 44 roles in a data analytics program, including business unit teams. Some corporate roles might require dozens or more FTEs (e.g., data engineer), pushing the number of full-time equivalent (FTE) employees in a data analytics program into the hundreds. For example, the Travelers has 170 employees on its BI team and another 85 on its data management team, which resides in the IT department. And that’s just the Commercial division of the Travelers. 

Jim Young, senior vice president of business intelligence (BI) at the Travelers says, “[Data and analytics] is a big priority here at the Travelers. We just don’t talk about analytics. It’s a top strategy in the company and executives want to compete on it and they invest accordingly. I couldn’t imagine being anywhere else.”

Every Program is Unique 

Teams. It’s important to recognize that every company organizes its data analytics program differently. Smaller or less mature companies often consolidate multiple data analytics functions into a single team, such as a BI team that handles data management, data governance, and data science. Larger organizations, on the other hand, are more likely to break out functions into separate teams. In addition, some functionality, such as data management and data infrastructure might be handled by the IT department and data innovation and data security might be handled by non-IT groups. This makes it imperative for data analytics leaders to establish strong partnerships with internal groups that manage data. (see figure 1.)

Roles. In addition, not all roles are created equal. Although some roles might have multiple FTEs, others may have none. For example, the BI Marketing role, as important as it is, might be handled by the director or manager of BI, who might hire a freelancer to help write or design marketing materials. In other cases, two or three FTEs may have the same title but play slightly different roles. For instance, a company might have two data architects, but one spends most of her time modeling databases, and hence would fulfill our “Data Modeler” role. Finally, volunteers or part-time staffers may fill some roles, such as the Community of Practice (CoP) Coordinator. 

Figure 1. Eckerson Group Organizational Framework – Teams and Roles

Centers of Excellence. A hallmark of a data analytics program is that it has one or more centers of excellence (or competency centers) staffed by experts in various facets of data and analytics. In general, mature data analytics programs have three centers of excellence: a corporate business intelligence (BI) team, a corporate data science team, and a corporate data management team. These teams build complex data and analytics applications for the enterprise and facilitate self-service by providing coaching and support to data analysts and data scientists in the field.

A hallmark of a data analytics program is that it has one or more centers of excellence (or competency centers) staffed by experts in various facets of data and analytics.

Chief Data Officer. In mature organizations, the teams listed in figure 1 report to a single executive, either a chief data officer (CDO) or chief data and analytics officer (CDAO). Since these functions intersect at some level, a CDO can coordinate activities to eliminate redundancies and optimize resources. However, given the broad scope of responsibility, some organizations spread the functions across several executives. For instance, data innovation is often part of an enterprise product incubator or it is a separate business unit with its own profit-and-loss (P&L) responsibility. MDM teams sometimes sit outside the data analytics program since those teams deliver capabilities used by all applications, not just analytics. 

Data infrastructure. Most organizations have a separate IT organization dedicated to managing data and infrastructure, including data center specialists, database administrators, security specialists, and so on. Given the number of complaints we hear from clients about the disconnect between IT and data analytics teams, our preference is to house data management and data infrastructure functions in the data analytics program. For instance, in our framework, the BI team has a BI administrator; the data science team has a data science administrator; and so on. These people serve as liaisons to the IT department responsible for managing enterprise data and infrastructure. 

Corporate Staffing

Each of the corporate teams pictured in figure 1 is staffed by people serving in one or more roles. Below is a description of each team and its primary roles. At large companies, multiple people might play the same role; for example, a company might have 30 or more data engineers and 20 or more BI analysts. Conversely, at small companies, one person might play multiple roles, such as a data architect that also builds data pipelines, models databases, and builds reports. 

Below is a description of the teams in figure 1 with a list of their key roles. 

BI Team. The BI team builds reports, dashboards, and custom applications for the enterprise and facilitates self-service reporting in the field. Business units may also fund projects with the BI team to build applications that it doesn’t have the skills, systems, or staff capacity to deliver.  The key roles on the BI team are: 

• BI Director. Manages the BI team; maintains close relationships with business units; facilitates self-service reporting; manages development projects.

• BI Manager. Assists the BI director in managing BI teams and customer relationships; serves as agile Scrum master.

• BI Developers. BI tool specialists who build reports, dashboards, and mobile apps. 

• BI Analysts. Business specialists with domain knowledge who consolidate and document requirements; they often conduct analyses, build BI prototypes and serve as agile product managers. 

• Project Manager. Person who schedules and manages projects; if agile methods are used, the person can serve as Scrum master. 

• BI Trainer. Develops and delivers training material (e.g. classroom, online, video)

• BI Help Desk. Provides first- and second-level support (i.e., triage and initial troubleshooting); often part of a larger IT help desk. 

• BI Administrator. Installs, configures, manages, and upgrades BI servers and applications; manages user access and permissions; builds integrations with third party tools and services; sometimes part of IT. 

• BI Marketing. Writes content for newsletters or web site; creates logos and posters; organizes online and in-person events; creates and tabulates surveys; manages internal CRM system for tracking BI team interactions with business customers.  

Data Management Team. The data management team designs, builds, and manages the data environment that extracts, cleans, secures, integrates, transforms, and delivers data for business consumption. The key roles on the data management team are: 

• Data Director. Manages the data management team; maintains relationships with internal IT partners, consultants, and software vendors. 

• Data Architect. Designs the modern data environment, including data processing and storage systems to capture, ingest, refine, secure, and deliver data. 

• Data Modeler. Often called a data architect, a data modeler creates data models and business views that present data in a user-friendly schema. 

• Technical Architect. Also, called a platform or analytics architect, this role designs the data and analytics infrastructure (i.e., APIs, services, containers, sandboxes, servers, and databases) to support self-service activities and data analytics solutions. 

• Data Engineer. Creates data pipelines required to support a use case or customer request; a data engineer that builds data warehouses is called an ETL developer or architect.

• Data Operations (DataOps) Engineer. Builds a continuous testing environment for production workloads; tracks and fixes errors; reengineers development, test, and production environments to speed deployment and improve output quality. 

• Quality Assurance Developer. Builds unit, systems, and integration tests to support data flows and pipelines and DataOps processes. 

• Data Quality Manager. Works with data governance managers and data stewards to track and manage data quality, making changes and fixes as needed. 

• Database Administrator. Installs, configures, manages, and upgrades data management systems; manages database security; tunes database performance; builds integrations with third-party tools and services. 

Data Science Team. The data science team builds analytic models from large volumes of historical data, deploys them for business consumption, monitors their accuracy to prevent data drift; evaluates models for bias and unintended consequences. The key roles on the data science team are: 

• Data Science Director. Manages a team of corporate data scientists; maintains a relationship (formal or informal) with embedded data scientists; evangelizes advanced analytics throughout the organization; offers continuous training and mentorship for data scientists; evaluates the effectiveness of data science projects; trains and mentors data analysts to run data science experiments; creates a culture of analytics. 

• Data Scientists. The statisticians, econometricians, mathematicians, operations researchers, social scientists, and citizens data scientists who work with business to create inferential and analytic models; best if aligned with individual business domains. 

• Data Engineers. These data engineers, aligned with a data science team, help individual data scientists build complex data pipelines that they can’t build themselves. 

• Data Science Analysts. Like BI analysts, these people work with business units to identify data science projects and help data scientists gather requirements. 

• Application Engineers. These engineers help deploy analytic models into production applications, including enterprise reports and real-time customer-facing applications. 

• Data Science Administrator. Installs, troubleshoots, and upgrade data science platforms; works closely with the application, data and DatOps engineers to ensure continuous processing. 

• Community of Practice (CoP) Coordinator. This person coordinates regular CoP events, such as meetups,  webcasts, and hackathons, and contributes ideas for training, certification, and governance policies. 

• Valuations Analyst. This role evaluates the costs/benefits of analytics projects based on initial success metrics, A/B testing, and other evaluation techniques. 

Data Innovation. The data innovation team can take many forms. It can be an internal laboratory where the company designs and prototypes new products, including data-driven products; it can be a formal incubator that attracts and nurtures new product ideas, both inside and outside the company; increasingly it’s a product team that delivers data or analytics as a product or service. Key roles on the data innovation team are: 

• Data Innovation Manager. Runs the data innovation lab, incubator, or product team. 

• Data Designers. Design and prototype new data-driven products and services; often on a 6- to 18-month rotation with internal business teams. 

• Analytic Designers. Design and prototype new analytic products and services; often on a 6- to 18-month rotation with internal business teams.

Data Governance. The data governance team guides the company to identify, prioritize, profile, define, and manage key data elements that must be managed as mission-critical assets, including KPIs, metrics, dimensions, master data, and reference data. Key roles here are: 

• Data Governance Program Manager. Runs the data governance team; facilitates data governance steering committees; delivers data governance process training; recruits and coordinates the activities of data owners and data stewards; coordinates the process of defining, managing, and monitoring key data assets.

• Data Governance Practice Lead. Assists the data governance program manager; skilled at facilitating committee meetings and defining, documenting, and updating data definitions. Provides coaching and support to data owners and data stewards.  

Master Data Management. This team designs, builds, and manages the system that maintains master and reference data for the organization, synchronizing master data across business applications. Key roles in this team are: 

• MDM Manager. Runs the MDM team; works closely with the data governance team to gain consensus among business parties about master data definitions and relationships; creates a strategic roadmap of MDM opportunities. 

• MDM Architect. Designs the MDM system using a hub-and-spoke, distributed, or other model best suited to the organization. 

• MDM Operations Manager. Manges development, test, and production environments and monitors production jobs, fixing errors in a timely fashion.

Security, Risk, and Compliance. This team often resides outside of the data analytics program, especially in large companies with big workforces and complex data processing and application environments. Key roles include: 

• Data Security Director (or Chief Security Officer). Oversees all security at the organization: data, cyber, and physical security. 

• Data Security Analysts. Help the Data Security Officer define and enforce policies, design projects to fix security issues; ensure compliance with all regulations. 

• Data Security Auditors. Identify data security holes and issues; work with data security analysts to minimize exposure and risk. 

Business Unit Staffing 

We just described the roles required to man corporate data analytics teams. But as we said in part two of this series, a data analytics program is federated with both corporate and business unit components. Believe it or not, there may be more data analytic resources in the business units than at corporate.  

But the setup here is more straightforward. Each business unit and department should have one or more data analytics teams comprised of a manager, multiple data analysts, and perhaps a data scientist and data engineer. The number of data analytics teams per department depends on a variety of factors: the size of the department; the data intensity of its work; and how data-driven its executives and managers are. As a rule of thumb, each business unit department should have one data analyst or data scientist for every 25 staff members. 

As a rule of thumb, each business unit department should have one data analyst or data scientist for every 25 staff members.

Data Analytics Team. The key data analytics roles in business units and departments are: 

• Data Analytics Manager. Manages the data analysts and data scientists in their department or workgroup; secures needed computing, data, and analytic resources for data analysts to succeed; works with business to better understand requirements and schedule resources; participates on the Working Committee of the Data Analytics Council that governs the enterprise data analytics program. 

• Data Analyst. The proverbial spreadsheet jockey with a master’s in business administration; works in finance, marketing, sales, and operations to answer ad hoc questions, managing pricing, build KPIs and performance metrics, develops plans and options. 

• Data Scientist. Helps optimize business processes using analytic modeling techniques. 

• Data Engineer. If there are not enough corporate resources and it has sufficient budget, the department can hire a data engineer to support its data analysts and scientists, making them more productive.  

Data Governance Team. Business units are responsible for managing the integrity of core data assets in their assigned domain. Core roles here are:

• Data Owner. Accountable for the definition, integrity, accuracy, and consistency of data elements assigned to it; works with data stewards who handle day-to-day data governance responsibilities. 

• Data Steward. Responsible for the definition, integrity, accuracy, and consistency of data elements assigned to it; informs data owner of any issues, including data quality issues and desired updates; works with data quality managers (on the corporate team) to track the quality of data elements in a data quality dashboard. 

Walk the Talk

It should be clear by now that data-driven organizations invest a significant amount of resources in people and teams. There are no shortcuts to data nirvana. Executives can’t talk about the importance of data and then allocate only a dozen or so resources to data analytics at corporate and elsewhere. In data analytics, you reap what you sow. To succeed, executives need to invest heavily in people, platforms, and processes.