BI success factors: Is your company ready to profit from big data?

An ancient two years ago – in 2010 – International Data Corporation (IDC) researchers estimated that the world’s digital data had surpassed 1021 bytes, or the equivalent of 125 billion fully loaded, eight gigabyte iPods. By the end of 2012, IDC forecasts that 2.7 zettabytes will be stored worldwide. How will companies profit from all these data? It’s going to take more than machines and number crunchers, says information systems Professor Uday Kulkarni.

An ancient two years ago — in 2010 — International Data Corporation (IDC) researchers estimated that the world’s digital data had surpassed a zettabyte. That’s 1021 bytes — one trillion gigabytes — or, as noted in the blogosphere, the equivalent of 125 billion fully loaded, eight gigabyte iPods. By the end of 2012, IDC forecasts that 2.7 zettabytes will be stored worldwide.

It’s easy to see how such data accumulate. A story that appeared in Information Management noted that the 90 million “tweets” zinging through Twitter each day total around eight terabytes. Facebook’s 750 million users create some 30 billion pieces of online content every month. The sensors on a jumbo jet can generate 640 TB in one transatlantic flight, and there are hundreds of these flights each day. Walmart handles more than a million transactions every hour and feeds an estimated 2.5 petabytes into its database.

How will companies profit from all these data? It’s going to take more than machines and number crunchers, says Uday Kulkarni, professor of information systems at the W. P. Carey School of Business. According to him, the factors that contribute to success with business intelligence (BI) are both human and technological, and he’s currently testing this hypothesis through surveys conducted around the world. Early results of his research suggest that corporate leaders must commit to data-driven decision making and encourage it among workers. Without leadership and a corporate culture that values business intelligence, organizations may not gain all the benefit BI systems can deliver.

Man vs. machine

Kulkarni partnered with José Antonio Robles of Universidad ESAN to begin his research in Lima, Peru. There, the team created a survey that evaluates the relationships between technology and human components of IT-based analytics. “We are looking at various factors that contribute to business intelligence success,” he says. “We break them down into BI capabilities, which include both data and systems capabilities. And then there are organizational factors we examine. One we call analytic culture, another is leadership support for business intelligence, and a third is user involvement.”

As Kulkarni and Robles note in a paper summarizing the Peruvian survey results, that although data are often overabundant, they are often imperfect, inconsistent or inaccurate. “It is estimated that more than 50 percent of BI projects have failed because of data quality issues,” the scholars write. “Data quality issues alone cost U.S. businesses over $600 billion a year.”

Along with data quality, the researchers looked at BI systems capability, which they bisected into usability and functionality. Usability is the system’s ability to support decision making. Functionality includes the system’s ability to perform activities, as well as its ability to be customized — a task that can involve feedback from users themselves.

And, then there’s the human factor in BI. “Leadership needs to consider their internal way of looking at data … the culture is part of it,” Kulkarni says. “This consists of encouraging people to make informed decisions based on data rather than gut feel. It involves asking people to provide supporting evidence for things they have done and expecting measurement when people evaluate decisions. It also means having people who are good at analysis and relying on them. All of this contributes to the analytic culture of the organization, and these things are necessary for technology to make a difference in decision making.”

When Kulkarni and Robles asked survey respondents if they had an analytic culture at their organizations, some 70 percent said yes, and median answer was 3.85 on a scale of one to five. A similar number of respondents indicated that their firms’ senior managers support BI initiatives. Overall, leadership-support questions earned a median score of 3.79.

Survey respondents were less positive about their companies’ data capabilities, which earned a mean score of 3.3. Not surprisingly, perceived usefulness of the systems was only rated 3.43, and overall user satisfaction came in at 3.0, meaning users were neither satisfied nor dissatisfied with the BI capabilities in their organizations.

User involvement with the BI systems also was rated at midpoint. It earned an overall score of 3.09. Given that the mean response for “BI success” in the various organizations was 3.22, Kulkarni and Robles see “a tremendous opportunity for BI professionals to understand the users’ needs and provide them with appropriate support for decision-making tasks.”

The researchers came to this conclusion because BI systems capability earned a mean score of 2.86, but cultural factors scored in the 3.8 range, indicating that organizations have started preaching evidence-based decision-making, and users see its value. Now, companies need to fine-tune the tools that support use of BI insights.

“There is a lot of research on user involvement when you build new ERP systems,” Kulkarni notes. He says a typical 18-month development cycles provide ample opportunity to get user input into requirements analysis, system design and testing, which is a good thing. “The more you involve users in development, the more acceptance and user satisfaction you’ll have.”

But, he adds, BI systems tend to be so pervasive, they don’t lend themselves to traditional user involvement in development stages. “Where you need user involvement is in improvement phases. Systems developers need to ask users how they are using the systems and what more they can do. It’s not a development process. It’s more of a learning process,” he explains.

That’s why their survey instrument tries to measure users’ knowledge of system features, as well as the users’ commitments to making their systems better. “These are necessary conditions for ultimate success,” he says.

Of course, there are other success factors beyond Kulkarni’s initial research, and these, too, are split between technology- and people-related conditions. On the technology side, Kulkarni points out: “Only about 5 percent of the data we collect today is ‘structured,’ which means that it’s in the form of numbers, text or other records that can be processed easily. The rest of the data are ‘unstructured’ and in the form of things like videos, music and photos.” We’re getting better and better at analyzing that unstructured data, but we still have a long way to go.”

That’s not the only trouble spot. “The weakest link today is people who can perform the types of analytics to make good decisions from data. We need people with degrees in ‘business analytics’ – statistics and mathematical modeling with business knowledge,” Kulkarni says.

In fact, some organizations are growing their own BI wizards. Capital One, for example, sponsors graduate students with free tuition for those who will join their team.

And, leaders themselves need training, which is something Kulkarni teaches in his W. P. Carey MBA courses. “We need a clearer understanding by corporate leadership about what problems can be solved with BI. We need more education at top corporate levels because tomorrow, these are the folks that will build the BI teams and invest in BI projects that will give their firms a sustainable competitive advantage.”

With this education, corporate leaders will also be more likely and able to provide the support and direction that fosters user acceptance of a data-driven corporate culture. Without leadership support and the analytical culture it creates, BI technology investments will probably be far less effective. At least, that’s what the data from Kulkarni’s research efforts indicate.

Bottom line

  • Data-driven decision making can boost a company’s bottom line, but companies must invest in certain things to make it happen.
  • W. P. Carey School Professor Uday Kulkarni believes effective business intelligence requires technology, such as data collection capabilities and powerful information systems, as well as human factors, including analytic culture and leadership support for BI.
  • Kulkarni has been testing his theories via surveys around the world.
  • Data collected from his surveys is indicating that, indeed, the people part of BI is a significant factor in creating a profitable BI initiative.