back Stories Overview

User Experience at SAP



Conference Reports

Future Scope


CHI 2006 – A Selective Reflection...

By Gerd Waloszek, SAP User Experience, SAP AG – June 26, 2006

CHI 2006 logoCHI 2006 is over, time to pick the leftovers and write a report. This year, however, my report is more a reflection on some of the issues that attracted my interest than an "experience report" – it is incomplete, subjective and maybe controversial.


Business Value Versus User Value (or: Business Goals Versus User Goals)

One observation that struck me at CHI 2006 was how business-oriented many discussions were. At least two panels touched this issue (Building User Value into the Business Case, The Route to the Sea for User Value). Some HCI advocates never tire of mentioning how important it is to "speak" the language of management (that is, of the business people) when trying to advance the usability case in companies. They say that the benefits of usability efforts need to be framed in measures, such as ROI or TCO, or rephrased in figures, preferably dollars – this is the only language that managers are said to understand. In this strategy, there is an implicit assumption, though, namely that business goals and user goals (or business value and user value) coincide – an assumption that is definitely worth questioning. Certainly, this assumption seems to hold for many examples: Most users want to perform their tasks as fast and efficiently as possible. Surely, their managers agree with this goal. However, it would be naive to believe that company and employee goals are always the same. In the case of a conflict, I would bet that business goals have priority over user goals, even though some of the panelists hesitated to admit this. I must add that I am somewhat tired of these stereotyped pleas for speaking the "dollars" language in order to be understood by management. I agree that there is a business case for usability. But there is a growing tendency to rephrase human values and characteristics as "capital," such as social or intellectual capital (see Leif Edvinsson, for example) – something that I oppose to. With this thinking, a picture of managers is maintained that characterizes them as somewhat narrow-minded and only able to understand one language, that of money. I am not willing to accept such an over-simplification.

Panel User Value

Figure 1: The panel Building User Value into the Business Case in search of the user value...

It is also interesting to see what usability work looks like in cases where business and user goals are regarded as coinciding. At the Building User Value into the Business Case panel, it was reported that at Yahoo! they measure where revenue is generated on a daily basis and adapt the user interface accordingly. Here, revenue is definitely the main driver for the usability work. A panelist from Microsoft was more cautious: He put user value first and revenue only third – but was he honest? (OK, what else could he say at an HCI conference...). At CHI 2005, Jared Spool had criticized ad hoc usability work that is performed without any theoretical orientation. He presented Amazon's A/B methodology as an example and concluded, "Nobody knows why something works: they don't know what they are learning." (A/B methodology: Two versions of the site using two different user interface variants were run in parallel and compared with respect to generated revenue; the less successful was discarded) In my opinion, such an ad hoc strategy is not acceptable if we want to maintain that HCI has at least some roots in science. I still have the ambition to start from a few general principles that tell me beforehand what the better user interface looks and feels like. (Of course, the design needs confirmation from user tests.) But I also know that HCI is still far from a satisfying state in this respect.

Last but not least, we should be careful not to transfer findings from heavily-used Websites, such as Amazon, Google, eBay or Yahoo!, to other software, for example, to complex business software as delivered by SAP, without reflection. Even though some "user advocates" present their finding with a lot of verve, it often turns out that they have a very narrow image of the software universe in mind. These Websites address the mass market, so everybody in the world should be able to use them, but they serve only highly specialized purposes. The simple truth is that controlling a company or managing a supply chain is not like bidding in an eBay auction. Thus, as Albert Einstein said, "Everything should be made as simple as possible, but not simpler."



One CHI 2006 panel debated the relationship between HCI and the Agile software development framework (Agile Development: Opportunity or Fad?). I must admit that I had heard of Agile only about a year ago, even though this methodology has been around for some years: The Agile Manifesto was formulated back in 2001 and, according to Wikipedia, is "widely regarded as the canonical definition of Agile development and accompanying Agile principles." Extreme programming, one of the methods used in Agile development, has existed even longer and was defined in 1996.

Agile manifesto

Figure 2: The Agile manifesto in a nutshell

It seems as if I was not alone in my ignorance. The whole HCI community came to recognize the Agile movement only just recently. And – to its dismay – the community is gradually realizing that usability methods and users have been left out of Agile. In response and fearing being left behind, some HCI professionals are eagerly trying to demonstrate that usability can be incorporated into Agile. For example, Karen Holtzblatt, whose methodology has often been praised but also criticized for being too costly and time-consuming, set out to show how contextual design can be streamlined to fit into Agile. (She gave tutorials at CHI 2006 in which she tried to demonstrate this.)

Many war stories of usability people who have worked with Agile teams go like this: " I tried to convince the Agile-driven development team that usability is important and they were quite receptive and thankful for that." That is indeed an interesting statement. With the CHI conference turning 25 next year and a HCI history of more than 50 years, usability should not be anything new to software developers. Why was usability left out of the Agile framework? I firmly believe that the Agile people did so on purpose. Perhaps they still believe that programming is an intellectual exercise that has nothing to do with users. They may also have a "the journey is the reward" way of thinking – the final product that users have to cope with is not on their radar screen, nor are design iterations with users during the development cycle. Maybe I am being a bit cynical here, but I wonder how a software development process can be defined these days without including users.


Human Information Interaction (HII)

In 1998, SAP's usability people had a meeting to launch the Enjoy initiative. At this meeting, everyone was asked to make a short statement on the future of user interfaces. My statement was, "I regard computers as 'information machines'." Most people looked at me somewhat puzzled after that statement. Nevertheless, it has become increasingly true during the last few years. In an abstract sense, everything that computers deal with is information. During the past decade, however, computers have turned into a true "informational medium" that competes with as well as complements traditional media, such as print, film, television, broadcast, and even the phone. Thanks to the Internet – but not only the Internet – we have traditional media on our computers, too.So, on the one hand, all information seems to gather on our computers. On the other hand, however, dealing with information has a much broader scope than focusing on computers provides. Computers are only a medium or tool. The latter has convinced a number of (mostly) researchers, among them Peter Pirolli from Xerox, to start an initiative promoting a new research direction called Human-Information Interaction (HII), which focuses on the ways in which humans interact with information. The downside of such a new direction would be a separation from the HCI community and establishment of a new one.

Peter Pirolli

Figure 3: Peter Pirolli making the case for Human-Iinformation Interaction

In the CHI program the HII promoters say,"The 'I' in HII implies a focus on information and not computing technology. But what does this mean? Is there any way to focus on information without also considering the supporting tools, applications, and gadgets that are enabled by computing technology?" Accordingly, one of the controversial issues during the panel's discussion was: "Do tools matter or not?" Or on other words, "Are the main information phenomena device-independent or not?" Not surprisingly, this issue was not resolved – both sides, the pro-HII and the anti-HII people, presented examples that supported their stance.

During the discussion, it also emerged that an HII research direction might become a "bottomless pit" because it is unclear what should it should cover. The traditional media already have their well established research fields, institutions, and communities. It is not to be expected that they will be willing to give up their identities and integrate themselves into the new HII direction. In the end, the panel discussion came to nothing: no separation, everybody happy... I am curious as to whether someone will be willing to resume the discussion and repeat the attempt at a split. Admittedly, I had some sympathy for it, but it is evident that the subject of HII needs to be defined as clearly and tightly as possible.



Of course, there were plenty of other topics at CHI 2006. Only afterwards, did I realize the breadth of the conference. During the conference, I was not so much aware of it – maybe because I had the feeling that some of the topics were not dealt with satisfactorily. To pick just one of the many topics, there was a panel on personas, which revealed to me that I am not alone in wondering what influence a persona's hobbies or number of kids have on the design of the user interface. Probably, there are also cultural differences that determine how much effort designers put into these aspects.

CHI madness

Figure 4: Patrick Baudisch introducing CHI Madness

Research overviews were one of the many new additions to the CHI conference. There were two such sessions: Mary Czerwinski gave an overview of large display research and Brad Myers one on end-user programming (including graphical programming languages, into which I did some research long ago). The latter leads me directly to Web 2.0. This appeared at CHI 2006 disguised as "mash-ups," to which a panel was devoted (Add a Dash of Interface: Taking Mash-Ups to the Next Level). Even though I "matured" in an era when people bought computers to program them, I have become skeptical with respect to end-user or "mass" programming. The proportion of users who do some sort of programming on computers is small compared to the proportion of "pure" users. If these "computer freaks" become highly productive and supply us with a lot of programs, we may have a whole host of problems, particularly if there is not such strict quality management as in the open source community. We will be flooded with buggy, insecure software that has minimal or no support – and no future; in other words, nothing that can be used in professional environments. The insufficient sustainability of software and computer documents (which are usually stored in application-specific formats) especially troubles me the longer I work with computers. Applications come and go, and I spend my valuable spare time converting documents (which are stored on media of unknown life expectancy).

While not all changes to the CHI conference were for the better, one innovation was welcomed by most attendees: CHI Madness, introduced and led by Patrick Baudisch, provided an overview of a day's presentations within half an hour. Some of my German colleagues appreciated this short run particularly, because it helped them to get an impression of the speakers' understandability and thus to finalize their choice of presentations for the day.


Final Remark...

One more highlight at CHI 2006 was a fire alarm during the hospitality events. I had a small problem with it, though: When the alarm sounded, I had still not eaten anything. At least I had had a bottle of beer or two – which can be proven: An SAP colleague and I appeared in the CHI 2006 slide show before the closing plenary.

CHI slide show

Figure 5: CHI 2006 reception shortly before the fire alarm



CHI 2006

Mary Czerwinski (Research overview large displays)

Brad Myers (Research overview end-user programming)

Mash-Up Panel

Human-Information Interaction

Leif Edvinsson

  • UNIC Website (offers definition of intellectual capital: What's IC >> Insight)


All photos by Gerd Waloszek, Minolta Dynax 5D w


To top top