Analytics and the Bottom line: How Organizations Build Success | Harvard Business Review

By Tom Davenport, Jeanne Harris

Analytics in business is about far more than just reporting. Thanks to bestsellers such as Moneyball and Super Crunchers, statistical analysis has broken through to public awareness. At the same time, the Internet and connected devices have brought a flood of valuable new data to analyze.

Those who view analytics as just reporting on past performance don’t understand the full scope and value of analytics. Analytics can be thought of as:

By Tom Davenport, Jeanne Harris

Analytics in business is about far more than just reporting. Thanks to bestsellers such as Moneyball and Super Crunchers, statistical analysis has broken through to public awareness. At the same time, the Internet and connected devices have brought a flood of valuable new data to analyze.

Those who view analytics as just reporting on past performance don’t understand the full scope and value of analytics. Analytics can be thought of as:

Descriptive analytics. This is “the what.” It describes what happened in the past through reports, queries, drill downs and alerts.

Predictive and prescriptive analytics. This uses data to understand the “so what.” Predictive analytics – which include forecasting, predictive modeling and optimization – are focused on the future. The use of predictive analytics takes an organization to a higher degree of intelligence and can yield competitive advantage. 
Analytics’ potential contribution is great, especially in light of exciting advances in behavioral economics, neurobiology, artificial intelligence, and even “the wisdom of crowds.”

The stages of analytical maturity

To become an analytical competitor, a company must evolve through stages of capability. There are five stages of analytical maturity. Unless data analysis is embedded in the corporate DNA, such as at Google, companies move through levels of progressive competence.

Stage 1 – Analytically Impaired. Executives aren’t asking analytics questions.

Stage 2 – Localized Analytics. Pockets of analysis exist in localized silos, but the focus is on the past rather than the future.

Stage 3 – Analytical Aspirations. These firms see the value of analytics and prepare a detailed road map to move from descriptive to predictive and prescriptive analytics.

Stage 4 – Analytical Companies. These companies are highly data-oriented and make use of analytics, but lack a top-level, passionate commitment to competing on analytics.

Stage 5 – Analytical Competitors. These “Zen masters” use analytics as a competitive differentiator. These analytical competitors have proven that the use of analytics is linked with positive bottom-line results.

Analytical competitors include:

Old hands who have used analytical methods for years, such as Marriott (revenue management) and Progressive Insurance (risk and pricing models).

Turnarounds that used analytics as a/the key strategy in restoring a company to health, such as Harrah’s (customer loyalty) and Tesco (Internet groceries).

Born analyticals like Capital One (data-based credit offers) and Netflix (recommendation engine) that have used analytics since their birth.

The DELTA model

The DELTA model describes the analytical capabilities model that companies need to get right.
A mix of interlocking capabilities defines successful companies’ analytical efforts. The symbol for change, DELTA, is the authors’ acronym for:

Data. A prerequisite for analytics, data must be clean, integrated, shared and accessible in a warehouse. Today’s data incorporates more than numbers: Text and images have a role, too. For competitive advantage, data should measure something new and important. Casinos, for instance, calculated that a staff member should smile at a customer every 10 minutes to optimize betting revenue.

Enterprise. To avoid multiple local versions of the facts, an enterprise perspective is critical. This enterprise perspective is necessary to determine the highest-impact performance factors, optimize investments across products/geographies/channels, and align decisions with company strategy.

Leadership. Senior management commitment to analytics is essential. If management is not committed (as is often the case), then secure support by running a pilot, measuring the benefit, and spreading the news.

Targets. While some analytical targets, such as pricing strategy and performance measurement, are common to most businesses, others are specific to particular industries or companies. Health care firms need to track drug interactions, for example, while financial services companies must focus on fraud detection, and online firms study website metrics.

Analysts. Most of a company’s analysts (70-80 percent) are “amateurs” who use spreadsheets and run queries. About 15-20 percent are “semipros” who can use basic statistical tools and create simple models, while “pros” (5-10 percent) can write their own algorithms. Significantly, the “champions” (the 1 percent who lead analytical initiatives) are not necessarily “quants” themselves, but leaders who know how to pose the right analytical questions.

These capabilities form the foundation of an analytical organization. In addition to analytical capabilities, an organization needs the appropriate context to compete on analytics.

It is not enough just to have the necessary analytical capabilities; competing on analytics requires institutionalizing analytics through culture and processes.

Culture. Facts, evidence and analysis are central to decision making in an analytical culture. When facts are thin, an analytical culture embraces a test-and-learn ethos, and it isn’t shy about demanding, “Where’s your data?” And the focus of a culture that competes on analytics isn’t just engaging in analysis; it is taking action after the analysis is done, then returning to the data to inform the next decision. At its best, an analytical culture also can integrate hunches and intuition based on experience.

Processes. An analytical business process is one in which analysis plays a role from start to finish. At Best Buy, for example, marketing begins with strategic customer segmentation and basket analysis; continues with promotional optimization, advertising metrics, in-store assortment and placement analysis; and concludes with post-transaction customer experience and loyalty research – all of which is raw data for improving the next marketing campaign.

Despite progress in analytical tools and data gathering, decisions are not improving

While the amount of data has increased exponentially and the use of analytics has grown, most corporate decision-making processes ignore or misuse this body of knowledge.

Although it is easy to mistake reports and scorecards for the goal, better decisions are analytics’ real purpose. By analytics, we mean more than data and quantitative methods; it’s also about better decision making.

Better decisions emerge when companies systematically:

• Identify their critical decisions.
• Inventory those decisions that require analytical help.
• Intervene where needed.
• Institutionalize what was learned.

Beyond analytics (which is the most commonly used intervention to drive better decisions), interventions being used to improve decisions include cultural and organizational changes, adjustments to processes and methods, and new training and communications.

An example of how analytics can be used to make better decisions comes from Stanley (tools). Using analytics to make better pricing decisions helped Stanley increase its gross margin from 34 percent to over 40 percent in six years.