Business Intelligence and Business Analytics, Big Time
In a world that’s awash in data, assembling the right information about customers and business conditions has never been more difficult. It’s a paradox that Jay Dittmann, vice president of marketing strategy at Hallmark Cards in Kansas City, Mo., knows all too well. Having 13 million loyalty program members who purchase thousands of different greeting cards and other items makes tracking trends a daunting proposition. “There are mountains of data to sift through and limits to our marketing budget,” he explains.
See Also SAS and IBM Execs Talk BI Analytics
About two years ago, Hallmark deployed a business intelligence (BI) and business analytics (BA) initiative to better understand buying patterns at more than 3,000 Hallmark Gold Crown stores across the United States. The company wanted to better nurture its relationship with its frequent buyers and, through predictive modeling, determine how to market to various consumer segments during holidays and special occasions. As a result of this initiative, Hallmark has boosted sales, while also simplifying and improving the analytics process.
See Also Analytics in a Nutshell
Business intelligence and business analytics aren’t new concepts. The idea of understanding the relationships between bits and bytes of data extends back to the late 1950s, and BI has been around in earnest since the late 1980s. However, today, the ability to aggregate, store, mine and analyze data can make or break an enterprise. As a result, BI and BA have emerged as core tools guiding decisions and strategies for areas as diverse as marketing, credit, research and development, customer care and inventory management.
In fact, BI and BA are evolving rapidly and meshing to meet business challenges and create new opportunities. Although nearly all Global 5000 organizations already use these tools, 35 percent of them fail to make insightful decisions about significant changes in business and market conditions, according to IT consulting firm Gartner. What’s more, the task isn’t getting any easier as data streams become more intertwined, and mashups and other Web 2.0 environments pull data from multiple sources.
The bottom line? “Business intelligence and business analytics are on the cusp of a major change,” observes Joseph Bugajski, senior analyst at Burton Group. “There is a shift toward providing deeper insight into business information. And there is a growing emphasis on better tools and putting more powerful and better software in the hands of business decision makers.”
Going Beyond Data
As more organizations attempt to sift through petabytes of data, it’s increasingly clear that BI and BA are as much about missed opportunities as about the ones companies tap into. Sifting and culling through data requires a robust infrastructure, effective data collection tools, and well-designed software for mining and analytics. Only then is it possible to identify hidden trends, customer relationships, buying behavior, operational and financial patterns, business opportunities and other vital information.
BI and BA are separate but connected tools. While BI provides a way to cull through data to find information—usually through querying, reporting, online analytical processing (OLAP) or other straightforward analysis tools—BA taps into statistical and quantitative data for explanatory and predictive modeling. For example, BA can predict which customers are likely to close accounts and can determine the optimal time to repair or replace a piece of equipment.
However, navigating BI and BA isn’t easy. That’s because numerous definitions exist, and the terms are loosely applied to a variety of tools and situations. Many companies continue to rely on spreadsheets as their primary BI tool, while others rely on dedicated and specialized applications from the likes of IBM, Oracle, PivotLink, SAS and Unica. (See “A View From the Top” on page 22.)
Steve Cranford, a managing director and practice leader at consulting firm PricewaterhouseCoopers (PwC), says BI typically involves the mechanics of turning data into information and then using dashboards and scorecards to disseminate it. Analytics, in contrast, centers on “solutions-oriented capabilities” that create value and transform information into knowledge.
Peering deep into data is certainly the goal at Independent Health, a Buffalo, N.Y., health maintenance organization with more than 300,000 members in eight counties in western New York. With health care reform looming and cost controls essential, Independent is continually searching for ways to become more efficient. “We look at health care as the business of information management,” says Joe Somma, director of market intelligence. “The better we perform, the better we’re able to deliver services and control costs.”
It’s no small task. At Independent, getting to results meant reshaping and rethinking things. “Putting the right data to use is essential,” Somma explains. Ultimately, the organization focused on three key initiatives: using predictive analytics to rate the probability of members’ contracting certain diseases, such as diabetes; using an application in-house to reduce costs; and adapting marketing campaigns to the preferences and demographics of different customer segments.
Independent Health turned to SPSS Predictive Analytics (recently acquired by IBM) to take its BA initiative to a higher level. Among other things, it developed a series of models that help customize communication with members, such as deciding the preferred method of contact, the scripts representatives should use and the services to suggest.
In fact, by overlaying demographic data from Nielsen Claritas, a market-ing and media communications company, with its internal customer data, Independent built a targeted marketing strategy that allows it to connect with members through a preferred channel—including telephone, conventional mail and e-mail (with Facebook and Twitter likely in the future).
The results have been impressive. Already, Independent has improved the efficiency and cost of campaigns by 27 percent, while achieving a 35 percent reduction in the use of outside vendors (saving approximately $350,000 a year). Moreover, the organization realizes as much as a 10 percent savings in direct costs by getting its members treatment earlier and on a preventive basis—particularly for conditions such as diabetes and heart disease.
In the future, Independent may also use BI and BA to identify fraud. “The technology is making it possible to become a far more efficient organization,” Somma adds.
One thing that makes newer BI and BA applications so powerful, says Boris Evelson, principal analyst for business intelligence at Forrester Research, is the ability of managers and others on the front lines of business to use these solutions to extract meaningful results. “They are no longer an exclusive tool for executives and power analysts,” he says. “They are enabling knowledge workers throughout organizations.”
BI and BA are evolving rapidly. For one thing, organizations are turning to these tools to incorporate larger and more diverse data sets, including unstructured data such as e-mail, audio files, video and documents. Evelson estimates that as much as 80 percent of the business intelligence residing in firms is held in spreadsheet files. “One of the biggest challenges organizations face is the amount of data sitting in silos,” he says.
“Too often, valuable data simply isn’t accessible or available.”
Part of the problem is that traditional BI tools aren’t all that flexible, and most databases aren’t designed for rapid change. Unfortunately, most mainstream BI vendors continue to produce products that don’t interface particularly well with unconventional sources of data, including social media.
“These tools are function-rich, robust, scalable and stable,” Evelson says, adding that most aren’t designed for today’s complex computing environment. In addition, many organizations lack the business processes, standards and governance procedures to develop a comprehensive BI/BA strategy.
Unfortunately, the situation isn’t getting any easier. As organizations interconnect data with business partners and outside sources, there’s a growing need to use mashups and Web 2.0 tools—as well as mobile solutions that push BI and BA into the field. Likewise, some enterprises are turning to clouds for data storage and cleansing—as well as to boost the flexibility and reach of their initiatives. And Burton Group’s Bugajski says younger workers are demanding more powerful Internet-based tools that extend far beyond OLAP and relational OLAP.
Plugging in diverse data sources and building a framework for effective BI is something the Kent County Council in the United Kingdom has grappled with over the last few years. Representing approximately 1.4 million citizens in southeastern England, the agency has looked for ways to make data more accessible to workers, citizens and other agencies—all while trimming costs.
“There is a lot of valuable information available,” says project manager Paula Rixon. “Unfortunately, it has been difficult to find it and put it to use.”
The council’s goal is to build an information-age distribution model for publicly available data. In 2009, it deployed IBM Mashup Center to connect to data residing in various spreadsheets and databases. It uses RSS to pull data into the BI application.
“We want to put citizens in control of information,” Rixon explains. An initial pilot project offers two dozen of the approximately 570 data sets available. The statistical data covers everything from auto accident and arrest statistics to the availability of medical doctors and clinics, and it’s accessible via a standard Web browser.
So far, the mashups have received favorable reviews from various constituents, Rixon says. As a result, the council plans to expand the use of the system significantly over the next few years and eventually put most of its data online and available for mashups.
“We are able to provide data at a lower cost, make it available more quickly and to a larger segment of users, and generate greater community interaction and involvement,” she explains.
Burton Group’s Bugajski says that new and better tools are ratcheting up the sophistication level for both BI and BA. “Organizations are looking to assemble data in new, more powerful ways in order to have a better understanding of events and conduct predictive analysis and what-if modeling.”
Meeting the Challenges
The road to BI and BA success is paved with plenty of potholes. One significant challenge is building the right IT infrastructure to manage data and ensure that it’s available and accessible throughout an enterprise. In many organizations, a veritable goldmine of data resides in software, systems and storage devices that simply aren’t accessible.
Worse, data may be stored and used differently in various organizational silos. If and when it’s reassembled, the data may wind up distorted or misinterpreted. Even tiny errors can result in drastically different results.
Another challenge lies in the design of these systems. An enterprise must create an interface that allows the individuals who are using reporting and analytics tools to navigate through systems and data. In some cases, it’s essential to mix and match traditional reporting tools with complex analytical functionality, while simultaneously providing an interface and environment that are simple enough for a wide range of businesspeople to use. What makes this task especially difficult is that different departments may have developed different tools and interfaces—in some cases, pulling data from disparate ERP and CRM systems and databases.
Knocking down the silos and combining all the data increasingly require an organization to adopt what Burton Group refers to as a data services platform. These middleware tools help connect a scattered BI infrastructure within a single interface and environment. Vendors such as Oracle (through its acquisition of BEA and its AquaLogic solutions) and Composite Software have begun to address the need for a more integrated environment and more sophisticated management capabilities.
In addition, as organizations adopt virtualization and cloud environments and use them to store and manage data, the ability to integrate disparate environments becomes even more vital. A big advantage of storing data in the cloud, Bugajski says, is that it reduces the demand on IT resources. In some cases, organizations can trim the number of servers required for BI and BA. There’s also the ability to quickly grab test data from several sources. “Oftentimes, when you do this kind of analytics, it beats up hardware and disk requirements for a one-time sort of use,” he adds.
Governance is an equally important factor in achieving success with BI and BA, notes PwC’s Cranford. “An organization must have an effective governance model and have policies and practices in place to ensure that technology is applied appropriately and that it enables the best possible results,” he says.
What’s more, executive sponsorship is a must, and a strategic road map is paramount. “These initiatives require more than technology,” Cranford points out. ”They require a focused approach.”
Hallmark’s Dittmann knows that very well. When the company amped up its BA strategy, it turned to SAS Analytics to mine data and spot patterns that could lead to improved sales. It runs the platform on a Teradata database that’s connected to a mainframe, with data updated nightly. “We have huge amounts of data to sift through, and understanding customer behavior and communicating with loyalty program members in a targeted way are vital to our success,” he says.
Hallmark turned to predictive modeling so it could adapt and adjust promotions on the fly. Today, it can determine which customer segments are most swayed by direct mail, which segments should be approached through e-mail and what specific messages to send to each group. The company is now able to assemble customer insights in 10 minutes rather than the two days or longer it used to take. Even more important, Hallmark has boosted sales to its loyalty program members by 5 percent to 10 percent.
“Picking the right customer to talk to at the right time is crucial,” Dittmann explains. “Understanding trends and the general health of the business is essential.”
The way Forrester’s Evelson sees it, the task of managing BI and BA isn’t getting any easier. Yet, as organizations seek to understand complex factors and variables in a more holistic and strategic way, putting data to maximum use will continue to grow in importance.
Evelson sums it up this way: “BI and BA aren’t routine applications; they are major differentiators in terms of frontline competitiveness. Real-time monitoring, split-second reporting and predictive modeling are both the present and the future of business.”