Using Your Data So It Really Does Some Good


There's a lot of talk about operational business intelligence, collecting and analyzing data to enable making better decisions, and ultimately boosting efficiency, improving service and cutting costs, all of those goals most companies ostensibly want to achieve. But is that how most or even many companies today use data? Sadly, it seems not.


Research by Jeanne Harris, Robert Morison and Tom Davenport, co-authors along with of "Analytics at Work: Smarter Decisions, Better Results," shows about 40 percent of business decisions aren't based on data. When I interviewed Harris and Morison in April, they told me this was due to several factors, including that managers traditionally rely on their intuition and feel overwhelmed by large amounts of data. And, Morison told me, "Most of those systems go in without much forethought [about] how they're going to use this information for decision-making."


I heard a similar opinion from Richard Snow, Ventana Research's global VP and research director, Customer and Contact Center, when I interviewed him recently about his new report on contact center analytics. He told me "people don't really know how to use" data analytics and said many contact centers could cut their call volumes if they bothered to analyze customer data. He said:

Call avoidance is a big point within the call-center industry. People say, "I've done everything I can to reduce average call time. I need to cut down the number of calls. I've tried sending people to the website, and it didn't work. I tried sending people to the IVR (interactive voice response system), and it didn't work." I always suggest using analytics to find out why people are calling.

He described a consulting gig in which he was hired to helped build a 3,000-seat contact center for a European cable company. The company didn't believe Snow when he shared his opinion that only about 500 seats were needed. Yet data analysis showed more than 70 percent of calls were about problems with the company's billing system. Improving the billing system would obviously cut down on calls, Snow said. He laughed when he concluded the anecdote:

They went, "What should we do?" And I said, "Well, I can sell you a billing system.

A CEO could probably garner similar insights by spending a day in a company contact center, Snow told me. (That's exactly what Microsoft CIO Tony Scott did, and I wrote about his visit to the contact center back in May.) Yet many senior executives simply can't spare the time, and there's always a risk that a single stint in the contact center might not accurately reflect larger trends. Calls might focus on a one-time anomaly like a network outage, for example. Said Snow:

If you get the analytics, especially things like speech analytics, you can get the same kind of information and put it in a PowerPoint chart. You'll find out things like the cause of a peak in activity might have been a marketing e-mail that the contact center wasn't told about. Understanding those things isn't exactly rocket science. You can use the data analysis to show trends. That gets back to the business metrics. If 15 percent of your customers say they don't understand your product guide, then go and rewrite the product guide.

Sounds easier said than done, doesn't it? There's that sense of feeling overwhelmed by big amounts of data that Harris and Morison mentioned in their interview. That's only going to get worse, as the number of channels people use to interact with companies grows, another subject addressed by Snow in his interview. For a so-called 360-degree view of a customer, companies need to aggregate and analyze data not just from contact centers, but from websites, bricks-and-mortar locations and social networks like Facebook and Twitter.


The good news, as Morison told me, is that data analysis is "not an all-or-nothing situation, go with your gut vs. full-blown modeling and simulation. There are a lot of stages in between." Before taking on a huge and costly data initiative, he said companies should "discipline themselves to be a little more attentive to the information they are using."


Companies can begin by looking at relatively small amounts of data in discrete systems and build on those successes to encompass more data, agreed Harris. She said:

In the 80s, it was popular in the IT world to talk about enterprise data modeling. It never happened because businesses are simply too dynamic. It makes sense to be more incremental, to build upon success and build an analytic capability over time at many different levels, not just the technology.