Executive Decisions: How to Organize Your Analytical Talent
Some models are more effective than others at engaging and retaining talent while also getting the most value from their efforts.
By Jeanne G. Harris (left), Elizabeth Craig (center) and Henry Egan (right)
Companies increasingly rely on a relatively scarce breed to maintain their competitive edge: the people who are able to use statistics, rigorous quantitative or qualitative analysis, and information-modeling techniques to make business decisions. Or, in shorthand, “analytical talent.” For business leaders, the importance of such people poses several challenges, but in our experience one question stands out: What’s the best way to organize our analysts? Executives have to decide whether they should be centralized or decentralized; “charged out” to the rest of the business as consultants or made available as a free resource; where and to whom they should report.
The stakes are high. Companies that choose the wrong model often relegate their best analytical minds to conduct simple analyses or to work on low-value projects. Such missteps are a sure recipe for analyst disengagement and defection. Companies that seek to increase the impact of analytics on the business – while building a highly engaged analytical workforce at the same time – also face a dilemma. They need to organize analysts so that they are working “close to the business” on critical analytical initiatives while also keeping them “close to each other” for purposes of coordination, mutual learning and support. Making both happen at once is the challenge.
Are companies up to the task? To find out, we interviewed dozens of executives and surveyed more than 700 analysts. While the right choice can vary, we found that companies with a strong commitment to building an analytical workforce are best served by greater centralization and coordination of analysts. The right organization model is also critical to engaging and retaining this ever more important segment of the workforce.
The key to getting started? Ask the right questions about your level of analytical maturity and the speed and level of change your organization is prepared to undertake.
Five Organizational Models
Our research revealed that large, multi-divisional corporations choose from one of five models to organize their analysts (Figure 1; these are described in greater detail in “Analytics at Work: Smarter Decisions, Better Results,” by Thomas H. Davenport, Jeanne G. Harris and Robert Morison, to be published in February 2010 by Harvard Business Press):
1. Centralized. Analysts reside in one central group where they serve a variety of functions and business units and work on diverse projects. Besides providing analytical expertise and support, a centralized unit sets the analytical direction of the entire organization. This structure makes it easier to deploy analysts to projects with strategic priority. Candy giant Mars uses this approach; its centralized Catalyst group has long-term funding and works on strategic initiatives across the company.
2. Center of excellence. Analysts work in units throughout the organization, but their activities are coordinated by a central entity. The center builds a community of analysts, primarily to share knowledge and best practices. It can also double as a project management office, looking at many analytical initiatives, setting priorities and staffing projects. Capital One and Wal-Mart both employ versions of this model.
3. Consulting. Analysts work together in a central group but act as internal consultants who charge business units for their services. The consolidation of analysts enables some enterprise-wide coordination of analytical activities. United Airlines, eBay and trucking firm Schneider National all employ the consulting approach.
4. Functional. Analysts are located in the functions where the most analytical activity takes place, such as marketing or the supply chain. Fidelity uses this approach: most analysts work in the Customer Knowledge group, which reports to marketing. Analysts are also found in large numbers in marketing at GE Money (although they also work for the risk management function and other groups) and at Carnival Cruise Lines.
5. Decentralized. Analysts are scattered across the organization in different functions and business units with little coordination. Companies with few analysts and little management support for analytics are most likely to be decentralized.
Our survey data reflects the variety of organizational models across industries. The decentralized model was most prevalent (reported by 42 percent of analysts surveyed), indicating the immaturity of most corporations’ analytical capabilities today. Consumer-oriented industries that we’d expect to have advanced analytical capabilities — including retail, consumer goods and services, financial services and health and life sciences — tend to centralize their analysts. Company size wasn’t a key factor, except in the most extreme cases: Companies with fewer than 1,000 employees were more likely to organize their analysts in one functional area, whereas companies with more than 25,000 employees were more likely to establish a center of excellence.
Analysts who work together in centralized groups or are connected through a center of excellence reported the highest levels of engagement and were the most likely to stay with their companies. In fact, analysts in centralized groups were twice as likely to be highly engaged as analysts who were decentralized (see Figure 2). Analysts in centers of excellence reported the strongest intentions to stay with their employer, whereas those in the functional, consulting and decentralized models reported significantly lower levels of engagement and weaker intentions to stay (see Figure 3).
If engaging analysts were the sole goal, our recommendation would be simple: centralize them. But in practice we’ve found that the ideal model depends on a company’s priorities, the maturity of its analytical capabilities, and the need to balance analyst supply and demand (see Figure 4). When just starting to use analytics, most companies don’t need to use them in every area of operations. Nor will they have enough analysts to justify centralizing resources — making a functional model a natural fit. As demand grows across the company, companies tend to hire more analysts. Once an organization gets a critical mass of experts and demand for these scarce resources grows, one of the more centralized models is the right choice.
How Organization Affects Analysts
Meaningful work and career opportunities are critical for engaging and retaining all types of employees, and analysts are no exception. How companies organize analytical talent affects whether they have access to the most meaningful opportunities — work that allows them to use their highly specialized skills, gain valuable experience and contribute to the organization’s goals.
Three key factors influence the quality of analysts’ work and career opportunities: 1) that their work is aligned with the organization’s strategy and goals and affects its success (“significant work”), 2) that they understand the business (“business insight”), and 3) that their skills and aspirations are a good match with the company culture and goals (“organizational fit”).
Without the chance to make a real impact on the organization’s success, analysts won’t find enough meaning in their work, and so they will be less engaged and less likely to stay. For the most sophisticated analysts, having to spend too much time on simple analyses and report generation quickly saps their motivation. We know of several organizations that have lost analysts because they were treated largely as “spreadsheet developers.” It’s essential to give your best analysts opportunities to apply their expertise to the company’s biggest problems.
One analyst, Sharon Frazee, gave us an example of how significant work was connected to engagement. At one point in her career, she led a successful and highly engaging initiative to standardize information and to streamline its distribution — changes that gave line managers more time to analyze potential growth opportunities. Frazee, now vice president of corporate healthcare analysis and research at Walgreen’s, explained her passion to contribute: “I want to make a difference. Being able to do the kind of informatics work that actually gets applied, and to see things changed because of it, is important to me.”
Analysts in centers of excellence and centralized units were most likely to report that the work they do is important. Analysts who are organized as internal consultants, housed in functional units, or dispersed throughout the business are more likely to be stuck working on peripheral projects.
Organizations don’t just want “quant geeks.” They need people who also understand the business and can develop strong relationships with business leaders. How analysts are organized can have a major impact on this dimension.
According to our survey, analysts in centralized units, functional groups and centers of excellence had more insight into their company’s business than analysts in internal consulting groups or analysts who were decentralized. These more centralized groups offer analysts an enterprise perspective: One in three analysts in centralized units has a clear understanding of how their group contributes to the organization’s success, whereas just one in five analysts in the other models has this degree of insight.
GE’s financial services unit, GE Money, operates analytical centers in Shanghai and Bangalore. Staff members from the centers routinely rotate through other parts of the business. These assignments help analysts learn about local operations and prepare them to meet their needs in the future. Equally important, they offer analysts a sense that they are making meaningful contributions to the business.
In many companies, analysts are viewed differently from the rest of the workforce; one executive we interviewed called them a “weird species.” Analysts themselves take pride in their uniqueness. Still, they want to work for companies that value analytics and with colleagues that appreciate and respect their unique talents. They are most engaged when they believe their rare and valuable skills are a good match with the company culture and goals.
We found that organizations with a centralized analytics function are better at communicating the value they place on analytics and at making analysts feel “at home.” Nearly half of all analysts working in centralized groups report a high degree of fit with their organization; only one in 10 feels their company is not a great fit for them. By contrast, analytical talent deployed as consultants or concentrated in one functional area often feel like misfits — isolated from the business and other analysts.
Grouping analysts together helps to maximize the fit between analysts and the organization. For example, Dr. Steven Udvarhelyi, senior vice president and chief medical officer at Independence Blue Cross, reports that their internal informatics organization is critical for engaging and retaining analytical talent: “It’s a defined center of excellence. It creates a fertile ground for people to work with other people. It creates a critical mass where you get career opportunities, growth opportunities and good professional interaction.”
How do you determine which model is best for you? That answer depends on the maturity of your company’s analytical capabilities as well as the quality and quantity of demand for analytical skills. If your organization has little demand for analytics, has only a handful of analysts and no executive sponsorship, then it’s best to house analysts wherever they have a willing sponsor. In more analytically sophisticated companies, demand for analytics quickly outstrips the supply. In this case a hybrid approach, with a centralized team of the most advanced analysts complemented by a center of excellence for analysts deployed to functions or business units, may be the best approach. An enterprise-level organizational model makes it much easier to deploy analysts to strategic business initiatives.
The next challenge: how fast can you get there, and what’s the ideal pace of change? Organizational change should be broken into stages — quick wins versus longer-term goals. For example, linking decentralized analysts through a community of interest is relatively easy to do. Setting up a more formalized center of excellence, with specialized training and development for its members, takes longer. The potential for change will be determined by the maturity of your organization’s analytics function, as well as the strength of leadership and the level of senior executive commitment to analytics.
What if you can’t easily change your organization’s model? Leave analysts wherever they have enough sponsorship to protect them and advocate for adopting a more enterprise-wide perspective (which would require, for example, consistency in skills, job descriptions, training, salary and career development opportunities). If you use the consulting model, find ways to demonstrate to analysts how their work fits with and contributes to organizational goals. Build a sense of community in order to share best practices and improve analysts’ sense of fit with the organization. If your analysts are concentrated in one functional area or are decentralized, you’ll need to find ways to provide them with an enterprise perspective and to build links between them and with other parts of the business. Building an informal community of interest may be the best place to start.
And if your analysts are already centrally organized, or in a center of excellence, that doesn’t mean all your problems have been solved. While centralized organizational structures are superior, they’re not perfect. Explore ways to link centralizedanalysts more closely to people in the business. If you have a center of excellence, it’s important to reinforce the links between analysts and to make sure that senior management fully supports the analytics function.
Company leaders need to recognize analytical talent as a special segment of high-value employees, whose preferences and motivations vary significantly from those of other employees. Organizing analytical talent in a way that not only addresses the strategic and operational needs of the business but also provides analysts with meaningful work and career opportunities is essential (see Figure 5).
Managers who value fact-based analysis create the kind of challenging and important analytical assignments that allow analysts to thrive. When managers don’t value analysts’ work, then no matter how they are organized, they will be under-used and unappreciated — two guaranteed ways to sap the motivation of the best analysts — and to send them to your competitors.
Jeanne G. Harris (firstname.lastname@example.org) is an executive research fellow and director of research at the Accenture Institute for High Performance. With Thomas H. Davenport, she is the author of “Competing on Analytics: The New Science of Winning” (Boston: Harvard Business Press, 2007) and, with Davenport and Robert Morison, “Analytics at Work: Smarter Decisions, Better Results” (Boston: Harvard Business Press, forthcoming, 2010). She is based in Chicago. Elizabeth Craig (email@example.com) is a research fellow at the Accenture Institute for High Performance in Boston. She is the author, with Peter Cheese and Robert J. Thomas, of “The Talent Powered Organization: Strategies for Globalization, Talent Management and High Performance” (New York: Kogan Page, 2007). Henry Egan (henry. firstname.lastname@example.org) is a senior specialist with the Accenture Institute for High Performance in London.