How to Improve Business Intelligence Software ROI

Monday Nov 1st 2010 by Jeff Vance
Share:

IT loves to blame vendors when business intelligence projects fail to deliver on lofty goals, but what factors are really to blame?

Business Intelligence (BI) software promises to deliver fast ROI and an array of corporate insights, but the promises often fall short of the reality. IT loves to blame vendors, but how many projects failed because they were underfunded, abandoned halfway through the process or ignored by the rest of the organization?

“One of the problems is that BI comes in a lot of different forms and shapes. Users scratch their heads and ask, ‘what the heck is it, exactly?’” said Anandan Jayaraman (A.J.), Chief Product and Marketing Officer for Connectiva Systems.

Connectiva Systems, a provider of subscriber data monetization solutions, believes that one of the main obstacles to BI ROI is that people look at it backwards. The typical BI project launches when someone within the organization realizes that they have a ton of valuable data, but it’s lost in a sea of other information in data warehousing systems.

The next logical step, then, is to unlock that data, right?

Not according to Jayaraman. “It’s a mistake to start with the data first. The first question to ask before any project, not just a BI one, is: what outcome do you want?”

This is not uncommon advice, but few seem to be following it. Jayaraman gave an example of a broadband promotion that failed because outcomes were not clearly defined.

A Canadian telecom company wanting to boost its new subscriber numbers launched a marketing campaign that offered new users 60 months of free service. As the campaign progressed, the company found that large numbers of new subscribers were canceling during the free trial.

The free trial hook had an outcome that was opposite of what the company wanted. Instead of luring in new customers who would become loyal, long-term ones, they attracted a slew of broadband mooches who bailed before they’d paid a single penny.

As the telecom company drilled down into demographic details, they figured out where they had gone wrong. They had assumed that one particular demographic segment would be swayed by 60 days of free broadband, so they marketed heavily in their direction.

That segment turned out to be the campaign’s Achilles’ heel. So, who were these people jumping ship on day 59? College students. Cash-strapped students love getting things for free and would trial hop indefinitely if they could.

With this information in hand, the telecom company shifted gears, stopped marketing the promotion to students and turned their numbers around.

“Too many companies believe that all they have to do is get the customer to sign up,” Jayaraman said. “If that’s all you want, often that is all you will get. A better approach would be to figure out how to convert customer-related data and the insights you clean from that data into cash.

“Once you know what appeals to your customers, what they respond to and what drives them away, you can turn that information into marketing campaigns, cross-selling strategies and up-selling offers – all in a way that has real value for both you and your customers.”

Most companies aren’t even measuring BI ROI

False assumptions have doomed many a project, but they can also prevent successes from being even more so. To know that you are achieving ROI, you have to measure it, obviously. But according to David White, Senior Research Analyst with Aberdeen Group’s BI Practice, “Most companies aren’t even calculating return. They’re just assuming they have one.”

Aberdeen Group studies the practices of BI adopters and places them into three categories: best in class, middle of the pack, and laggards. As you would expect, best-in-class adopters are much better at delivering BI projects on time and on budget.

One of the key behaviors they exhibit is that they take the time to build cost projections into the deployment process and develop early warning signals for when costs threaten to spiral out of control.

Conversely, laggard organizations fail to set concrete budgets, establish clear goals and, most importantly, fail to get the buy-in from CEOs for the projects.

Another barrier to the kind of high ROI that vendors hype and adopters hope for is lack of coordination across the organization. If one part of the organization has a solid BI plan in place, it would make sense to repurpose that plan for the rest of the organization. Many intend to do just that, but few actually follow through.

Worse, this lack of organizational coordination can mean paying for the same thing over and over again. One of my sources, who preferred not to be quoted on this particular point, found examples of companies paying for software licenses that they already had on the shelves.

The BI buyers simply didn’t know what was available to them within their own companies, so they bought the same things over and over again.

Turner Broadcasting saves time and money by leveraging what they already own

Turner Broadcasting avoided falling into the buy-it-twice trap simply because an executive, before buying a new BI solution, paused and thought, “you know what, I bet we already have something in place.”

That executive, Robert Copenhaver, Vice President, Finance, was struggling to calculate year-end bonuses for Turner Broadcasting employees. As with so many BI projects, one of the main problems Copenhaver sought to address was the lack of visibility into the process.

“We have operational groups all over the world, and coordinating everyone in a coherent bonus program is trying, to say the least,” Copenhaver said. Copenhaver would have to deal with “dueling spreadsheets,” indecipherable faxes, questionable claims that had little documentation to back them up and “mounds of information that didn’t match up.”

To fix this mess, Copenhaver began looking at some of the business intelligence software Turner Broadcasting had already deployed. “The math isn’t that difficult, and we have many business tools at our disposal. I knew we should have something that we could use for bonuses,” he said.

The company had previously deployed BI software from Arcplan to improve financial reporting. With a few modifications, the software was able to deliver actionable information about the metrics used to calculate bonuses.

By centralizing the information fed into bonus calculations and forcing employees to have concrete data to back up their sometimes exaggerated claims, Copenhaver managed to shave at least a couple of weeks off of the bonus-calculation process.

That is serious ROI. Not as concrete but just as valuable, Copenhaver no longer wakes up in a cold sweat at night when bonus time approaches. Instead, he has the luxury of a streamlined, straightforward, fact-based bonus-calculation process.

Knowing where to start seems simple but isn’t

Another common behavior that harpoons BI projects is trying to do everything at once. As companies begin digging into their data, they find all sorts of potential opportunities. If they let their optimism run wild, they risk trying to do too much, too soon and often end up failing at everything along the way.

Ohio-based wireless provider Revol Wireless started their BI efforts with a simple goal: figuring out which of their retail stores were their top performers.

“We had a lot of assumptions about performance in each of our sales channels, but we didn’t have many facts,” said George K. Mehok, CIO of Revol Wireless. “The wireless market is hyper-competitive. If you don’t know what your customers are doing, and what your competitors are doing, you won’t keep up. You have to be able to pinpoint your opportunities and root out your weaknesses in order to stay competitive.”

Revol Wireless deployed the open source BI suite from Jaspersoft and began measuring the variables that determine retail success. “After deployment, we developed a sophisticated retail ranking system. Our rankings go beyond revenue and sales to such measures as foot traffic, sales conversion and customer churn,” Mehok said.

Based on this information, Revol Wireless launched a large-scale retail redesign effort, using BI information to determine which stores to keep, which to close, where they should expand and where they needed to modernize.

“Sentiment analysis” and the future of BI

While the main hurdles to BI success tend to be cultural and behavioral issues, once the value of data is unlocked, organizations can push forward into far more profitable analytic efforts.

For instance, call center representatives are exposed to plenty of information during calls that – until now – has never been translated into data. If a caller is angry, irrational or simply frustrated, that emotional information means something. It can help companies deliver better customer service, reduce churn and retain high-value customers.

When tied to actionable back-end data, this “sentiment analysis” can be extremely valuable. A.J. Jayaraman from Connectiva Systems noted that with the proper data capture in place, companies will know ahead of time when they risk losing high-value customers. The company can leverage that information to set up systematic, proactive retention processes that make their most valuable customers happy once again.

“Often, it’s as simple as having a manager follow up with the customer in a timely fashion,” Jayaraman said. “If a customer service manager takes the time to call you and tell you exactly what is being done to resolve your problem, you are less likely to defect to a competitor.”

On the flip side, low-value customers, the ones who, say, think their mobile provider should replace the phone they just dropped in the toilet, may find themselves spending a long time on hold waiting for that elusive “next available operator.”

Summary: 5 ways to achieve BI ROI

1. Focus first on desired outcomes, not raw data.

2. Establish clear budget parameters and set up an early warning system for cost overruns.

3. Get senior-level, non-IT (CEO, COO) buy-in for BI projects.

4. Start with something simple and easy to monetize, such as retail store rankings.

5. Then, go beyond the easy stuff and embrace BI for cutting-edge analytics like “sentiment analysis” and “subscriber data monetization.”

Share:
Home
Mobile Site | Full Site
Copyright 2017 © QuinStreet Inc. All Rights Reserved