What’s a guy who describes himself as a ‘classic New York street fighter’ doing in the executive boardroom of Rogers Communications Inc? The short answer would be: helping Ted Rogers ‘fly the corporate plane’. The long answer is a little more complicated, but well worth investigating.
In June of 2002, Bob Terdeman was something of a burned out case. As Chief Data Warehousing Architect for EMC Corp, the native New Yorker had just finished logging a grueling 300,000 air miles in nine months, sharing his extensive knowledge of databases and information architecture. So when a small New York-based division of Rogers came knocking on his door, he grabbed the chance put a little stability back into his life and signed on as Chief Technology Officer of Rogers Medical Intelligence Solutions, responsible for rebuilding the firm’s technology infrastructure.
A scant six months later, the rebuild was done, pretty much in record time for a job of this complexity at a Rogers company. And the folks at head office took notice. They began calling him up to Toronto – for a day here, a couple of days there, and eventually for a day every week. And still it wasn’t enough. Finally, the man in the corner office decided he wanted him there three times a week – and as Terdeman himself noted, Ted Rogers is not a man you say no to.
Rogers vision was to build a world-class customercentric organization. This organization would leverage the expertise and leadership of the individual operating companies while providing enhanced capability back to the operating companies. The strategy covers every aspect of the interaction with the customer, but at the nucleus is a new ability to understand the customer, plus a dashboard to “fly the plane”.
“Ted hates flying blind,” said Terdeman. “He understands that he needs the control panel to fly the plane. And he sees business intelligence as the tool to provide information to the control panel.”
The vision not only includes business intelligence but new technologies in customer service, call centres, and new business processes which integrate the billing processes between the operating companies and the customer care (call centre) platform. These have already yielded an improvement in customer service.
In order to realize his vision, Rogers needed the right point man in the technical architecture and business intelligence arena – someone with superlative BI skills and enough steel in his backbone to lead the company through the difficult cultural change that would come with such an undertaking. Terdeman – coauthor of several seminal books on data warehousing – was a standout choice on both counts.
“I’m not a gray guy. I’m not a ‘go along’ guy,” he said. “If I’m doing a fairly good job, some people will be upset. If I’m doing a really good job, everybody will be upset.”
Naturally, some discomfort was bound to occur by using his blunt approach. So when he took the job of Chief Information Architect in September 2003, he warned CIO Ronan McGrath that he had a strong personality and would no doubt cause some disruption.
Unfazed, McGrath smiled back at him and said, “Sometimes a little disruption is good.”
Understanding the customer
One of the key corporate imperatives at Rogers is customer satisfaction. As Terdeman says, “We want to be a trusted partner to the consumer, not ‘those damned people at Rogers!'”
For a company to satisfy its customers, it has to understand them. That’s where business intelligence comes in.
“Clearly, the customer is king in our business,” said McGrath. “Our objective is to be able to offer the customer a very broad range of products, either individually or in combinations. To do this we need to understand more about the customer base – who the likely buyer of those products is, who the different influencers in families are, and so on.”
Added Terdeman, “The world we’re going into is going to be far more complex, so it’s becoming more important for companies to tailor or individualize their offerings. A customer is not going to sit and choose from a menu of options. He expects you to understand what his need is, and that’s really the challenge.”
To this end, Rogers has implemented what it calls its ‘global customer data mart’, an enterprise data warehouse, spanning all of the business units, that gathers data on the customer and congeals it – logically connects it – around the customer and his family. The data includes such things as what customers have bought and how their buying patterns have changed. With the help of BI tools, the data can help Rogers understand a range of influences that may affect buying patterns, such as geography, socio-economic status, lifestyle, and community affiliation.
The BI tools Rogers uses with the data mart are from SAS. As Terdeman noted, “When you try to do something of this nature, you want to pick the technology that could be used by the most people as quickly as possible, with the most effect. And SAS is the most common product you find in graduate schools. Most students who study statistics know how to use it. So it’s a way of raising the bar without raising the training costs.”
Making the company transparent
One of the goals of Rogers’ BI initiatives is to make the company transparent to its customers.
“The customer only knows the word Rogers. He doesn’t know about Rogers Video, Rogers Cable, Rogers Wireless, Rogers Media,” said Terdeman. “When the customer contacts Rogers, he expects us to know who he is. And if we upset him, we upset him across the board. So in the world we’re now in, the products are replaceable, the customers are not. The customers are forever.”
According to McGrath, one of the company’s objectives is to be able to handle all customer issues with the minimum possible number of steps.
“Whether the customer contacts us through the Web, by phone, by fax, or by email, our goal is to keep a proper and complete understanding of that customer transaction so that we get it right the first time, or, if we don’t, we learn from any mistakes so that we get it right the next time,” said McGrath.
Another key issue is customer retention. Acquiring customers is largely a marketing effort, but retaining them is a combination of marketing and service. Doing the best possible job in this area again comes back to having the right customer knowledge.
“We need to understand who our best customers are, and do more premium things for them,” said McGrath. “The development of customer intelligence is really a way for us to improve levels of service and to improve the targeting of our marketing efforts.”
Towards real-time marketing
Beyond the ability to better target its marketing efforts, Rogers’ BI capability will also shorten the time it takes to put its marketing solutions into play by an order of magnitude. Typically, it takes weeks or months to engineer a campaign. In the world Rogers is moving to, those weeks and months will be reduced to hours and days.
According to Terdeman, a typical disastrous pattern in data warehousing is for the system to be designed incorrectly at the outset. Initially it is planned as quarterly; then the push is on to make it monthly; and when it’s monthly, the plan becomes to make it weekly. With each change, there is a significant investment of time and money in re-engineering.
To avoid this trap, Rogers planned at the outset for its system to be daily. And the company made a substantial up-front investment in the infrastructure that would be required. Now Rogers is able to set its sights on virtually real-time marketing.
Empowering the users
Following the old adage “Give a man a fish and feed him for a day. Teach a man to fish and feed him for a lifetime”, Rogers is putting the power of its BI tools directly into the hands of its end users.
“A classic mistake made in this world is having the IT people run the queries for the end users,” said Terdeman. “It’s a partnership. I’ll build you the house, I’ll get the data there. But I don’t know what’s in your mind. You need to use the tool.”
Understanding this, Rogers is moving its IT people out from between the end users and the data. Users must run their own queries, and learn how to use the tools.
“This is a radical change for Rogers,” admitted Terdeman. “Younger people that have come from outside of the company totally embrace this. The older people are struggling with it. ‘You mean you want me to get my hands dirty?’ You bet, pal. That’s the challenge. And by the way, this also creates a level of discomfort for IT people. Tough. Get over it.”
Cleaning up the data
In any implementation of this nature, legacy data is probably the most complex issue, according to McGrath.
“We’ve removed a significant number of legacy systems already. We’re now down to two major billing systems and one call-centre infrastructure,” he said. “The process we’re dealing with now is to move from legacy data into enterprise data, and that’s a lot of effort – understanding where data is, understanding where the deficiencies in data are, and building remedial plans to deal with it.”
To make sure the warehouse is populated with the right data, Rogers has put a federated data policy in place. While the various operating companies may acquire the data, it is the property of parent Rogers Communications Inc. (RCI). The operating companies have to certify to RCI that the data is accurate and clean. And this is a stepping stone to data governance.
“Now we can put the right people together as a data governance committee and we can coach and counsel them as to measures that ensure the cleanliness of the data,” explained Terdeman. “This puts us on the road to cleaning up the data, which is a major issue, and which, typically, the businesses will not do. The businesses will only acquire the data they absolutely think they need, and not one bit more, because they don’t see the strategic value of it. They only look at the tactical value – what made the sale.”
The result? When the businesses discover they need it, it’s too late to collect it.
Moving forward
With the implementation of its first real enterprise data warehouse, Rogers is finding opportunity it didn’t see before – overlaps between the publishing and the wireless businesses that weren’t apparent, for example.
Right now the warehouse is being hand run, but in the next two months the plan is to turn it into a fully operational system, running on its own – not a trivial task, considering the amounts of data being collected. The company’s second key task is to educate the various users to the value of that data and how to leverage it. Once that happens the benefits should really start to flow.
So will Ted Rogers finally have the control panel he’s been looking for? Terdeman certainly thinks he will.
“I believe that the tool we’re building here will rival any in the world,” he concluded, perhaps a little immodestly.
But hey, he’s a New Yorker. We should cut him some slack.
David Carey is a veteran journalist specializing in information technology and IT management. Based in Toronto, he is editor of CIO Canada.
Key issues in business intelligence
There are many issues that can make or break a business intelligence project. Here are a few key ones identified by Bob Terdeman, Chief Information Architect of Rogers Communications Inc.
1. Bring in the right advisors
There are individual strategic business intelligence advisors out there that are world class. Bring them in and pay them for a day of their time. It’s the best money you’ll invest and it’s the cheapest fee you’ll ever pay. Get them in and let them tell you what they can do. It’s not a six-month engagement with a consulting company for a million dollars; it’s a two- or three-day exercise in looking at your data, talking to your executives, and getting a quick and dirty overview of what can be done. Then tighten your seatbelt for the ride. Hire the best people you can. Fund it. And then stay out of the way.
2. Find a champion
There’s a commitment that companies need to make in order to do this, and often that commitment is lacking. This is a business and not an IT initiative, so you must have an executive sponsor that is not the IT person. The executive sponsor doesn’t get involved in the day-to-day activities, but when things get to a confrontational level – which they will – the champion is the one who says, “This is what we’re doing; this is what I want to get done.” The champion must be willing to cross over the business lines and carry the faith. He is like a proselytizer or a preacher. A lot of them wind up burned at the stake.
3. Hire the right point person
When creating a BI system, you have to know how to overcome obstacles. Some processes work and some processes you have to go around – change the rules. Not many people constitutionally can do that, so you have to recruit the right person to lead the initiative. You only need one guy like me in an organization. Two are dangerous. But you need somebody who is willing to risk of the ire of his peers to get the job done. And you need someone to be personally accountable – who’s willing to put his head on the plate.
4. Don’t underfund it
Understand that you’re making a strategic investment in the future of the company. Don’t underfund it. This is not something that you can do on the cheap. You must do it right the first time – bring in the best and let it go.
5. Don’t underestimate system needs
When building a large-scale BI system, many companies underestimate the capacity required for their computer platform. In the transactional world, reserve capacity on the platform is usually in the range of 10 to 20 percent. But in the BI world, the load is not predictable. Bad performance in data warehousing is a good thing. It means that the users are using the system. You want the end users to drive the system. You want them extracting every bit of business value – which means that you cannot predict the load. If you’re predicting the load, you’ve done the wrong design, because what you’re building is a tool to drive more business value, and what you want is to let the business move as hard and as fast as they can, because every pearl is worth a lot of money. In order to do this, you have to have enough reserve horsepower to let it rip. Therefore, the rule of thumb is 4X the predicted power – let the users suck up that extra power. That’s extremely hard to understand in a cost-conscious company.
6. Embed the knowledge
Some companies make the mistake of relying on consulting companies. You must embed the knowledge into your own organization. One of the first things I did when I came to work for Rogers in Toronto was to inspect my consulting partners. I found out who could do the hands-on work and then I fired everyone else. I’m a big believer in hands-on knowledge, not hands-off management.