Dan Power, Editor of DSSResources.com, conducted an email interview with Hugh Watson in October 2005.
Q1: How do you define the concept "Decision Support System?"
Watson's Response: DSS is both a generic term that describes a field of study and a term that describes a specific type of computer application.
When used to describe a field, it refers to the study of any system that is used to support decision making (e.g., EIS, OLAP). When conceptualized this way, applications can be categorized as being either for (1) transaction processing (i.e., operational systems) or (2) decision support. The focus of attention of DSS is on the latter. This is somewhat simplistic, however, because of the increasing trend to embed or integrate decision support applications into operational systems (e.g., fraud detection systems embedded in credit card processing), but it provides a good starting point.
As a specific application, there are several alternative conceptualizations, but I prefer the one initially provided by Ralph Sprague. Using Ralph’s conceptualization, a DSS is characterized by a user-friendly dialog (i.e., interface), a database that supports the application, and a model base that manages the models (e.g., the analytical tools and the models that are created). The technologies have changed considerably since when Ralph first developed his conceptualization (e.g., web browsers, data warehouses), but it still holds up.
Steve Alter provided another early conceptualization that is more in line with the thinking that a DSS is any system that supports decision making. Steve’s framework identifies the various kinds of systems that can be used to support decision making (e.g., model or data based systems) and is very broad and inclusive of a variety of applications. I sometimes find Steve’s conceptualization useful.
Q2: How did you get interested in DSS, EIS, BI, and data warehousing?
Watson's Response: When I was in graduate school (I finished my doctoral degree in 1969) there wasn’t as much specialization as there is today, and there wasn’t much coursework in information systems. Because of my interests and undergraduate engineering degree, I gravitated to management science and initially taught and researched primarily management science courses at the University of Georgia.
In 1973-74, I took a leave of absence to be a visiting professor at the University of Hawaii and Ralph Sprague introduced me to information systems in general, and DSS in particular. We ultimately wrote four or five DSS articles together and several editions of a DSS book. At about the same time, the management science field was becoming overly specialized (in my opinion) and less relevant to business. When I returned to Georgia, I switched to IS, and continued to do DSS research.
In the early 1980s, I attended an Atlanta SIM meeting and met George Houdeshel, the EIS manager at Lockheed-Georgia. He invited me to see his system, and what I saw “knocked my socks off.” Though the technology was primitive by today’s standards (it was DOS based), the thinking about how to support executives, the understanding of what information executives need, and how to best present information was way beyond what I had seen before. For about a year, I went to see George and learned from him. Ultimately, we entered the SIM competition and won an award for the EIS at Lockheed-Georgia. This kicked off my EIS program of research.
In the early 1990s, EIS was becoming the “old news” as Peter Keen used to describe the maturing of a topic, but data warehousing was just starting to take off. I was invited to join the faculty of The Data Warehousing Institute, and using that organization and the contacts that it provided, I started my research on data warehousing. I continue to specialize in BI and data warehousing today.
The common thread throughout my career has been decision support -– management science, DSS, EIS, and now data warehousing. Because of my interest in studying the “new news,” I move on to new areas as the old became less interesting and important.
Q3: You were a leader in investigating Executive Information Systems (EIS). What has happened to EIS?
Watson's Response: Executive information systems are alive and well. I find them in many of the organizations that I work with. EIS just don’t get as much attention as they used to; they are the “old news”. Vendors and technology publications always promote what is new.
In a lot of ways, it is easier now to have a successful EIS than it used to be. A major problem for EIS was the difficulty of supplying timely, consistent, accurate data, but data warehouses now make this much easier. Web technology makes it easier to deploy and maintain systems, because everyone had access to a browser from almost any location. The Internet provides easy access to external data, while corporate intranets make it possible to access vast amounts of internal data.
It is important to recognize that EIS have also morphed into corporate portals and business performance management (BPM) systems. Dashboards and especially scorecards have many characteristics that we usually associate with EIS – a wide range of data (to provide a balanced perspective), the focusing of organizational attention, an intuitive interface, the ability to drilldown to underlying detail data, KPIs (similar to critical success factors), and a link to business strategy (at least in the case of scorecards). The business need for EIS has not gone away. It is just in some cases the needs are being met with systems that are both different and similar to EIS.
Q4: What keeps you interested in computerized decision support systems?
Watson's Response: I continue to find DSS interesting, intellectually challenging, and important to companies. I’ve always preferred to research topics that have “market value,” in that companies have an interest in the research findings. It makes it easier to find support for research, creates opportunities to work with companies in various ways (e.g., consulting), and improves the relevancy of classroom teaching. Some aspect of DSS is always important to industry; but it is constantly changing, so it is important to change one’s interests as well.
Q5: You are recognized for conducting programmatic research on EIS and data warehousing. Do you have a “model” for doing it?
Watson's Response: Yes, I do, though it is not completely my own. I’ve benefited from discussions with Gary Dickson about the “Minnesota Experiments,” Jay Nunamaker about his research on group support systems, and Jack Rockart of MIT’s Center for Information Systems Research. Each has provided useful insights that I’ve incorporated into what I do.
As a starting point, it is important to select a topic that is of emerging importance. It needs to be important to the business community and of interest to you. You want to write some of the first articles on the topic in order to have a “first mover” effect.
It is also important to become an expert on the topic. The best way to do this is to identify, work with, and learn from the best people in the field. These are usually leading-edge practitioners, not other academics.
Because the topic is new, there is the potential to develop frameworks and conduct survey research that makes sense out of what is taking place. Versions of this research should be published in leading practitioner outlets (e.g., Computerworld) so that you get recognized. This can lead to future research funding and other opportunities.
I’ve also benefited from being affiliated with the leading practitioner organization (e.g., EIS Institute, The Data Warehousing Institute) for the topic that I’m studying. Be a speaker at their conferences, help with their research, be a judge for their competitions, or work with their publications (in my case, the Business Intelligence Journal). This affiliation puts you into contact with leading practitioners, consultants, and vendors.
It is also important to get graduate students and other faculty involved in the research, because you can’t do it all yourself. This isn’t hard to do, however. You have the expertise, an understanding of what merits study, access to experts and funding, and the ability to use your contacts to collect data. In other words, you can do the things that are difficult for graduate students and junior faculty to do, and as a result, they are pleased to work with you.
Q6: In your role as Senior Director of the Teradata University Network (www.teradatauniversitynetwork.com), what is your vision and goal?
Watson's Response: The mission of the Teradata University Network is “To be a premier academic resource for knowledge about data warehousing, DSS/BI, and database. To build an international community whose members share their ideas, experiences, and resources with one another. To serve as a bridge between academia and the world of practice.”
Currently, we have over 900 faculty members at 515 universities in 53 countries. So, to some extent, through the efforts of Teradata, our software partners, and the contributions of our Board (especially Jeff Hoffer and Barb Wixom) we have achieved our goal. However, we want to do more. We want TUN to become so useful to faculty that it becomes an integral part of all the data warehousing, BI/DSS, and database courses that faculty teach. For example, I teach a data warehousing course for MBA students, and all of the instructional resources that I use come from TUN and its companion site for students, the Teradata Student Network.
I’m really excited about continuing to add state-of-the-art software for faculty and students to use. Currently, we have software from Teradata (database), Hyperion (BPM), and MicroStrategy (reporting and OLAP) and we plan to add at least one major software partner each year. The software is available through an ASP arrangement and can be accessed using a Web browser. Most universities don’t have the human and technical resources to support state-of-the-art, n-tier architectures well, and I’m convinced that the ASP model is the way to go.
Q7: What do we really know about how computerized decision support impacts decision making? What generalizations are supported by research?
Watson's Response: DSS impacts decision making in many ways, depending to some extent on the nature of the application. To illustrate, optimal solutions may be identified, more alternatives may be explored, decisions may be made faster or even automated, group decision making may be facilitated, and so on. These impacts have been found in many studies and I think that there is no doubt that there are positive impacts from DSS.
Personally, I find assessing the success of decision support systems to be interesting. For my work, the success framework of DeLone and McLean, with its focus on information quality, system quality, system use, user satisfaction, individual impacts, and organizational impacts, has been very useful. Having said this, I’ve never found it possible to use the success metrics from other studies without significant modification. Success constructs and questions need to be customized for the phenomenon under study. While giving proper respect to theory, literature, and the need to build a cumulative body of knowledge, I find it very helpful to involve experts (e.g., consultants, managers) in determining the constructs and especially the questions.
Q8: Why are Business Intelligence systems so popular?
Watson's Response: BI is a collection of different types of technologies and applications. Collectively, they provide a powerful set of analytical capabilities that can affect organizations’ ability to do more with less; better understand customers, their needs, and their profitability; integrate and optimize supply chains; find previously unknown relationships; and so on. Most fundamentally, they give organizations the ability to compete more effectively in the marketplace.
Q9: Do we need a theory of computerized decision support systems? If so, what are the main elements or a starting point?
Watson's Response: I don’t think that we need to find a single theory or framework. Furthermore, I don’t think that we will see a single overarching theory emerge. Rather, there will be multiple theories, each one being appropriate for specific situations.
Theories are both good and bad. They are good in that they help to understand and predict a phenomenon, provide a conceptual framework for organizing ideas, organize previous research and knowledge, and link current research to the larger body of knowledge about the phenomenon. On the other hand, too much focus on theory-based research can lead researchers to use a lens that isn’t the best from a practitioner perspective. Also, too often researchers doing theory-based research rely too much on the theory and don’t spend enough time fully understanding the domain they are studying. I always require my doctoral students to spend time with practitioners who actually do for a living what is going to be studied. They always come back with an enhanced and different understanding than what theory provides. Without exception, the final research model benefits from and is different from what the theory suggests.
Q10: What role should theory play in DSS research?
Watson's Response: As I mentioned, there are many good reasons for conducting theory-based research, whether it is theory building or theory testing. The problem occurs, however, when the use of theory becomes such an important consideration that it gets in the way of doing good research, rather than supporting it. This is most likely to be the case when a new phenomenon is being studied or when theory does not provide an appropriate lens for framing the research questions and model.
Let me give you a personal example. Several years ago I collected data about the inclusion of soft data in EIS and the success of the system. I found that as EIS mature they include more soft data. I also found the inclusion of soft data is positively correlated with the success of the system. This was an important finding, but because I could not find a good theoretical basis, I was not able to publish the research in a leading academic journal.
It is appropriate to use theory in DSS research, but we should accept and encourage important research, whether or not it is theory based. We should not apply a theory litmus test as to whether or not research gets published in our best journals.
Q11: Where and what do you think the next generation for DSS will be as we move ahead?
Watson's Response: I see multiple developments in DSS. I’m especially high on real-time BI and data warehousing. They have the potential to significantly change the scope of decision support from strategic and tactical to include support for operational decision making. With real time data and analyses, current decision-making and operational processes can be improved. This is especially important for customer-facing and supply chain applications.
Directly related to real-time BI and data warehousing is the ability to monitor processes in real time. Business activity monitoring (BAM) is the name given to this and it is the equivalent to process control in a production environment. With this monitoring, the use of alert technology becomes more feasible and useful.
I also like the emergence of guided analytics where the software helps the user take the next logical steps after an initial analysis. It is like having the best analyst as your helper. Related to this is the emergence of workflow analytic software. We have long understood the importance of being able to communicate and share the result of an analysis with others, and software is now starting to support this much better.
Data visualization has been around for a long time, but there have been recent developments in terms of thinking of how to best present data and there are now many new software products for doing this.
ReferencesAlter, S.L., "A Taxonomy of Decision Support Systems", Sloan Management Review, vol. 19, no. 1, Fall 1977, pp. 39-56.
Delone, W.H. and E. R. Maclean, “Information systems success: The quest for the dependent variable”. Information Systems Research. vol. 3, no. 1, 1992, 60-95.
Sprague, R, H., Jr., "A Framework for the Development of Decision Support Systems," Management Information Systems Quarterly, vol. 4, no. 4, Dec. 1980, pp. 1-26.
About Hugh Watson
Dr. Hugh J. Watson is a Professor of MIS and a holder of a C. Herman and Mary Virginia Terry Chair of Business Administration in the Terry College of Business at the University of Georgia. Hugh is a leading scholar and authority on decision support, having authored 22 books and over 100 scholarly journal articles. Hugh helped develop the conceptual foundation for decision support systems in the 1970’s, researched the development and implementation of executive information systems in the 1980’s, and most recently, specializes in BI and data warehousing. Hugh is a Fellow of the Association for Information Systems and The Data Warehousing Institute and is the Senior Editor of the Business Intelligence Journal. He is also the Senior Director of the Teradata University Network, a free portal for faculty who teach and research data warehousing, BI/DSS, and database. For the past 18 years, Hugh has been the consulting editor for John Wiley & Sons’ MIS series.
Power, D., "Hugh Watson Interview: Understanding Computerized Decision Support", DSSResources.COM, 11/04/2005.