Who’s Counting?

by Susan Cohen

Analytics can benefit so many groups throughout an organization – customer service, inventory, forecasting, direct marketing and human resources, to name a few. With so many groups that benefit, whose responsibility is analytics?

Some companies have formed centralized analytics teams that report directly to the CEO. Others appoint an analytics guru to each functional team. And some individuals might practice do-it-yourself analytics in addition to following the advice of their company’s quantitative experts. Which option is right for your company?

This column will define each option and discuss some of the advantages and disadvantages of each.

Centralized analytics team
A dedicated analytics group is a powerful asset for any executive team. It provides analysis and insight that enable managers and functional team members to make better decisions – decisions based on careful analysis.

A centralized analytics team often works like an agency to corporate clients, providing structure, methodology and expertise to each project.

Anne Milley, director of analytical strategy at SAS, succinctly summarizes the benefits of a centralized team as providing "economies of scale and skill." Not only do economies of scale reduce duplicate work, but they also enable faster ramp-up for new projects that might be similar to past projects, promote the sharing of best practices and provide for established communication channels for disseminating findings.

Economies of skill create a strong talent pool of quantitative experts. Since team members work together on projects, they can learn from each other and share significant findings more easily. Furthermore, managers are likely to be quantitative experts as well, providing mentoring and staff development.

On the downside, a centralized team can lack intimate knowledge about each business unit, requiring more input from business mangers for each project. Also, projects that are important to a single unit but not to the larger corporation may never make it to the top of the priority list, leaving units to fend for themselves.

Embedded analytics experts
In some companies, analytic experts are scattered throughout an organization, reporting to and sitting with the functional teams. Since the quantitative expert is part of the business unit, she understands the business issues more intently. She can suggest projects that will benefit her team and will quickly understand the business issues surrounding each project.

Moreover, since the analyst reports to the business unit, the unit leader can prioritize analytics projects, leading to a more flexible and responsive system for each unit.

This structure can also lead to a less formal relationship between the business manager and the quantitative expert. For example, casual discussions might replace more formal reports and presentations. Or business managers might be more comfortable sitting with the manager to examine "what-if" scenarios, adjust a model or prioritize upcoming projects.

There are, however, some significant drawbacks to a decentralized system. Companies have complained about higher costs due to work duplicated across multiple business units. Others report difficulties in communicating best practices and insights throughout the larger organization.

Do-it-yourself applications
New offerings from analytic software solution vendors like SAS bring analytic power to business users. These solutions, in addition to solutions developed by in-house analytics and information technology teams, put the power of analytics in a business context. This enables business users to complete routine tasks with better information, such as predictive modeling for employee recruitment and retention planning; customer segmentation analysis for marketing campaign planning; and strategic sourcing and supplier portolio management.

Once an analyst develops the model and creates conduits for the data, business users can use the model to make business decisions, freeing up analysts for other projects.

Bringing analytical power to business users enables the business unit to respond even faster to changing business environments. And empowered business managers are able to make better decisions.

Not all analytic projects should be turned into do-it-yourself (DIY) applications, which work well with repetitive tasks and static models. DIY applications also make sense when business users are making decisions with no quantitative input. A DIY application may provide enough information for better decision making when a quantitative expert isn’t available.

SAS’ Milley warns that companies must choose the right skill level for each application. Pushing routine tasks to business users can free up analysts for other tasks. But pushing a complex function to business users might deliver inaccurate results. Additionally, business users need to understand when to apply the application and how to interpret results.

Organizations that employ do-it-yourself analytical applications often find they require quantitative experts and analysts to answer questions and perform follow-up analyses. Surprisingly, many organizations find an increase in demand for quantitative experts when they roll out DIY applications.

There is no single organizational structure that makes sense for all companies. In fact, many companies are most successful when they create their own unique blend of structures – a centralized analytics team for larger, strategic projects, embedded quantitative experts for daily operations and business managers who can run models to make routine decisions.

Regardless of the structure, good communication and teamwork are essential.