Companies are returning to data, especially predictive analytics data, in order to achieve business goals.

Big data is back.

Not that it ever left, according to Kally Pan, accounts director for Leftronic, a San Francisco-based real-time data visualization platform provider. But Pan says companies are finally returning to the notion that data can be leveraged in order to help companies meet business goals. The problem, however, could be in what data they’re using.

Back in the 1990s, Pan says big data spawned by the business intelligence movement created a metrics-driven culture. Now companies are returning to data, especially predictive analytics data, in order to achieve business goals. However, Pan says the metrics aren’t always used properly.

“Essentially, people think of business intelligence as some sort of magic bullet that comes in and tells you what to do,” he says. “It’s not — it’s a tool. It’s like saying ‘I’ve bought a typewriter, now where’s my term paper?’”

One big disconnect Pan sees is the specificity of the data and the timing in which it’s delivered. Companies, he says, need data to make decisions, but only so much of the mountains of data coming in is useful to the decision-making process. He says companies should understand how to get relevant data from such analytics.

“If you’re making decisions on a current trend, data from last year is going to be irrelevant. A lot of businesses, even when they’re trying to make timely, fast decisions, are relying on data that’s months, even years old,” Pan says.

Scott Horowitz, senior director at FICO, says there’s a wealth of data companies have that they might not know they have or not be utilizing most efficiently. Also, companies are adding new rules or strategies, yet not removing those outdated methods that aren’t working.

Companies, he says, are simply layering on procedures for underwriting, pricing, or managing claims. “They’re doing things they put in place ten years ago and adding a new thing they put in five years ago and then adding the thing they put in yesterday.” He recommends companies:

 Take a data inventory: Get a sense in a more documented manner of what data exists. Horowitz says insurers tend to have siloed data in different parts of the organization. By identifying the information, insurers can then address how best to share information across all business areas.

 Locate information that is being gathered but not stored: Horowitz suggests looking for information in those hard-to-access formats, such as written notes, conversations, text documents saved by one person, and other documents stored by with no efficient way to access them.

 Look at how data is being captured but not systematically: Start gathering it in a more systematic way. “Call center information may be in employees’ heads, but not gathered anywhere. Maybe that’s because there’s no way for employees to document it,” says Horowitz.

 Work with companies that have done it before: Horowitz cautions that embarking upon data gathering and dissemination singlehandedly could be difficult. “You have only the experience of your company and there may be things other companies are doing you can take advantage of.”

Pan suggests that companies looking to instill a metrics-driven culture should focus first on the people.

“You need to get more people involved to understand the data and manipulate it, and you need to have the tools to help them do that,” he says. However, how the tools are used can also be a missed opportunity.

Predictive analysis tools, Pan says, were designed to make data dissemination easier. But they were quickly misused. Especially at the enterprise level, the solutions became merely compliance tools. “That’s kind of missing the point of getting all that data.”