Mining Data to Make Better Decisions

John Baltes hates the term "business intelligence". Most people associate the term with analytics, but in Baltes' view, it's really about changing business behaviors.
Baltes, principal/managing partner for SVA Consulting, thinks "BI" has led many an organization – business, government, and nonprofit – down an expensive and failed path. The ability to pull revealing and actionable data from dozens of internal business systems, for the express purpose of supporting business decisions, is definitely enticing, but there is a method to monitoring, reporting, or analyzing data.

"Frankly, business intelligence is misnamed," Baltes said. "We're talking about the process of getting information back to the organization about performance. It's about taking data and turning it into information to answer questions. It's not completely about what people think, which is analytics.
"It's more about being able to get data and change behavior in a real-time fashion."

Judgment calls

Before rushing out to purchase dashboard-building tools, think about program design. It might require some patience, but the results of a well-designed system will pay for themselves over time. It starts with management's understanding of the most impactful information for decision-makers, and building a flexible system that enables seamless modification and the retention of all necessary corporate information, even if a certain business system is no longer available.

As organizations ponder business intelligence, they will approach a fork in the road. The first management judgment, which comes before deciding whether to invest in a tool like Business Objects or a host of others, is whether to warehouse data or attach a tool to business applications.

While this is not a universally shared view, Baltes is not a fan of the latter. "The reason is because a lot of times people get that data delivered to them, and they don't know what to do with it," he said. "If we give them data, and they can't relate the data to a change in behavior, it's all for naught."

In contrast, he believes data warehousing is more accommodating to the process of understanding how data relates to managing business and what behaviors must be changed.

The next judgment call pertains to sorting data. Are there systems available to actually sort the data you want? In many cases, companies want data, but they don't have an application to catch it, in which case they must back up and start creating systems to accomplish this.

At that point, it's time to review the technology behind how data will be delivered, and where it's stored. "At the end of it, you look at how you are going to present the data," Baltes said. "You present that data in three ways: either through an analytics tool, through a reporting tool, or through a dashboard."

Baltes believes that when data warehousing is done correctly, these tools don't matter because they should produce the same answers. The key to data warehousing is that answers are stored in the databases, and "you don't make the presentation layer calculate the answers," he said. "A prime reason projects fail is because they started at the presentation layer. They want to know what their dashboard is going to look like.

"They start with that, versus starting with the business questions they are trying to answer, and they got sold a bag of goods by someone who gave them a lot of pretty pictures and didn't do the business analysis of correlating the data with how the business is being run and with how decisions are being made."

Correlations, he added, are among the most important information that can be gleaned. "There is a whole new method of management that is coming about," Baltes stated. "Due to the overhead, we can no longer build these historic management levels that we've had in companies, so being able to present data back to people, and let them measure their own performance against what the organization wants from them, is a new way of managing."

Tool time

The "design first" mantra is not meant to suggest that technology tools are unimportant. In fact, there is a software evolution occurring that better enables organizations to tailor their own solutions, according to Ginney McAdams, national business development manager for CoreBTS. The proliferation of data produced by cloud computing, social media, and mobile devices has sped the emergence of self-service business intelligence and discovery.

Traditionally, there has been limited access to the powerful analysis required to make sense of data, McAdams said, and it has taken a long time to get answers. With new self-service tools being developed by new market players, and components of existing players like Microsoft, the evolution of intelligence software is changing how data is melded together.

"You now have, because hardware is a lot cheaper, tools that are designed to handle large data sets, so you don't necessarily have to consolidate all of your data sources into this massive data warehouse," McAdams said. "Many of these newer tools allow you to actually pull the information from disparate data sources into a single model and then build your reports and dashboards based on that information. It's an improvement in the process, and it allows you to be more flexible in the self-service environment."

Traditional reporting solutions generally are more static with columns and rows and data, and there is a place for those. In a true self-service dashboard, McAdams said users will be able to create their own ad-hoc queries, conduct analyses, and perform what-if scenarios because they can slice and dice the data in a way that makes sense to them. "They can answer more questions on the fly, rather than figuring out what kind of reports they need IT to build," she said. "If IT has to figure out what field to pull into that report, then it's static."

Analytical technology is more commonly associated with mid-sized and large businesses, but McAdams notes that vendors are making BI tools for the small business space. "The small business probably needs to be even more agile to be competitive," she said, "so having ready access to information that is pertinent to their business is a key to their success."

Mastering the curriculum

Much of the BI focus is on business organizations, but units of government are finding uses for it. Margo Sickele, director of accountability and school improvement for the Harlem School District in Machesney Park, Ill., located just south of the Wisconsin border, said district teachers wanted an analytical way to inform classroom instruction and help teachers make instructional decisions for the district's 7,500 students.

Before building a data warehouse to capture information from various business systems, the district established critical needs and created a plan based on those needs. A number of companies provide pre-packaged software, but the district felt it would have more control over how it pulled data with an in-house system, which was built by a data analyst in the district's information technology services department.

The data warehouse can pull and aggregate data from systems like Skyward, the district's student system, and from academic services like the Diagnostic Reading Assessment to provide both individual and group pictures of how students are performing.

Teachers now have access to data on individual students' strengths and weaknesses, which enables them to tailor individual instruction plans. If they find that an entire classroom is struggling with something, they can design a whole-class remediation, and they can do so without the inefficiencies associated with poring through multiple data sources.

Sickele's advice to other organizations is to start with a needs analysis in every layer of the organization. "They [teachers] can really problem-solve now," Sickele said. "It's really powerful. They love using an electronic format instead of putting together packets of paper, so it's much more efficient."

Fewer layers

For organizations contemplating this investment, Baltes always asks why. "To me, the reason business intelligence is becoming so prolific is because business systems are getting very fragmented," he said. "They have many systems in place, not just one or two, but maybe dozens in place. The second thing is the cost has come down so dramatically in terms of the ability to do this, compared to what it was even four or five years ago.

"The third thing is more C-suites understand they need to give people information so they can monitor their own performance, not put levels of management in place that look at data and then try to get people to change their behaviors."

Sign up for the free IB Update – your weekly resource for local business news, analysis, voices, and the names you need to know. Click here. If you are not already a subscriber to In Business magazine, be sure to sign up for our monthly print edition here.