ASIS & T 2002 Information Architecture Summit: Refining the craft

This text documents my experiences at the ASIS&T 2002 Information Architecture Summit, March 15-17, 2002, Baltimore, MD.

airport tunnel
airport tunnel
aquarium
aquarium
bail bond devil
bail bond devil
baltimore water
baltimore water
city hall
city hall

Friday, March 15 - Preconference day

I attended a preconference facilitated by Noel Franus and Christina Wodtke of Carbon IQ ( http://www.carboniq.com/ ) on the topic of usability testing. Since this conference is being attended by "IA's" (information architects), this preconference did not so much discuss and describe what usability testing was but more on how to do it.

The first step in usability testing is to learn about the site to be tested. This entails interviewing the stakeholders of the site and learning who the intended audience is and what the site is expected to accomplish. It helps to have a face-to-face interview with these stakeholders.

The next step is the recruitment of usability testing participants. Franus and Wodtke seemed to advocated the use of a recruiter to assist in this process, but in actuality, this will be more expensive than most people/institutions can afford or want to spend. Francus and Wodtke advocated the creation of a few "persona" representing the types of people expected to use the site and then the creation of screening documents -- questionnaires -- to be given to people in an effort to decide who are good candidates for testing.

Third comes the creation of a protocol document. This document is an outline of how the usability testing will be conducted. It describes what sorts of things are to be tested (i.e. content structure, navigation, validation, etc.). It is a place where the business goals of the site are echoed, and a script of how the process will be conducted. It is a document where the task list will be enumerated. The presenters stressed the test itself should be more like a narrative process usually consisting of the following parts:

  1. a welcome
  2. a non-disclosure agreement (NDA)
  3. thinking out loud exercise (i.e. "This is how I use a stapler.")
  4. testing
  5. expression of general impressions
  6. natural use of the site
  7. conclusion

The meat of the matter includes the testing itself where a set of tasks are to be performed. There are any number of ways this can be accomplished but the "behind the elbow" approach was advocated as the easiest and least expensive. In this case the moderator -- the person asking the questions and/or directing the test -- sits behind and at the elbow of the tester. This way the moderator can see the face of the participant as well as the computer screen. Another person, the note taker or "scribe", records the entire process being very careful not to make any judgments about user behavior. During the test it is important to use open-ended questions, be calm, cool, collected, and be sure to stress the user is not being tested but the interface instead. Good questions to continually ask during the test process include:

Before test participants leave they should be compensated for their time. Cash seems to be the best compensation, but food, t-shirts, and gift certificates work as well. Discounts on future products are not necessarily good compensation because the testers might not like the site, service, or product being tested.

The last step of usability testing is a report. The report can be short or long, formal or informal, written or oral. Usually a combination of all the above is in order. The developers of a site will want to know what functions to add to the site (even though functions should probably be removed). Designers will want to know what colors to add or change. The IA will want to know how to improve navigation. All of these things should be addressed in the report. The goal is to meet all of these needs.

Start the report with the text/notes from the tests. Look for patterns in the notes and use quotes and/or video clips to support observations. Then make useful recommendations -- recommendations that are specific. At this point it is a good idea to call a meeting of the stakeholders to present the report. It is important to hilight the successes of the site and the testing process. This is a political game where the stakeholders must feel good about what they are taking away.

coast guard
coast guard
colonist
colonist
DAIAD
DAIAD
ducks
ducks
fish
fish

Saturday, March 16 - Day #1

Steve Krug - "Confessions of a SIGIA-L Lurker: A pinhead's view of IA"

After introductions by Dick Hill of ASIS and Andrew Dillion , a primary organizer of the conference who said, "Information architecture is a craft because the idea of a craft connotes the non-separation of design and manufacture." Steve Krug gave the opening remarks. Krug is the author of the popular usability book entitled Don't Make Me Think. He said he started out as an English major doing typesetting after graduation. From there he was a technical writer, specifically about the Kurzweil Reader. After that he got involved in Apple Computer's eWorld service, and when that service ended its lifetime he moved into usability testing for Internet-based services.

He went on to "confess" he does not know about information architecture as much as he knows about usability. More specifically, confesses to know about some of the more typical aspects of information architecture, namely: search, labeling, navigation, and hierarchies/organization. At this point he contrasted two different aspects of information design. On one hand there is the information design of libraries where the principle characteristic is organization. On the other hand there is the process of making accessed information more understandable. I thought this was a very interesting point.

It was around this time Krug alluded to a couple of publications written by Jesse James Garrett. After perusing Garrett's site, I believe Krug was alluding to the following items which look particularly interesting:

Krug does subscribe to the SIGIA-L mailing list: http://www.asis.org/Conferences/Summit2000/Information_Architecture/listserv.html

He enumerated the top five things the IA community can't seem to stop talking about:

  1. Tools - When is the next tool going to rear its head and make our job easier
  2. Definitions - Specifically, "What is IA?"
  3. IA vs. ia - "What is the difference between Big IA and Little IA?"
  4. Research findings - Krug did not seem to give very much credence to the formal research process. He thought usability and information architecture was a process of refining, and what is needed is to communicate in ways that make sense to people. In his opinion, research does not necessarily do that.
  5. Proof of ROI - In other words, how does information architecture help the bottom line? Does information architecture provide a return on investment?

When Krug does consulting (www.sensible.com) he is always sure to educate people about usability and to put forward best practices whenever possible. "Do not tell people their websites are poorly done. Communicate the need for change by illustrating other ways to accomplish the client's goals. Show how things can be improved." He stressed the importance of providing usability/information architecture services not because you are the expert, but because you have the time to provide the service.

Keith Instone, Andrew Dillon, and Christina Wodtke - "Responding to the keynote or 'ROI 4 IA'"

Each of these people very briefly responded to the keynote address, and the theme most echoed was, "It is time to give IA away. We have to teach IA as quickly as possible to the people who want to learn." There was some discussion whether or not this was giving away the farm, but the growing consensus was that education makes people more aware of information architecture processes, and through this process people who see the value of information architecture will come to hire the educators anyway and consequently everybody's purposes will be addressed. The information architect consultant will have a job and the client will have a better website -- ROI 4 IA.

flaming sculpture
flaming sculpture
float
float
general
general
heavy building
heavy building
lighthouse
lighthouse

Peter Meholz and Louis Rosenfeld - "Information architecture for the enterprize"

In the afternoon I attended this session in an effort to learn how to get information architecture a more understood activity in an institution. Meholz sees the corporate world as a set of decentralized departments. With the advent of the Internet, these decentralized departments create their own content and present it to the world. Ultimately this gets in the way of a corporate identity. Information architecture can bring things back into focus using the following formula:

information architecture goals = mission statement + business goals + branded design

Put another way, the goal is to strike a balance of consistency. The steps to this consistency include:

  1. centralized Web development
  2. building organizational awareness
  3. study of customers
  4. development of style guide
  5. implementation of a content management system making it easy to do the "right" thing

Rosenfeld believes institutions are getting smarter about their Web services because they are learning about user-centered design, metadata, and proper search engine configuration. This is resulting in less duplication of effort, consistency, branding, clear communication, shared expertise, and reduced number of re-organizations. Even though these things are becoming more and more in the forefront of people's minds, it is not possible to make (read force) it to work for the enterprise. IA must take a more velvet glove approach to providing their services.

He went on to enumerate five deadly sins of people who dislike IA:

  1. greed - "This is my turf!"
  2. ignorance - "What is IA?"
  3. slothfulness - "IA is not a high priority."
  4. fear - "Something is being taken away from me."
  5. loathing - "Just piss off!"

Similarly, the enumerated five deadly since of IA's:

  1. over-reaching - "We can fix all problems."
  2. haste - "This things can be fixed quickly."
  3. over-extending - "Working for everybody all the time."
  4. presumptuous - "This is the only way to do it."
  5. naive - "Working with no political support."

Thus, he posited a "Hyper-Evolution Model" for information architecture, a model where things are modular, offer services, work in an economic model, take marketing/entrepreneurship into account, incorporate staff issues like a board of directors, and consider timing by working with one unit at time. He suggested starting out small -- working on a staff directory for example -- and offering a "Chinese menu" of service selections. Some of these choices could include usability testing, graphic design, navigation models, taxonomy creation, etc. Some of these services would be offered for free, but others require a fee. Using this business model information architects can generate revenue and garner respect for their services at the same time. "People value it more if they must pay money for it."

Finally, he advocated the necessity of effective communication skills when talking to stakeholders. He said telling a story, as opposed to talking about the bottom line, would be the most effective way to get the ideas of information architecture across.

mascot wanabee
mascot wanabee
memorial
memorial
Mercury and Apollo
Mercury and Apollo
ominous pair
ominous pair
pansies
pansies

Matt Jones - "BBCi Search: Why search is not a technical problem"

Jones , who works for the BBC (http://www.bbc.co.uk/ ) described the creation of the BBC's latest search interface. He was given a mandate -- a strategic mandate -- to provide a search mechanism against the BBC's site and allowing the BBC to become the center of people's information universe. After looking at all the sites across the BBC he felt it looked more like Geocites as opposed to a unified whole.

Based on market research he learned about the people he was suppose to be reaching, and like the balance of the information architects at the conference, these people were called "personas." He then began creating a system based on user logs (things people has searched before), drew a lot of paper prototypes, and had a lot of usability testing done. Their final result was the creation of a system that searches many things with one query and it returns results based on the where the user is located in the BBC's site. He said the creation of the taxonomy was the key to the system's success, and the taxonomy is created by people. Thus, search is not a technical problem.

pansy
pansy
parade
parade
sail boat
sail boat
science museum
science museum
seal
seal

Sunday, March 17, St. Patrick's Day - Day #2

Lisa Chan, Bob Boiko, Paula Thornton, Samantha Bailey, Amy Warner - "Business context of information architecture in content management systems"

Chan introduced the panel discussion and set the stage for the presentation.

Boiko began by describing a metaphor he thought was appropriate for information architecture. He thought the acceptance of information architecture principles could be likened to a super saturated glass of sugar water. Given enough dissolved sugar in a volume of water, crystals will form on a string hanging in the water creating rock candy. He thought more information architecture principles need to be dissolved into institutions before the beautiful crystals will form. "Education is the key to the success of information architecture." He then went on to advocate for the implementation of content management systems (CMS). These systems allow the information architect to abstract the principles of the "craft." A CMS embodies the means for creating look and feel, navigation, and processes for content acquisition; it is a software system for creating a site.

Thornton , following-up on a previous discussion, suggested the mascot for the profession might be an ant as in informANT. In her opinion, a CMS is the latest incarnation of the data warehouse but coming together from a single point of entry. When buying a CMS it is imperative to make sure it fits with your requirements, and before it can do this you must be able to understand your problems. She also compared and contrasted activities within information architecture. On one hand there was the aspect of finding (a library process). On the other hand was the process of doing. These two things are both necessary to make a system work effectively.

Bailey was pleased with her particular CMS -- Vignette ( http://www.vignette.com/ ) -- although she did say she needed to cobble it together in order to make it meet her needs. In her opinion, a CMS can help separate content from presentation. She stressed the importance of not implementing a system that puts more burden on the user even if the user is an internal user. Otherwise the work will not get done.

Warner talked a lot about taxonomies and how they can improve the information retrieval of a site, but she also stressed much of this would be useless unless a site's particular indexer was able to take the taxonomy/metadata into account when presenting search results. She explained time (and therefore money) would be saved if taxonomies were not put into place, but this would put a burden on the end user. Which is more important? Your time or your users'? Since it is usually not an all or nothing thing, it is a good idea to tag the more important information first and work on the less important information later.

Louise Gruenberg - "Facet Analysis: Using facetted classification techniques to organize site content and structure"

Gruenberg provided an overview of S.R. Ranganathan's facetted classification method. She said it is important to understand he created the method within his particular cultural background, and his system was intended to be able to classify everything in the universe. With the advent of computer technology it became more and more possible to implement a facetted classification scheme because classified items did not have to reside in a particular physical location. Many researchers, including Vickery and Fawcett, thought the system could be used if it were not necessarily applied to the whole universe but particular problems. Attendees of this presentation then worked at designing the very beginnings of a facetted classification system.

another seal
another seal
shrine
shrine
sky line
sky line
soldier
soldier
swirling sculpture
swirling sculpture

Conclusion

I am very glad I attended this conference. It validated and re-enforced concepts I learned through previous readings and experience.

I enjoyed the preconference. A few years ago a mentor and boss -- Caroline Beebee -- seemingly could not complete a sentence about digital library initiatives without mentioning the "U" word -- usability. At that time the concepts were difficult to swallow but with time and reflection I have come to understand the value of the "U" word. The preconference provided an outline for future work.

Krug was entertaining and thought provoking. His comments successfully framed subsequent discussions throughout the weekend.

The topic of content management systems was discussed frequently. Through this conference I have become more aware of what these things are, and they sure do look like the integrated library systems of libraries. A CMS is a system of software allowing people to manage their content. The difference between a CMS and an ILS is the type of content being managed. In a CMS the content is usually Web pages. In an ILS the content is overwhelmingly books.

IA's seem to have a bit of an identity crisis in the same way librarians have an identity crisis. "We get no respect, and we are the first to go when there are budget cut backs." I believe this is true because of the intangible nature of information and knowledge. It is very difficult to put a price on these things, and consequently it is very difficult to demonstrate how they have contributed to the bottom line.

I met quite a number of people at the conference and re-introduced myself to previous acquaintances. I hope I can keep in touch.

Finally, I heard a lot about user-centered design, but I'm wondering whether or not users should be responsible for "bringing something to the party." Maybe users should be trained or educated on how to use a website. How much of website design should be centered on users and making it easy, or how much less work could I do if I taught people how to use the tool. Put another way, how do I design a site that is used by true experts and others who just want to find the information and don't want to know or care how it is found? How do I design a site where primary users have a very wide set of expectations and skills? What principles of information architecture can be applied here? I did not find an answer to these questions at the conference, but after time I believe one or more answers will become evident.

I am proud to have the phrase "information architecture" in my title.

tall ship
tall ship
tractor
tractor

Creator: Eric Lease Morgan <eric_morgan@infomotions.com>
Source: Based on personal experience; this text was never formally published.
Date created: 2002-03-19
Date updated: 2004-11-19
Subject(s): travel log; Baltimore, MD; information architecture;
URL: http://infomotions.com/musings/ia-2002/