When implemented and utilized properly, business intelligence software can help enterprises save thousands of dollars, retain and win customers, and ultimately make a firm more competitive. For CIOs charged with overseeing such a deployment, however, the task can represent one of the tallest challenges of their career, rife with technological, political and communication-related barriers that must be hurdled. Here are four tales of battle-hardened CIOs who completed their missions -- and emerged with ample advice for those embarking down a similar path.
PHARMASCIENCE: A HEALTHY DOSE OF BI
When David Lavallée came to Pharmascience as the drug developer and manufacturer's CIO in 1997, he immediately knew something was wrong. Unscrewing the cap of the firm's data management infrastructure, he saw a system that was well past its expiry date and lacking in potency. So he set about filling a new data management prescription -- in a hurry.
"I was telling the executive management team a week after I joined that we had to change things immediately," says Lavallée from the firm's headquarters in Montreal. "They had badly implemented an ERP system, and they were doing all their reporting off of it. The first thing I did was inform them of business intelligence and data warehousing trends, and got them almost immediately on the path to those."
Lavallée's confidence in prescribing an updated dose of IT medicine to Pharmascience's operation was born out of experience. He had just left a large pulp and paper operation that had "reams of data to get out to customers," and had turned to BI years earlier to handle the load.
As a manufacturer of generic drugs, Pharmascience is in a commodity business where even a slight edge that one player can attain over its competitors may be extremely valuable.
Lavallée knew that a modern BI system would help the 24-year-old firm better understand its marketplace and help reduce costs throughout its supply chain, and his approach was refreshing to those power data-users who, in his words, were desperately seeking better information and were "queued up, begging on their knees for new reports."
The strategy he implemented was focused on three key areas, the first of which involved taking all the data-crunching processes off the transactional system and putting it on a separate infrastructure. This, he reasoned, would not only improve performance but also decentralize the way the data was accessed and put it in a friendly format so that business analysts could access their own information -- "sort of the ATM approach," he says.
Next, Lavallée went about eliminating Pharmascience's dependence on paper-based reports by empowering the business as a whole to analyze its information and ultimately make better decisions.
Finally, he decided to provide his executive management team with corporate scorecards to help them focus on the exceptions within the data that would require their attention, be they good exceptions or bad. "They had to start doing this, instead of reading through voluminous monthly operating status documents," he said.
Because Pharmascience had an Oracle base already installed, Lavallée naturally implemented a data mart from that vendor and expanded on it for the next few years. By 2003, he had switched the firm to an SAP shop, at which time he brought in a BI suite from Information Builders. He likes the IBI suite because it allows his team to pull together info from multiple software programs and roll it up into the executive scorecards.
The benefits have been numerous, but they have been accompanied by a number of challenges. The quality of the master data with which his team was working was perhaps the biggest. Once unit leaders started seeing the cleansed numbers, they began to notice some "funky things" going on with many of the company's internal processes.
"They started saying 'What is this?', which pointed to the underlying misuse of the ERP system," Lavallée says. "They questioned the numbers, such as average costs and the amount of customer returns, which weren't jiving [with the numbers they already had]. So BI helped us improve our internal processes."
Another hurdle was getting the numbers distributed to more people. Some of those who had the info, and realized the value of it, became a tad possessive.
"They kind of became kings of that information and were sometimes unwilling to share it, because knowledge is power," Lavallée observes. The problem resolved itself when those looking to be added to the data guest list were the company's top senior managers.
"That drives adoption pretty quickly," he says with a laugh.
Lavallée advises those embarking on new BI implementations to "start small and build on your successes...don't tackle too much at once." Perhaps most importantly, though, he warns those ramping up a BI rollout to never underestimate the importance of their master data.
"Until you have [clean data], you are not going to have high-quality information in your whole BI space. And the moment the users start to doubt the quality of the information, well, then you are starting to fight an uphill battle."
GRAND & TOY: FILING AWAY ITS INFORMATION HEADACHES
When office supply giant Grand & Toy really began sinking its teeth into business intelligence, the problem for then-CIO John Melodysta wasn't getting buy-in from those who would be using it. The headaches arose when the users actually got their hands on it.
"Acceptance of this stuff is easy," says Melodysta, who has since moved up the corporate ladder and now sits in the Chief Administrative Officer's chair. Since his team built a tool for sales people to view their performance on a daily basis, "there isn't a rep around who doesn't get up in the morning, log on from home, and see how they did against their sales targets of the day before."
One of the main challenges Melodysta faced, and still comes up against, was arriving at a common language around the data that the BI system was spewing. Without it, he says, confusion can reign supreme.
"There are a lot of people who want to analyze sales, for instance, but what does sales mean? Is it invoice sales at the ship-to level, pretax sales or after-tax sales? You have to get your whole company thinking in one way about what sales are."
The data, served up through software from Clarity Systems, can be interpreted many ways, Melodysta adds. "The sales guys have an interest in one venue and the finance guys have a different view of the same data. It's hard to get people away from the old way of thinking."
The role of the IT department, then, is often to play interpreter to the different requests from all departments who want to use the BI system to their own advantage.
"It's hard for an IT department to get one set of directions," he says. "You might have six different business units and they all have different levels of understanding, different interpretations, they can or cannot articulate what they want. It just drives you nuts."
To best solve the problems this traffic jam of requests creates, Melodysta suggests creating a group dedicated to BI that can work specifically with the other departments.
"If they are smart enough to interpret the needs and wants of the business, then that is the ideal situation."
Despite the challenges, the payoffs that have accompanied the implementing of BI at Grand & Toy have been numerous, primarily in the area of cost savings. "If we have to send a truck once with a large order to a customer rather than multiple times with smaller orders, that's good for us and the customer because we can keep our costs and our prices down."
Getting deep into the data to see what's really going on around the most common business practices has also been a difference-maker. Sales analysis, for example, can be done on as granular a basis as products' SKU codes, Melodysta says. Often, this info is taken back to the customer to help them with future orders.
"We can do an intelligent review with [them] and help them save, such as in the area of returns, which costs them and us money. We create all of that using BI."
LONDON POLICE: PUTTING COPS BACK ON THE BEAT
In the old Barney Miller sitcom, viewers never saw the cast of policemen out on the street making arrests and chasing the bad guys. Instead, all the action took place within the walls of the Twelfth Precinct, where most of the cops' work, it seemed, consisted of filing reports from behind their typewriters.
For the police force of London, Ont. a few years ago, it was starting to seem that many of the patrol officers' efforts were similarly being eaten up by paperwork -- and to the force's CIO, Eldon Amoroso, there was nothing funny about it.
"Our average call time went up significantly, from 105 minutes...to 119," says Amoroso. "We pushed the panic button and decided to do something about it. But first of all, you have to understand what's causing it."
Enter business intelligence. Using a set of proprietary tools, Amoroso and the BI committee he formed asked themselves what factors were causing the grief. Analyzing "the data that had been sitting there all the time," they realized that four common offenses were taking longer to process than they should: drinking and driving, arrests, family violence calls and mental health issues.
In most of these cases, Amoroso discovered, officers were drowning in the unnecessary paperwork that had grown up around the task of processing them. With the 10 most common charges, then, the team streamlined the reports to include only the most pertinent information that would help the Crown in its efforts to prosecute the case.
"So for a drinking and driving charge, instead of keying 14 pages of information in, it would be a two-page template with information that the Crown attorney would really need."
The result? "Our patrol time dropped back to where it was," Amoroso says. "There was a saving of 25 officers' time, which was just huge for us." Another area where the administration demons were taking over was in crime analysis. "Analysts were doing a lot of clerical work and we missed trends that were emerging," says Amoroso. "A person can't look everywhere, and you can miss things that are emerging."
In one instance, Amoroso received a call from a local politician who was hearing complaints from many of his constituents about the sudden spike in car break-ins in the area. "Sure enough, there was a problem there that we missed," says the 27-year veteran of the force's IT department. He again turned to BI to help reduce the paperwork overhead. Using police trend analyzing software from Ottawa-based Versaterm Inc., the London Police began letting "the computer do the time-consuming clerical stuff," Amoroso says.
While the BI software has certainly come off looking as cool and capable as Dirty Harry throughout the process, Amoroso cautions that it would never have been able to show its stuff had the business side of the force not seen the value in it.
"This process has to be business-driven, about finding the weak spots in our business and allowing our people to work smarter....But the business has to know they have a problem. If that side of the organization isn't watching, BI probably won't succeed."
Getting that buy-in wasn't always automatic. "Police by nature are probably fairly conservative people, but most people, if you show them good results...they don't care what software you're running or what's behind it. They will realize the benefits," Amoroso says
As the London Police force continues to expand its use of BI, lines of communication have to be constantly maintained between IT and the officers to make sure no disconnects arise. To that end, Amoroso uses a daily management meeting to keep all concerned parties up to speed on what's happening with the data and what to do with all of it. A weekly e-mail sent to the senior people serves to reinforce whatever items may have come up throughout the week.
"They have to see the results," Amoroso says. "Communication is really important."
What the BI process really boils down to for Amoroso, though, is that police officers now have more time to work in the place where they are most effective: on the streets.
"You want the officers to have a certain amount of time where they are literally just driving around. That's where they can get out of the car and walk down the street, and respond to all those things they see happening."
CHALLENGER MOTOR FREIGHT: ON THE ROAD TO SMOOTHER OPERATIONS
Challenger Motor Freight is in the business of connecting people and corporations with their goods, from Montreal to Mexico City. Five years ago, however, the technology infrastructure that was supposed to make that process smoother was causing nothing but disconnections among the firm's business leaders.
According to Director of Information Technology Eveline Gaede, each department would be bringing different numbers to the table, causing considerable confusion and frustration. When it came time to figure out how many trucks were available for a certain day's run, for instance, each business leader had their own number, and none of them were the same.
"The Operations department sees 15 trucks, but Maintenance knows some are not available, so they see 12. And Driver Services knows there are only 10 drivers there, so they see only 10 trucks available. So what really is the true number?" Gaede says. "It was at that point that I realized we should do something."
What she did was implement a business intelligence system from Microsoft to help make sense of the robust data that was populating Challenger's servers. Through a satellite tracking system, figures were being collected on where a driver was located. Detailed elements such as fuel consumption and tire pressures could also be monitored by the technology in place.
Initially, Gaede says, the BI rollout was done cautiously, with only five people in the executive management group using it for the first two years. Eventually, however, more of the front-line staff began clamoring for the ability to use the data.
"So in the past year it's just gone crazy," says Gaede. "It's very easy once you have the whole data warehouse built. If someone wants [a particular set of data], the turnaround time to provide them with a cube to analyze is maybe a day."
While that speed on the inside of the organization is listed by Gaede as one of the top benefits of the BI implementation, another has been a more outward-facing one -- namely, the ability to provide certain high-end clients with customized reports.
"We have a lot of customers who will call us for reports rather than using their own system," she says.
The next step in the coming year will be to inject more predictive qualities into the process. Gaede and her team are looking to develop driver scorecards which will rate each individual's performance and hopefully prevent accidents.
The chore of getting all members of the organization who should be using the BI capabilities up to speed is an ongoing one for Gaede.
"We still find it difficult for some people to grasp it. The financial group jumped right on it, but it's a different way of looking at the data, and a lot of people are still used to looking at just paper reports and they kind of get lost."
For those entering into the BI world in their organizations, Gaede has a few hard-won lessons to pass along. First, she says, consider what you want to get out of the BI system before you jump in, and "really think through how you want the data laid out. I started by just building cubes, and I've had to rebuild them over the years because it wasn't the best way to do it."
Also, Gaede advises, don't underestimate the importance of getting your users trained on the software. "We would have got a lot better adoption if we had spent more time training and working with the users."
She advises going beyond a mere cursory introduction session and letting them get their hands dirty with the BI tools. "I can set up the views for the user, give them an hour training session, and after they've played with it for a bit I'll have another session."