Data Warehousing Dialectic

Roundtable Participants:

    Tom Atkins (facilitator), President, TWA ConsultingKelvin Cantafio, CIO, Foster Parents PlanMichael Cuddy, CIO, Toromont IndustriesFariba Rawhani, CIO, VP, CIBC MortgagesNorm Rosner, Senior Director, IT, The Loyalty GroupSteve Seki, Director, eCommerce Development, Equinox Financial GroupJohn Shoesmith, Research Manager, Software, IDC CanadaDwight Willett, VP, Information Services, Enbridge Consumers Gas


ATKINS:What kind of solutions, or what kind of business problems, have you deployed your data warehouse to deal with?

CUDDY: Toromont’s business is heavy equipment for construction. We’re very much tied to the general health of the economy, and we want to be very well prepared if our business cycles downward. So having the information current as of this morning – versus having it six or eight weeks old – is very important. A lot of our information used to be on paper and was only available once a month, without much detail. Now you can go in daily and see, for example, what machines are out on rent.

All of our EIS is built under a Web browser, so it’s very easy for anyone, including the executive group, to grab the information and drill down. And there’s no question, it creates discomfort in certain parts of our organization. It opens up the financial and accounting area tremendously, and the questions become much tougher. It’s getting to the point there’s no place to hide.

WILLETT: At Enbridge Consumers Gas, we could drown pretty quickly in customer data. We’ve got 1.4 million customers, most of whom have gas meters that spit out data that we need to prepare bills every month. We’ve stored huge amounts of that data in old technologies. It would really be useful for us to be able, for example, to provide our broker customers some consumption data around geographical areas, such as ‘what’s the gas usage pattern for people in south Toronto?’ This would be useful information to sell. But even though it’s sitting there we don’t yet have the capability to pull it out by any meaningful sort of segregation. Now we’re beginning to say: let’s start to build some interfaces that would at least let us pull out some subsets of the data and establish, in effect, smaller warehouses in which it’s easier to find things. Let’s learn something in a kind of data mart, and then we’ll move forward into larger warehouse-type applications. We have a couple of pilot projects under way that are doing that.

SEKI: The data warehouse has two main uses at Equinox Financial Group: one is to provide information to help management make key business decisions. The other is on the e-commerce side, where we have a strategy of self-served information. We need to communicate out to our agents and we want to offer them a self-serve approach, rather than having them calling in to support groups to get information. They can tap into the warehouse and get appropriate authorized data, and that minimizes the infrastructure we need to support them.

CANTAFIO: Foster Parents Plan does work in the developing world and our primary objective is to get as much money into the field as possible. Currently, over 83 per cent of all the money we raise goes into our programs for overseas children. So for us, a really important side of data warehousing is to help us with our operational effectiveness. It’s become a driver for our re-engineering and it’s extended the life of our legacy system by pulling out information and allowing us to look at it in a different way. It’s even allowed us to model, in some ways, how our business should be different.

ROSNER: In The Loyalty Group’s case, we have some pretty unique uses for our data by virtue of the fact that we are a direct marketing company. Our tagline is “Rewards – Results – Knowledge”. The Rewards part is obvious, with the Air Miles program. For Results and Knowledge, we use the database. Key to all of our operations is, firstly, to prove to our sponsors that the program is working and, secondly, to learn more about our collectors on behalf of our sponsors. We don’t have to sell the value of the information across the company – it’s fundamental to what we do. One of the challenges that we have is that we aren’t using the warehouse to get summary, aggregate information the way that many people use it. We try to use it to actually learn more about our individual customers and the individual collectors. So we’ve got some technological challenges to face that perhaps some don’t.


ATKINS:I’d like to know the kind of involvement you’re getting from the business – where the drivers are. Are you pushing data warehousing out or are you getting demands from the business for data warehousing?

RAWHANI: At CIBC Mortgages, we provided the business with information as to the value of the data warehouse. It went from them not being on board to the point where now they really want it. Once they understood it, the feeling was “how have we done business without this?” I think sometimes IT assumes that the business knows about a particular technology, and we assume they should get on this band wagon because the benefits are obvious. But education has to take place, and then it becomes a user initiative.

It goes from the CIO telling you there’s great value in this, to the business really wanting this because it’s losing competitive advantage by not having it.

CUDDY: That’s hard, though, because by saying, “I need better information, I need better tools,” the marketing VP or the sales VP may imply they don’t know enough about what they’re looking for. Some people will be comfortable saying, “I don’t have what I need to do my job effectively. Can you help me get it?”, while others will be very reluctant to do so. Not everybody wants to say they don’t know enough.

SEKI: It’s definitely the business that’s driving data warehousing at Equinox. But I think if you used such words as data marts, data warehousing and data mining, the business wouldn’t have any idea what you’re talking about. It’s a really good situation for us in IT because it’s a classic model where the business is really just identifying a business need – we want to consolidate this, we want consolidate that – and it’s really obvious from a technical perspective how to deliver that solution. When we come to the table and say, “This is how we can address your need, and in addition here are some things that you may not have thought about,” you can see their eyes open.

CANTAFIO: We were fortunate in that we got buy-in right from the board level, and it’s been driven from the top. Certain factors are at play continually. It’s a question of asking the business where they want to go, and then doing a lot of hard work to get the project implemented. The first time people see the numbers there can be a lot of skepticism in terms of giving up the old ways and looking at the new ways. In our case, we put together a cross-functional group very early on and a lot of decisions on priorities were done cross-functionally. Each representative from each area of business was basically looking at one of the data marts. Sometimes it gets very sticky if you don’t have agreement on how you define things, so our focus was on metrics right from the start. We called ours the Metrics Data Mart and said, “What do you need to measure your business to increase your effectiveness?”


ATKINS:From the time that you realize you have the need for data warehousing, or you have a business need for new information systems that data warehousing supports, how long is the decision cycle?

SHOESMITH: As a research manager for IDC, I’ve interviewed organizations where the decision-making process has taken a matter of two to three weeks – where IT will develop a small pilot project, show it to the executive board, and there’s almost immediate buy-in. And I’ve talked to organizations where it’s a struggle, where they can’t quite see the business case at first. Every company is different.

WILLETT: Typically, if we look at something that is a business opportunity, we would take anywhere from six months to a year to get something from the gleam in someone’s eye to a business case, and through approval at the executive level. In the case of data warehousing, we’ve never made it that whole year. I think one of the reasons that a data warehouse in it’s full-blown magnificence has never gotten off the ground is because when we find ourselves talking about a generic solution, like a data mart or a data warehouse, our business knowledgeable people who buy into it say they want product X. And we find ourselves going from pushing something to defending the position that we shouldn’t choose the solution yet. And that really muddies the decision-making process.

RAWHANI: In mortgage banking, you have to make quick decisions. That goes for everything – from spending millions of dollars buying hardware, to launching a new product. Large projects can get approved very quickly – any executive can bring it to the project steering committee for approval. Then the challenge is what must be taken off the table if we proceed, or can we do it in addition to what’s on our plate? The data warehouse was approved very quickly. It was like, we took it to a meeting and within a month it was approved. But keeping it alive has been hard. Most of our investments have been on the operational side, positioning the company to deliver products that other institutions can’t deliver. Projects like the data warehouse are easy to cancel, and right now we are trying to keep the momentum going, and keep everyone excited about the initiative.


ATKINS:How do you deal with vendor selection at Loyalty Group? Does the business poke their nose in to give you a little help?

ROSNER: We’re a business that attracts marketing people and users that are very technologically savvy, so we definitely have to deal with that issue. Our solution is to build very tight partnerships with those end-users in the actual projects, to make sure that we’ve established very clear objectives for every initiative, both short and long term, and then make sure those approaches are consistent. Users have great interest in some technology areas but in others they have difficulty understanding the relevance to the business, and you have to drag them into the decision process. The real trick is to make sure that everybody is working off the same page.

RAWHANI: If users feel that a solution can take a year to implement internally, but they want it in two months, they become great purchasers. They go and see what their colleagues are using and they bring it to the table. In the past, they would actually go buy the solution, bring it in, and then drop it in IT’s lap to support. This would increase our infrastructure costs and IT was always on the defensive. We’ve changed those dynamics – now it’s a partnership. Today our clients consult IT before they buy, and we look at whatever they’re bringing to the table and evaluate it against the overall IT blueprint, which the executive team has collectively approved. The blueprint’s purpose is to get us to our destination while managing our costs. So as colleagues we consult on such things as: How does it fit the infrastructure? Can we support it? How does it fit with the other active initiatives?

SEKI: In the last three or four years, the tide has really been changing. For instance, our U.S. parent company used to launch huge projects just to have a consistent office, so we could support the same spreadsheet and word-processing packages. At this point, we’ve come to realize that the key is really the data. When we select products now, we realize that these are almost throw-away solutions, and they port so well it doesn’t matter what vendor’s products you’re using. So we focus on the data, and there are multiple technologies to get at it. If we have to bring a technology in, we’ll just pick a best-of-breed product.

SHOESMITH: At a hockey game recently, I ran into a VP of a vendor that does a lot of data warehousing. They had a box, and they’d invited some key customers that they wanted to sell their solutions to. And from what I gathered, they had invited the business executives, not the IT department. So I think vendors are beginning to recognize that it’s really the end-users that are starting to make some of these decisions. And because of that, they’re starting to woo the end-user community, rather than the people in the IT department. I think you’re going to see, over the next couple of years, a lot of the vendor community really attaching themselves to the executives, because they’re making key decisions now, especially around data warehousing.

RAWHANI: I think it’s a no-win situation for the vendor if they bypass IT. Ultimately, this is the group they have to deal with. Even when the business falls in love with what they have to offer, IT has to implement it, support it, and maintain the relationship. It’s not a good business decision for them to go around IT.


ATKINS:What about standards versus proprietary products in the area of data warehousing?

ROSNER: We’ve experienced both sides of the coin. When we built our last generation three years ago, we chose a proprietary platform as a throw-away for a few years, simply because we found that in the open technologies the performance wasn’t there for the amount of data that we had to move around. That was a good decision at the time and now we’re replacing that with open technologies. We’re finding that what the open interconnectedness of the technology gives you is the ability to pick best-of-breed. But at the same time, given our requirements and demands, we don’t have that many choices. For a particular type of tool, we’re lucky if we can find two to choose from. So generally, we don’t have the luxury of going after price as a big criterion.

CANTAFIO: At Foster Parents Plan, we’re very much going for an open architecture. We’ve taken a three-tiered approach to data warehousing, and in each tier we’re looking for best-of-breed. We don’t have the issue of having to deal with large databases and so we’ve really been able to focus on the end-user side – finding what we felt was the best cost/performance for us, which were the best tools, etc. Our view would be, if another tool came around that was more cost effective or gave us better performance, we should be able to pull out the one we were using and replace it with another. And as an organization, we’re looking at potentially using a similar approach worldwide. We’ll then be able to really get our economies of scale going as we extend our data warehousing concept out to other countries.

CUDDY: I believe the success of any project is the product of two factors: one is how slick an idea it is, and the other is how well you can get it in place and fully utilized. I will trade off functional excellence in a product in return for other factors that I think will make it easier for us to get it usable. One such factor: Is the product likely to become predominant in the market place? Because if it is, the likelihood of bugs will be lower, there will be more bug fixes, and the availability of resources to support the product will be higher. My objective is generating more value in the company. If you give me a product that’s not quite as good, but for these other reasons I can likely get it in place faster, quicker, better, with lower support costs, then we will do that. Our standards are largely driven by that type of philosophy, which is one of the reasons why all of our data warehousing and EIS information is being delivered through a browser and Internet-based technology, even though some of the functionality that we’ve got in the EIS isn’t as good as it might be with some other vendors’ products.


ATKINS:That’s a good segue into Internet versus intranet and extranet. In terms of access, are you looking at using browsers; and if you are, are you looking at internal versus external, or both?

WILLETT: For employees who may be working from home, we use both browser and traditional dial-in remote access. What that means is we have two security systems, and a help desk that spends 75 per cent of its time on password problems. We are moving to the browser interface more and more. Our issue is the old legacy systems that weren’t ever intended to be browser accessed, but we’re looking at products that allow us to do that. For our customers, we already have the ability to dial in and get online bill presentment. They can’t pay their bill but they can look at it. Like other organizations with large bill streams, we’d like eventually for them to be able to pay it without ever having to give us any physical transaction.

ROSNER: Right now we don’t have access to the warehouse from outside the enterprise. We need to get a lot of our ducks in a row in terms of security and what information we’re actually prepared to give to our sponsors before we do that. But we do have our strategies in place for how we’re going to deliver that. The challenge we’re confronting is that finding best-of-breed tools, and confining yourself only to tools that are available over the browser, may not be consistent goals. So if we’re going to be looking, for example, for Windows-based clients, as opposed to browser-based clients, we’re looking for vendors that have credible plans in place for porting to browser technology. We’ve also invested fairly heavily in other thin-client technologies, like Windows terminal servers, which are stop-gap ways to deliver Windows client technology to a large audience. And we’re trying to establish the same standards both internally and externally.

SEKI: We’re heavily into both the Internet and extranet. An extranet site is in the works right now and it’s going to play a big part for us. Our customers will get various value-adds through the extranet site. Right now we have components built, but the thing we’re all wrestling with is the authentication and the security. We’re spending a lot of time to make sure we’ve got that nailed down.


ATKINS:What’s next for data warehousing at your various organizations?

RAWHANI: The most critical thing in business today that makes you competitive and sets you apart is customer service. That means putting power in the hands of your knowledge workers. The future for us is to make sure data warehousing gets deployed across the board, that we have a very powerful repository of all our information. Pilot the data warehouse in a couple of key areas and then open up access to all users at CIBC Mortgages. At a later date, we can consider giving access to our external clients and vendors.

CUDDY: Ultimately, the biggest thing I’m struggling with is the concept of embodying rules into some of these things that allow us to trigger alerts when thresholds are met. How do you do that effectively? If my exception-reporting system says to trigger an alarm when a certain customer stops buying a product they’ve always been buying, is that a good thing or a bad thing? Maybe they switched to a higher-margin product.

SHOESMITH: One thing we’ve been tracking at IDC is the fact that organizations are looking to get more data and information into users’ hands, but that doesn’t necessarily translate into more knowledge. Organizations have to undertake a process of educating their end-users as to the kind of information that they can derive from these systems and really turn it into something that’s meaningful.

SEKI: I’m not convinced we’re there yet in terms of the warehouse driving behaviour on business decisions. I think the journey to get there involves finding out business-process issues that have led to bad operational data. The warehouse helps you clean house; as you’re building it you have the opportunity to clean up your own operations. From there, and from the new systems we’ve built, I believe that the benefits of data warehousing will become top-of-mind for the business owners. All business people need proof.

CANTAFIO: The future for us is to continue with this iterative process: evolve our data warehouse in terms of depth and breadth of the data, and identify and address the operational issues. Our next step at some point will be to replace components of our legacy system with more client-server solutions that will enhance our day-to-day operations and continue to feed the warehouse, so the users don’t have to get used to a different set of reporting requirements. The major future step for us is moving towards an international data warehouse, which will allow us to share the program information from the more than 40 developing countries in which we work.

WILLETT: We’re going through what consultants call a “discontinuity” in our industry, where we’re moving from being the retailer to the wholesaler, and yet at the same time our ultimate end customer – the person that takes gas into their house or into their business – is still going to want access to the data. We see data warehouse technology as the solution that let’s you have different tiers of access. We’re looking at it as the way that we’ll be able to survive and be successful in the new form that our business is taking – where internal customers and internal users can access whatever detail they need to do their jobs, where our broker customers can only look at the aggregate data to which they’re entitled plus the detailed data to which they’re entitled, and at the same time our individual customer or end-user – who may not be our customer any more, in one sense – can still access their consumption data and make informed choices about which middle broker they choose to buy from. Now we’ve got to get the business to buy in, because this would be a major investment; we’d have to put in place major new technology to support that kind of model.

CIO Canada thanks all of the roundtable members for participating in the event, as well as Tom Atkins, president of TWA Consulting, for facilitating it.