To anyone who studies or works in the field of data warehousing, the name Ralph Kimball should need no introduction. Through his research and innovations, his books, speeches and now 12 years of teaching more than 10,000 students at Kimball University, Ralph Kimball has become a data warehousing icon. For those needing a quick refresher, Kimball followed his 1973 Stanford Ph.D. thesis on man/machine systems with a long stint at Xerox's legendary Palo Alto Research Center (PARC) where he was part of the 1970s research team that hashed out the first iterations of 'windows,' icons and the computer mouse. These innovations were revolutionary - and consigned only to be a fancy word processor - until Kimball and some others recognized the missing element: data. With former Xerox colleagues, Ralph was a member of the founding team of Metaphor Computer Systems, and later, he personally founded Red Brick Systems. The work that ensued helped set the stage for the data warehousing industry that exists today. Tiring of executive roles, Kimball returned to his roots in 1994 when he founded Kimball Group, where he began to teach the lessons learned in his consulting work on some 30 enterprise data warehouses. Despite his many reference points, the consistent thread for Kimball has always been to start with the end-user experience, which will come as a surprise to pure technologists. "Only when I am pretty confident that I have the end-user part right do I start working back to eventually arrive at the technology," Kimball says. While his task-oriented philosophy has occasionally put him at odds with others in the data warehousing industry, Kimball has never sought controversy on his own. DM Review Editorial Director Jim Ericson recently caught up with Ralph Kimball for an update on the current state of business intelligence and data warehousing.
DMR: Across the span of your career, how have perceptions changed in terms of the ways people use data?
RK: It's not that the data has changed so much as the expectations of end users have changed. This is kind of an old story now, but I still find it interesting. In 1980, end users had not been educated in business school to use data, and they didn't have a computer to carry around. We did a lot of evangelizing in those early days at Xerox and Metaphor to just wake people up to the idea that they could do something with data. Today, there is an implicit assumption that people can analyze and do better, which leads to a certainty of needing data. But at the same time, the end users' ability to use the computer has not changed very much. I ask my IT students how many of their end users actually program a spreadsheet cell, use the little "=" sign and sum "A1" plus "B2." I get them to admit that maybe 10 percent of the population has reached that level. So there's your 10 percent who can actually use an ad hoc tool. Everybody else needs support in terms of an end-to-end user interface and a forgiving computer environment.
DMR: From that perspective, it sounds like there is an expectation gap.
RK: Well, there has been a huge move to a much more operational involvement with data and information. I think data warehousing has stepped up to the top level of the podium in the last two or four years. I call it data warehousing, but it's really the platform for BI. The need for data and data warehousing has become hugely serious. It has penetrated to all levels of the organization, so if we look at who the average user of a data warehouse is, you're right. The original concept of the data warehouse 10 or 15 years ago was a kind of research facility. Now it is very much more penetrating at every level of the organization, 24x7, real time, integrated, tactical, operational and compliant. It has made the job distressingly difficult.
DMR: Who is a typical student at Kimball University, and how do you address this "difficulty" with them?
RK: Most of our students are designers and builders of enterprise data warehouses. We close the door and level with them. We tell them their job is ridiculously overconstrained, and to survive, they have to work in an organization that understands the enormity of their role and gives them a lot of executive support, boundaries and definitions that will let them get their job done. You cannot integrate the data in a huge organization all by yourself. You can't define the boundaries of compliance by yourself and determine what the acceptable practice is. If you don't have executive awareness, executive boundaries and executive responsibility for any of those things, you will have an impossible job.
DMR: It sounds like your students don't have much control of their destiny.
RK: Many who attend our classes and read our books manage the data warehouse team, a small group of people. They typically have multidisciplinary interests. They got to where they are because they like both the business world and the technology world. They are good at communicating with people. In my view, the prototypical "nerd" type of person doesn't usually last long in data warehousing. Our students are idealistic, but they are realizing the enormity of their task and there is always a new challenge on the horizon. Lately, issues of security and the threats of exposing data on the Web are distractions I wish we didn't have.
DMR: Should security be the data warehouse team's problem?
RK: There is only one small group (the data warehouse team) that can understand simultaneously who the end users are and what are their legitimate needs for data. They also understand the databases themselves, so that little department is the only place where you can mediate security needs and manage roles. In addition, you have auditors looking over your shoulder asking who used the data and who changed it. You cannot outsource the security function to another group because they don't have the context or the ability to make the decisions. I actually don't like security and rarely teach it. I am sympathetic, but I point out that we have more basic issues, such as interviewing end users and building models for them.
DMR: Has all the focus on BI tools and platforms obscured the underlying role of data specialists?
RK: Standing back from that question, I would say that the fact people are having this discussion using the words "business intelligence" rather than "database" is very good. It reflects the shift of power to the end-user departments that are now saying, "IT, if you are worth anything, stop being so damn technology-oriented and come to understand our issues and help us solve them." I think that is a healthy challenge. That doesn't mean the onus is only on IT to do it right. I think the business needs to meet them halfway.
DMR: A lot of executives seem to expect that data should be automated and available like the dial tone on their telephone.
RK: That is an unfortunate analogy; in some sense, the opposite is true because the data itself is the problem. It is very labor-intensive to sort out, to wrestle to the ground, to get it into the ETL pipeline and eventually expose it to the business intelligence tools. That's why I have been in the back room for so long even though my interest is more with end users. We have figured out what makes end users happy, but the data thing is a nightmarish problem I honestly would have preferred to leave in the back room.
DMR: Is there a strategy that can succeed here?
RK: I tell my students never to ask, "What do you want in a data warehouse?" It would force the end user to interpret technology, which is not their job. Instead, if the person is an advertising manager, you ask, "How does an advertising manager know when he's doing a great job?" It completely twists the interview around and gives them confidence to answer in their own vocabulary. Another issue is general experience. A very formative time for me at Red Brick was spent at Procter & Gamble. I got to know everybody there and spent time with Bob Herbold, a rising star at P&G who went on to Microsoft. Bob and the team at P&G had a wonderful management strategy. They told everyone in the company that if they wanted to be more than a first-level manager, they had to go through a management culture that might send them from product marketing, to one of their factories, to a tour of duty in finance and then maybe 18 months in IT. The thing I like most about what P&G did was they had trust and understanding between departments and personal contacts that simply made the company collaborate better. It wasn't so much about changing careers as remembering what it was like to work in finance.
DMR: Is that still a viable strategy in our specialized, outsourced world?
RK: I tell students the best thing they can do is to live in the end-user departments during part of their career, become a member of that department and then come back. The best data warehousing person is permanently confused about whether they like the technology or the business better. I don't think you can outsource that, and overall I think the pendulum needs to swing back in favor of IT people who are simultaneously connected to the business and the technology.
This article originally appeared on www.dmreview.com.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access