Chapter 3
Best Practice #2
Build the High Performing Analytics Team

“Talent wins games, but teamwork and intelligence win championships.”

Michael Jordan

Once the goal statement for the analytics initiative is defined, refined, and validated, implementing best practice # 1, the goal needs to be realized by the analytics team. The second-best analytics practice is on building high performing analytics teams. Whether the company is a big corporation with thousands of employees or a small company with just twenty employees, high performing teams inevitably offer superior business results. “No matter how brilliant your mind or strategy is, if you are playing a solo game, you will always lose out to a team,” is the way Reid Hoffman, LinkedIn co-founder, sums it up. Successful analytics initiatives are no exception and are also dependent on high performing teams to provide good business insights to the insight consumers.

Why is this a best practice?

While most companies understand the importance of analytics, according to a recent McKinsey survey, fewer than 20% have maximized the potential and achieved analytics at scale [Miranda, 2018]. There are many reasons for this dismal statistic, and one important reason is the organizational design or team structure for data analytics. This is important in today’s business environment as every company is a data company that strives to have an operating model that focuses on innovation, scale, and value creation. Achieving this requires a set-up with new skills, new roles, and new organizational structures.

Building the right data analytics team starts at the highest level in the company. At a strategic or board level, there is still a lack of awareness on the potential of digital technologies on business performance. McKinsey’s research says just 16% of the board members fully understand how the industry dynamics of their companies were changing due to digital technologies [Mckinsey, 2018]. At the operational level, most enterprise data analytics teams today are a shadow of the old MIS (Management information system) or BI (Business Intelligence) team structure and typically reporting into the CFO (Chief Financial Officer) function. These “CFO-centric” teams are organized around the specific IT skills that are often a combination of ETL (Extract-Transform-Loading) developers who build and maintain data-marts and data-warehouses, business analysts who capture the needs of business users for operational and BI reports and report builders who run queries and build reports. In addition, most of the current data analytics teams report to the CFO, who is usually averse to innovation and change as it is a cost controlling and regulatory function. This makes it difficult for the analytics team to deviate from this mind-set.

Realizing the best practice

There is no one standard way to build a high performing analytics team. Building a strong data analytics team varies from one organization to another and is usually contextual. Some of the characteristics of effective teams like communication, trust, a clear sense of purpose, and mutual support, are applicable to data analytics teams as well. But there are some characteristics that are specific to analytics teams. Top-performing analytics organizations are enabled by deep functional expertise, strategic partnerships, and a clear center of gravity for organizing analytics talent [Miranda, 2018]. These are the strategic elements. But what are the tactical aspects of building a high performing analytics team? Below are five key tactical elements or capabilities that are required in realizing this best practice of building high performing data analytics teams.

  1. Data literacy as the foundation
  2. A strong analytics leader
  3. Staffing the team across the entire data lifecycle
  4. Hypothesis-based methodology
  5. Execution mechanism for data analytics

This brings us back to the earlier question on where the data analytics team should be positioned in the organization? Should the data analytics team report to the CFO, or should it report to a different function? Fundamentally, data analytics is a value creation function; it is not a controlling or a regulatory function. In other words, the data analytics team and the leader should come from the “create” or customer-facing business functions like Sales, Marketing, Innovation, Operations, or even better – the data team headed by the Chief Data Officer (CDO).

A 2018 KPMG study found businesses that have a CDO are twice as likely to have a clear digital strategy [KPMG, 2018]. According to IBM, two-thirds of the firms that have a CDO outperform rivals in market share and data-driven innovation [IBM, 2016]. Even during the COVID-19 pandemic crisis, the Center for Disease Control (CDC) was looking at recruiting a CDO, and this highlights the importance of the CDO role [Vincent, 2020]. Mario Faria of Gartner Research Board (GRB) says, “Most CDOs care about solutions and how they can impact revenue” [Torres, 2019].

1. Data Literacy as the foundation

Along with having the CDO, the success of the enterprise data analytics team rests heavily on establishing a culture of data literacy in the entire organization. Data literacy is essential to position the analytics team for success. Data literacy in the analytics context can be achieved by creating an environment where the organization strives to use insights from data over intuition to augment their decision-making process.

Senior management support is essential for achieving data literacy in the company. But how does one convince the senior management if they lack an analytics background? The C-suite is not always thinking of data analytics as they assume that analytics is covered and managed by the operational team. Tactically, you build awareness on data literacy to the senior management by highlighting the business opportunities lost and the compliance risk that exists due to a lack of data analytics products and solutions in the organization. Business practice #6, discussed in chapter 7, looks at building data literacy in the company, and chapter 9 looks at building data monetization capabilities in the company.

2. A strong analytics leader

Once the culture of data literacy initiative is in place, a strong leader is required on the ground to run the data analytics initiatives. The analytics team should be led by one who has a solid understanding of data, technology, and the business to translate the vision and the needs of the business stakeholders into measurable results. In most organizations today, this analytics leader is the CDO – Chief Data Officer. Just like the CFO manages money, the CMO (Chief Marketing Officer) manages products, the CDO should manage one of the key business assets – data.

According to Florian Zettelmeyer and Matthias Bolling of Kellogg School of Management, “Getting value from data is not a technical challenge. It is a leadership challenge that demands developing and deploying data strategy throughout the organization” [Zettelmeyer and Bolling, 2015]. According to them, the analytics leader role demands excellence in three strikingly different areas:

  1. Strategic orientation: The analytics leader must be able to find new opportunities to add value, not simply oversee analytics operations.
  2. Change leadership: This will involve developing processes to break data silos, drive data-driven projects across those functions, and link analytics initiatives to operations.
  3. Collaboration and influencing: The analytics leader must cultivate a compelling vision, earn buy-in from key stakeholders to weave analytics into the fabric of the organization.

Mckinsey consulting calls the analytics leader as a catalyst—who embraces a style of leadership addressing the current demands and roadblocks and deploys analytics solutions at scale. The analytics leader should have the skills and experience not only to build the culture of data literacy but also to educate and drive implementation of insights throughout the organization.

3. Staffing the team across the entire data lifecycle

Traditional enterprise data analytics teams focus on technical capabilities like ETL (Extract-Transform-Loading data) and report building, while the best analytics practice is to build a multi-disciplinary analytics team across the entire data life cycle (DLC). DLC is the sequence of stages a data element goes through from its initial generation to its eventual archival and/or purging at the end of its useful life. From an analytics perspective, there are four sequential stages in DLC: Data Capture, Data Integration, Data Science, and Data Visualization.

The four stages of the DLC and the key IT products in each of these DLC stages are shown below.

Value stewardship over showmanship.

With this backdrop, the analytics team should be led by the Product Manager given that the delivery of data-analytics solutions often demands strategic planning, capital investment, and management of complex development cycles, including ongoing maintenance. The emphasis is on product management over project management as analytics product managers focus on scale and reusability, while project managers usually focus on the instantiation of the analytics products in a time-bound manner. An analytics product manager holds the key if the organization intends to be a long-term player in leveraging data analytics for business performance.

This requires the analytics team should constitute team members who bring expertise on data capture, data engineering, data science, and data visualization and should be supported by the data governance team. While the data capture and the data engineering teams are IT teams, the data science and data visualization teams should be from the data team. Regardless, the team members from both these data and IT functions should integrate seamlessly and work collaboratively with business stakeholders.

4. Hypothesis-based methodology

The next phase is to devise a methodology specifically for delivering analytics solutions. Analytics solutions cannot be often delivered in a big-bang approach. It must be delivered iteratively and incrementally as the business is an evolving entity, continuously adapting itself to be relevant in the market. In this regard, a hypothesis-based methodology will offer the data analytics team early and quick insights and sets the direction for iterative and incremental analytics processes. Fundamentally, the hypothesis-driven approach recognizes that there are multiple possible alternatives for any given problem, and it examines each using data to test and, ultimately, prove or disprove the assertions.

While there are many techniques for developing hypothesis-based thinking, one key technique is the McKinsey’s thought process called MECE - an acronym for Mutually Exclusive, Collectively Exhaustive, which separates the problem into distinct, non-overlapping issues while making sure that no issues relevant to the problem have been overlooked. MECE works by grouping elements that are mutually exclusive (ME) and collectively exhaustive (CE), which can then be used to logically categorize issues that can be analyzed systematically and minutely [Chevallier, 2016]. In simple words, MECE ensures that all elements listed cover the entire range of ideas while being unique from each other.

5. Execution mechanism on data analytics

Finally, how will the analytics team deliver? How will this delivery be different from the conventional IT-centric team? The execution mechanism of the analytics team will be different from the conventional IT-centric team in three main ways:

  1. They focus on data and the way the data is managed in its lifecycle
  2. They translate stakeholder goals into hypotheses and continuously define and redefine the hypothesis that can be verified using data.
  3. They deliver insights in an iterative and incremental manner.

In this backdrop, there are three key aspects in delivering or executing data analytics solutions in a business enterprise. Firstly, the analytics team should work closely with business stakeholders who believe in leveraging data for business performance. In some organizations, there might be some managers who are not convinced that data will improve business performance. Instead of educating and convincing these types of managers, it would be more effective to collaborate with leaders who believe in leveraging data for business results to start with.

Secondly, the analytics team should start on a small scale and focus on building trust and credibility with the business. What exactly does small mean for a data analytics team? Small could be a small number of use cases, engaging a small number of business stakeholders, working on small data sets with sample data, smaller time frames, small budgets, smaller projects like proofs-of-concept (PoC), and so on. Thirdly, the analytics team should focus on “good enough” analytics solutions. Analytics initiatives will rarely be fulfilling all the needs of all the stakeholders, given that the needs of the stakeholders and varied and diverse. Analytics solutions take refinement in an iterative and incremental manner, and the analytics teams should work on showing some small and significant wins quickly so that they can be positioned for bigger success in the enterprise.

When working on analytics initiatives, think big, start small, and act fast.

Conclusion

Building high performing enterprise data analytics teams is more than staffing people with ETL and report building skills. Today, as data and analytics extend their footprint across the organization, it is very important for all stakeholders to have a shared understanding of what will drive success. The business expectations today from the analytics team are on value creation along with compliance and performance reporting. Since data analytics in business enterprises today is the new language of business communication, the data analytics teams should work on ensuring that the data and insights are in the hands of competent leaders and front-line knowledge workers who will use data and insights to drive better business results.

References