in this issue

Beyond E-Learning

Human-Centered Design Considered Harmful

That Sinking Feeling

Stupid Marriott Marquis Elevators

Workforce Performance Solutions

LMS Survey Results

Amazon: No Longer the Role Model for E-Commerce Design

Customer Storytelling at the Heart of Business Success

Model Problems in Technologies for Interoperability: Model-Driven Architecture

Reasoning Frameworks

by Marc Rosenberg

Amazon's book description:

Is e-learning just a “flash in the pan” or is it a viable strategy for organizational learning? Why does traditional training continually fall short of its goals? In this new book organizational learning guru Marc Rosenberg shows that real performance and learning happens through a confluence of strategies – not just e-learning or classroom training and provides a well-timed roadmap to integrating training, e-learning, knowledge management, and performance support to bring about successful and sustainable learning and performance in organizations.
 

A number of our readers asked EPSScentral's opinion of Donald Norman's latest article, Human-Centered Design Considered Harmful (online version is at this link and another version is available at Interactions of the ACM (Interactions 12.4, July-Aug 2005, pp 14 - 19). In this article Norman appears to take an about-face as he begins questioning the wisdom of the human-centered design process inherent in his own work and in that of usability professionals. Salient comments:

• "If it is so critical to understand the particular users of a product, then what happens when a product is designed to be used by almost anyone in the world? There are many designs that do well for everyone. This is paradoxical...[and led me to] ...re-examine common dogma."
• The automobile configuration is common world-wide - and it works. Kitchens haven't changed significantly in a century - and people "manage quite well."
• Activity-centered design is the key - where an activity is a "coordinated set of tasks."
• The historical record shows that people adapt to good tools, not the other way around.
• Human-centered design can be harmful in many contexts.

If we compare this to what Norman has to say about the power of representation and the role of distributed cognition in making us "smart" (e.g., from his 1993 book Things That Make Us Smart), I would argue that there is much similarity - in spite of the Human-Centered Design evangelism throughout his work. As a matter of fact, it was Norman's inspiration that motivated EPSScentral's Performance-Centered Design (PCD) model, represented by the diagram. What I noticed over the years was that there are really three distinct disciplines necessary to foster performance - which, in the new language, might be called completing activities. The three disciplines are (1) Process Improvement (e.g., Quality, which we trace to Deming), (2) Human (User)-Centered Design and (3) Knowledge Management (including information, content and learning management). In this context, PCD is a process that fosters the proverbial clash of the titans (the three disciplines). Whenever we take any one of these disciplines as religion for design, the things we produce typically fall short of achieving their goals in significant ways. What distributed cognition requires is an iterative process in which activities (dynamic processes, not static, individual tasks), human factors and information/knowledge are represented, and how they are represented is refined until the thing works. The moment we wear one of the three hats to the exclusion of the others, our products are doomed to fail.

Examples? Early altimeter designs in aircraft met the task need, but failed to address the whole activity: They accurately displayed the altitude, but humans could not interpret the readings fast enough to avoid crashing. The process (activity) is know your altitude and do not crash. Most enterprise computer systems (human resources, customer relationship management, financials) require labor-intensive training on "navigation" and business rules. Static tasks are represented, but dynamic processes (activities) are not in any explicit way. The response to this is usually lots of information added to the environment (context-sensitive help, wizards etc.) that mostly make the problem worse. And the list goes on - underscoring that it is not just the human dimension that is at issue, but the representation of the whole dynamic process (activity) and what constitutes minimal knowledge to be successful.

In other words - to quote the brilliant work of Paul Simon - "One Man's Ceiling Is Another Man's Floor." We must diligently and responsibly represent process/activity, human factors and content/knowledge optimally for performance to happen.

Further to this point, consider the following two articles on the topic of New York City's Marriott Marquis' elevator redesign: That Sinking Feeling and Stupid Marriott Marquis Elevators. In the first article, we learn how brilliantly Schindler's Miconic 10 Elevator solved the abysmal queuing problem at the Marquis, where guests used to suffer 45 minute waits:

"...the Miconic 10 organizes passengers according to their registered destinations. In other words, you key in your destination before you get in, then leave it to the elevator. There are no buttons inside the elevator."

But immediately following implementation, Doc Seals writes about his experience:

"...this ridiculous system with a central keypad that you type your floor number in, that then tells you which elevator will be coming to pick you up. The elevator has no keypad, because it supposedly knows where to go. Unless it doesn't, which you can't tell, since there's no display, in which case you are trapped into going to some random floor."

Again, “One Man’s Ceiling Is Another Man’s Floor.” To Don Norman’s point, there are apparently a few artifacts missing from the activity space of Miconic 10 that transcend human factors shortcomings. Yes, Doc Seals needs more feedback. But the activity of getting from point A to point B is dynamic. It is not a static task. Whatever design insight made Schindler dispense with conventional buttons and displays is refreshing – but the implementation fell short of the complete activity. For the entire system (that would be human being, process and information) much was omitted from the design beyond the task of calling the elevator to the appointed floor.

As usual, I will continue to sing praises to Donald Norman in spite of the apparent 180-reverse. He inspired EPSScentral about twelve years ago in a way that is completely consistent with this 2005 inspiration. We knew you’d get there, Dr. Norman. Keep up the brilliant work!

Regards,

Gary Dickelman

PS Watch for Marc Rosenberg's latest book, Beyond E-Learning. It represents the greatest concentration of contemporary wisdom and experience around e-learning to date - including his own and that of the most seasoned professionals who truly make performance happen. Beyond E-Learning will therefore make better decision-makers of leaders such as CEOs, CFOs and CTOs and, equally, will make better managers and practitioners of those charged with ensuring organizational performance through human performance.

Many design elements work for Amazon.com mainly because of its status as the world's largest and most established e-commerce site. Normal sites should not copy Amazon's design.

It looks like a usability expert got lost in the world of business modeling and made a home nonetheless. If we replace the word “customer” with “user” this article reads like a usability professional’s 101 course. But think about it: For years our mantra has been “Business performance through human performance.” Human beings are the customers, right? A business is a thing that must support activities for human beings, right? Cross-discipline pollination between marketing professionals and human performance / usability professionals must therefore be a good thing. Great article!

Abstract: Model-driven architecture (MDA) is a technology produced and maintained by the Object Management Group (OMG), an open membership, not-for-profit consortium that produces and maintains computer industry specifications for interoperable enterprise applications. This technical note examines two claims regarding the benefits of MDA, namely, that it (1) reduces development time, and (2) allows the developer to focus on business logic rather than on details about the target platform and architecture. Such advantages would greatly benefit interoperability; as target platforms and underlying infrastructure change, deployment of applications would be quick and easy. This note presents the results of applying the model problem approach to verify these claims.

Abstract: Determining whether a system will satisfy critical quality attribute requirements in areas such as performance, modifiability, and reliability is a complicated task that often requires the use of many complex theories and tools to arrive at reliable answers. This report describes a vehicle for encapsulating the quality attribute knowledge needed to understand a system's quality behavior as a reasoning framework that can be used by nonexperts. A reasoning framework includes the mechanisms needed to use sound analytic theories to analyze the behavior of a system with respect to some quality attribute. This report defines the elements of a reasoning framework and illustrates the reasoning framework concept by describing several reasoning frameworks and how they realize these elements.