you want to know when new interviews go online,
subscribe to the WebWord.com
Face of Information Architecture
An interview with
Christina Wodtke, the mighty woman behind Elegant
Conducted via email by John S. Rhodes
Why is there a
need to architect information? That is, why does everyone assume that
information needs to be structured and categorized and formatted? Are there
any problems with this core assumption?
It's a human
tendency to organize items for retrieval. Look around your house; you've got
a sock drawer and an underwear drawer, and a t-shirt area to promote easy
retrieval of clothes in order to get dressed; perhaps you have a silverware
organizer to separate spoons from forks to facilitate ease of table setting;
perhaps you alphabetize your CD collection so you can find the band you're
looking for. Think about bookmarks (or favorites, if you use Internet
Explorer). At first when you saw a site you liked, you probably just clicked
on bookmark, not worrying about organization. After a while, you probably
started to have so many bookmarks, you had to create some sort of
organization: work and play perhaps. then organization had to become more
granular when those categories because big: work because html, inspiration,
usability and competitors perhaps.... Finally, most people I know got too
many bookmarks and too many categories and can't keep up with organizing and
go back to simply using Google.
It's a simple fact
that once you have a certain amount of stuff, you have to organize it or you
can't find anything. Information Architecture does that for information
gathered on a given website.
To answer the
second part of your question: I'd say the only problem with the assumption
that information has to be organized is that sometimes designers (and
by that I mean the broadest sense of the word-- graphic design, IA,
programmers) often jump into structure before its necessary. Have you ever
clicked on a category link and found only one item in it? Pretty annoying. I
think most of the time it's better to keep an undifferentiated mass on one
page than to jump the gun and over-organize.
One thing I've
learned from usability testing is that humans are remarkably comfortable in
a mess-- pages you would swear were examples of information overload turn
out to be more usable than a simple page with a few choices on it.
Of course user
testing can tell you where that breaking point lies.
What are the main differences between doing information architecture for
small sites versus large sites? What is "large", what is
"small", and how does that change the nature of the game?
architecture on a small site is pretty straightforward, and anyone with some
training can play the role of information architect: a designer, a producer
or an engineer. On a small site typically only a site map and page layout is
needed. A big site might contain thousands of pages of data, and need a lot
of work to make sure people can find what they are looking for-- much like
the example of bookmarks I gave before. There is a point where organization
becomes more and more difficult, and retrieval more challenging, and then
you have to depend on more sophisticated methods whip the information into a
malleable form: hard-core Information Architects talk about taxonomies,
controlled vocabularies, thesaurus thesauri. They have a lot of tools to
make the site useful to the visitors.
What's big and
small is hard to determine at first glance. A site might have thousands of
pages of new articles, but they might be based on just five templates,
making it a small site for design and a big job for information
architecture. Or a site might be a web application and have only five
screens to design, but have dozens of possible configurations, with
attendant error messaging and contextual help. Big and small mean less and
less as the concept of a web "page" slowly disappears.
Exactly how does information architecture apply to web application
design, versus doing information architecture on the content of a site? What
are the important things to keep in mind?
There are two
schools of thought on this. Some folks would say that the two are not at all
the same. They would suggest that you need an interaction designer for an
application, and an IA for a content site. But the other school of thought,
which I subscribe to, suggest that in reality these two are never quite so
neatly separated, and you need an information architect who has the skills
of content architecture, interaction design and information design.
You ask about the
important things to keep in mind: same as ever, with a task-based twist:
what are the users trying to accomplish, what does the business need
them to successfully accomplish, and what will the technology allow? If you
can balance these three forces, you'll have a solid product.
That said, there
are tools that are used for certain types of problems: if you architecting a
content heavy site, you will want to do a card sort. If you are doing an
application, you will want to map the user flow, or do a use case. A
taxonomy will probably we useless on many applications, and task analysis
may prove invaluable. I suppose it's a bit like being a plumber: sometimes
you need a wrench, sometime a whack with a hammer does it. It's pretty
important for an IA to know their tools, and when it's the right time to use
Information Architecture, Some Details
What do you
think about web services? How do you define them? Do you think information
architecture will be important in that space? If so, how?
It depends on the
web service. To start with, web service is hard to define... think of an
online calendar: is it an online service, or a weblication? The lines are
My experience is
that IA is very important and incredibly challenging to web services. I
worked on Snapfish.com. It is a place where people send in their film to
have it developed and scanned, then the film and prints are returned to
them, and their pictures are online to be viewed and shared. The IA there
was incredibly challenging: we never knew our content: what sort of pictures
would people have? how would they organize it? Would they keyword it? (of
How could we
create organizational structures for content we didn't know anything about?
And once you've got six or seven rolls of film, you definitely need some
kind of organization scheme. "Where was that picture of fluffy with the
red bow?" Services are fun, because your design partner is the end
user, and users can be an unpredictable bunch.
Jesse James Garrett states that information architecture is the
structural design of the information space to facilitate intuitive access to
content. Can web sites, applications, and services ever really be intuitive?
After all, what is intuitive to me might not be intuitive to you.
That's not a
leading question! <grin> Don't you want to ask Jesse that?
<sigh> Okay, in my opinion, intuition does not exist. What we have
instead is a huge database of all of our experiences in our brain, and when
we interact with an interface, the terrific instrument that is our brain
checks all those experiences and makes an educated guess about how to deal
with the new interface.
I've been in user
tests where the person testing the interface simply shakes their head and
says, 'I have no idea what to do'. They have no experiences that match what
they are seeing, and their brain could not intuit how to use the site.
So to make an
"intuitive" interface, one should look at the experience base of
the typical users and then try to leverage what those users know in order to
then design an intuitive system. That's why illustrator has the pen tool,
and Microsoft word a highlighter; that's why AtomFilms has a thriller and an
extreme section. These are concepts that their respective audiences have
already be introduced to, and their audience has certain expectations when
dealing with these items. A pen tool says "draw" a highlight tool
says "make a transparent yellow foreground," extreme says
"sex and violence." Why reinvent the wheel if there's already
something that works well?
How does structure interact with context? For example, let's say that we
build a web site using all of the latest and greatest information
architecture tools and techniques. How will I know that I have developed a
structure that is flexible enough to handle the nearly infinite contexts? In
general, how is variation handled?
You're making my
brain melt. Not because there isn't an answer, but mostly because there
really isn't a short one. I'll do my best.
First: the best is
the enemy of the good, trying to be perfect or trying to meet infinite
context will both make you crazy and make you miss your launch date. The
only thing to do is to create a product for your users. Not "The
User" but the specific audience for whom your product is intended. Too
many insane dot-coms thought they had the same audience as Amazon.
e-greetings might have. e-pets didn't. e-pantyhose definitely didn't.
Second: you can't
just use information architecture techniques-- you have to add in
user-centered methodologies. Personas are a wonderful technique for
prioritization of user groups. There is a nifty little system we've
developed at Carbon IQ where we start with market research and contextual
inquiry, create personas, run scenarios, break them into task analysis...
then go back to the inquiry data, design cognitive models, align the
cognitive model with the task analysis and create the structure from that.
It's actually faster than it sounds, and creates a very solid targeted IA.
Third: 'How do I
know I got it right?' You know this one, John: test with your potential user
The Ultimate Things
What is the
best example of information architecture you have ever seen? Think broadly,
beyond the web. Now, why did you choose this example? What is the key
I never think
beyond the web. <grin> I'm still fond of Wurman's definition:
"the individual who organizes the patterns inherent in data, making the
complex clear. It's a person who creates the structure or map of information
which allows others to find their personal paths to knowledge, and it's also
the name of the emerging 21st century professional occupation which
addresses the needs of an age, focused upon clarity, human understanding and
the science of the organization of information."
That gives me a
lot of room to play. I entered this profession because I wanted to make
technology better for humans. Wurman contrasts the human need for clarity
with information's tendency to overwhelm, and he states that somebody needed
to intervene. It makes his definition very wonderful.
What is your description of the ultimate information architecture job?
Describe the company, team, tools, and work environment. What about the
position would make you feel most valuable?
I'm pretty happy
at Carbon IQ: I've got brilliant partners who challenge me, and I'm doing
what I love, which is IA and user research. However I do occasionally
fantasize about the dream project: I'm brought on as CXO (chief experience
officer) for a big decentralized company with a wealth of products, both web
and real world. I stay on for one year to make their website work. I
evangelize the importance of a user-centered process, I deal with the
politics that comes from decentralization, I hire a crack team from the
hordes of talent that are on the streets right now, I choose a toolset that
suits the workflow.... my dream job is always a worst case scenario, because
challenges keep you from getting bored, and I hate being bored.
Final words? Last comments? Did I miss anything? What should every person
remember about this interview?
Architecture is a new science, based on old principles. Lots of folks
beginning to learn IA look to me for rules, like "a menu should only
have seven items" but that's a lie and lazy, and will result in bad
work. You don't want rules. You want to sharpen your mind.
Observe the world,
observe people interacting with all the things that have already been
designed in the world. And ask yourself "why is this like this?"
and "how could it be better?" The more critical you are of the
stuff that surrounds you, the more solutions will start appearing before
If I had one thing
I wanted people to remember, it's that technology is here to serve people,
not the other way around. When I hear "stupid users" I cringe.
Everything you design will be used by some poor human being. It's good to
try to love that end user, and build something good for them, they way you
might for your mom or your husband.
Of course it's
particularly useful to know who that user is: I don't think I'd build the
same site for my mom that I would for my husband.... same reason you don't
build the same site for Schwab as you would for Pokeman. But I would love
that investment banker just like I'd love that five-year-old screaming
"pikachu!" and try to make a pleasant experience for each.
carbon IQ user experience group http://www.carboniq.com
tel: 415 824 7090 fax: 415 824
(1) I've been
reading the Elegant Hack
weblog for a long time. Check it out!
(2) Christina is
the face of
-- John S.