|
If
you want to know when new interviews go online,
subscribe to the WebWord.com
Usability Newsletter!
The
Dynamic Duo of Information Architecture
An Interview
with Peter Morville (President, Semantic
Studios) and Lou Rosenfeld (Louis
Rosenfeld, LLC)
Conducted via email by John S. Rhodes
(18-October-2002)
How has information architecture (IA) changed in the past few years?
Any new discoveries? Any exciting research?
Peter Morville:
The first question has many answers because IA has followed totally
different trajectories in different organizations. Leading
technology firms such as HP, IBM, Microsoft and PeopleSoft have taken a
holistic perspective that connects IA, content management and user
experience design to real business goals and metrics.
On the other hand,
I've seen a dangerous trend toward reductionism in several old-guard
Fortune 500 corporations, where IA components are being created in
ignorance of the big picture. For example, I worked briefly with an
international, cross-departmental team earlier this year that had been
assembled to create an enterprise taxonomy. They had no idea or
interest in how this taxonomy would actually be used within the broader
intranet environment, and they didn't really care how it aligned with
other closely related initiatives. I fear the inevitable backlash
against IA when these companies discover their next generation web sites
and intranets are composed of small pieces badly joined.
As far new
discoveries and research, no single change comes to mind. What's exciting
is to see the many small changes in the world of the web and to explore
how they might fit together. Blogs, google, google news, social
software, topic maps, ambient devices, web on the wall, nanotechnology and
MEMS...all of this stuff will impact the future practice of information
architecture.
When does search work best? When does it fail? How does information
architecture fit into the search picture? Exactly how do IA and search
relate?
Lou Rosenfeld:
I'll start with your first question and work backward from there. Peter
and I see IA as being about findability, the same way usability
engineering is (naturally) about usability. Users find information
by browsing, searching, asking questions, often employing a combination of
these approaches. An information architecture should ideally support
the configuration of approaches users employ when finding information. So
really, search fits into the information architecture picture, and not the
other way around.
When we're
designing, it's important to consider this bigger picture of finding,
rather than focusing on something more specific, like search. Otherwise we
often end up with imbalanced designs, such as a robust search system
coupled with a puny browsable taxonomy. Or we don't integrate our
architecture's searching and browsing components, which is too bad:
users often want to use these *together*, and the sum (finding) is often
greater than the parts (searching, browsing, and asking).
So when does
search work best? When there are three conditions in place:
1) Users
who know what they're looking for and know how to articulate what it is
they're looking for; information scientists refer to as known-item
searching. For example, searching the staff directory for Vint
Cerf's email address is likely to be successful because you know where to
look, you know that the directory will provide a phone number, and you
know how to articulate your need (assuming you can spell Vint Cerf
correctly).
2) Huge
volumes of content that afford high levels of structure and/or metadata.
For example, if I can search an index of names in that staff directory,
I'll probably have a better chance for success than if I search the full
text of each entry in that directory. Structure and metadata are
tools we use to expose the answers that lie buried deep within our
content. Essentially, the more we can make our content--which, on most
sites, is primarily textual--behave like databases, the more successful
our searches will be.
3) An
economic incentive to spend money to make search work. Realism
always manages to rear its ugly head, but yes, if there isn't a compelling
business reason to invest in making the content more accessible to users,
there will be fewer resources to devote to making search work well.
So, search works
less well with unstructured textual content that is not especially
strategic to the organization. And it works less well when users
have open-ended information needs, in which case other techniques for
finding information may be more appropriate.
Why is it so hard to "sell" information architecture and
usability? If information architecture and usability are so wonderful, why
do they need to be sold at all? Are designers, project managers and even
CEOs total idiots? What is going on?
Peter Morville:
If you take a ten year perspective, we've collectively had tremendous
success selling IA and usability. There are far more people who get
paid to do this work in 2002 than in 1992.
Admittedly, it's
tough to feel good about the big picture when we're experiencing the pain
of a recession that has wreaked havoc throughout the whole IT industry.
It's hard to sell anything in IT right now.
We shouldn't give
up but we must focus our energies. We should steer clear of lost
causes where there simply aren't compelling business cases to be made and
concentrate our efforts on the low-hanging fruit.
For example, using
controlled vocabularies to improve searching through synonym mapping can
provide huge bang for the buck when focused on the most important and
common searches. To pick on HP's web site for a minute, it seems
absurd that a search on "journada" is not mapped onto
"jornada." This is a common misspelling of a key product
name. Given that HP already has a mechanism for supporting best bet
recommendations, the cost of tackling synonyms would be tiny.
Is it possible to automatically generate an excellent information
architecture? If yes, what tools are available? If no, why does IA demand
so much human energy and attention?
Lou Rosenfeld: No,
it's not possible to automatically generate an excellent information
architecture.
OK, let me step
back from that statement, but just a little. There are aspects of a
site--a shopping basket system, for example, that may be specialized
enough to come out of the box. But when you compare the
architectures of Dell and Gateway, you'll find plenty of differences, even
if they're in the same business. There are simply too many variables
in content, users, and business context to assume these sites could have
the same architecture.
Other
architectural aspects may be automatically generated, as the folks at
vendors like Autonomy will tell you. What they didn't used to tell
you was that you'd have to "train" the software to work well,
and in many cases you'd need a custom taxonomy to get the software to
work. Even if such tools get you 80% of the way there, which is great,
you'll still have to deal with the gnarly architectural problems that make
up the remaining 20%. In other words, you'll be doing something manual and
custom.
Shrink wrapped
architectures don't even work for something as specialized as blogs, which
come with templates that essentially constitute a default architecture.
For those of you who blog with tools like Movable Type, ask yourself how
much you've configured and customized your blog when compared with the
default templates. If you haven't, I'd bet you'd like to.
Maybe in twenty
years, we'll have many more useful IA conventions, much like we have the
wonderful architectural conventions of books and magazines in the world of
print content. But still, automatically structuring, organizing, and
labeling content is essentially an artificial intelligence challenge, and
like much else with AI, tools can get us part way there, but manual
efforts are needed for that tough last mile.
What information architecture differences are there between small web
sites and large web sites? Does IA even matter for small web sites?
Peter Morville:
On very small web sites, the impact of IA on brand identity outweighs
issues of findability. Good organization and clear, consistent
labels present a positive image of the person or company behind the site.
However, sites don't have to grow very large before findability becomes a
key usability problem. The web is heavily populated by 100 page web
sites that frustrate users' attempts to find what they need. These
sites may not merit a professional information architect, but they do need
help from someone who understands the basics of information architecture.
You have written about brand loyalty and information architecture a
couple of times recently. Can you explain how IA generates and maintains
brand loyalty? Is this just fluff or does it *really* make
difference?
Lou Rosenfeld: Providing
easy access to useful information is the primary goal of many of our
sites. On the Web, competition is brutal. One way our sites
can stand out is by achieving this goal of findability better than their
competitors. Amazon does a better job than Barnes & Noble at
helping me find titles easily and suggesting relevant new ones.
That's why it has my loyalty and repeat business. But if Amazon ran
physical bookstores, I might prefer a Barnes & Noble because the
coffee is better and the parking more abundant. But we're talking
Web here, and findability makes a huge contribution to the brand value of
many web sites. In fact, IA can provide a stealthy competitive
advantage, as people often notice it only when it stinks.
Can I prove the
brand impact of IA? Not really; measuring the factors that
contribute to branding is very difficult. But common sense seems to
indicate that findability is likely a large component of online
branding.
I care a lot about usability. When I travel and when I am in new
environments I notice usability problems all the time. Does information
architecture impact your life in a similar fashion? What are your IA pet
peeves? Do you think about IA in your personal life?
Peter Morville:
Yes. Wherever I am, there's IA :-) My major pet peeve relates
to the tremendous economic and psychological losses we experience as a
society due to bad street signs. How many minutes per day are lost
due to missed exits? How many accidents are caused by panicked
drivers who went the wrong way? This is why I love Richard Ankrom's
"Guerrilla Public Service"
initiative to improve signs on L.A. freeways.
More than three years ago, we did Information
Architecture Revealed! What do you think about it now? How did you
feel then versus now?
Lou Rosenfeld: An
old Web-based interview is like your crazy aunt who lives down the street:
she crops up from time to time to embarrass you in public, and there's
nothing you can do to get rid of her.
Really, aside from
some wrong ideas I had about usability, and now painful mentions of two
dead endeavors (Argus and the Clearinghouse), reading it wasn't as bad as
I'd imagined. Some of the ideas actually weren't so far off.
I hear so many Web
designers discuss search and metadata these days, so I know that people
with library backgrounds have indeed made an impact on the industry over
the past three years. I'm more concerned than ever that information
architects, as well as others in the industry, really don't design for
users' broad information needs. And I'm no less pessimistic
regarding the future of search engines.
Overall, I'm
thrilled that the field of information architecture has matured so much in
such a short time. There's more recognition of and respect for the
field, more community among its practitioners, more books coming out on
the topic... And it's more obvious than ever that our future is
bright. Yep, the economy is awful right now. But I can't see how one
can't be wildly optimistic about information architecture's future.
Some people already know that I love your book, Information
Architecture for the World Wide Web. How is the 2nd edition different than
the 1st edition? What's new? Why did you decide to write a second edition?
Can you offer WebWord readers something from the book that can't be found
any other place on the web? Maybe a free book chapter, or something else?
;-)
Peter Morville:
Well, for starters, you need to buy the 2nd edition because the 1st
edition is totally wrong. Just kidding! Seriously, the 2nd
edition is a much better book. It's almost twice as long as the
first and is packed with the lessons we learned while working on some of
the world's biggest and hairiest web sites and intranets. We wrote
it because we're getting old and don't trust our memories. Also, we
wanted to drive home the message that information architecture has
survived the dotcom implosion and will be around for the next 10,000 years
or so. Finally, we wanted to give something back to the IA and
usability communities, and to the people who have supported our efforts
through thick and thin. In this spirit of generosity peppered with a touch
of enlightened self-interest, we offer a free
online book chapter, available exclusively from WebWord.com.
Any last minute thoughts or comments?
Lou Rosenfeld: Just
something we'd like to point out about our book: there are a lot of
differences between the two editions, but perhaps most important (beyond
sheer volume) is our focus on business context. This is a direct
response to what we've encountered in our own consulting and what we've
learned from our peers: focusing on getting content to users is an
almost pointless pursuit if you don't figure in such factors as resource
constraints, organizational culture, and politics into your design.
Maybe in a few years management gurus, organizational psychologists,
change management experts and operations engineers will have joined the IA
realm and will revolutionize IA practices around issues of business
context. In the meantime, we hope the relevant material in our
second edition will be of use.
Thanks very much
for the opportunity to be interviewed John; as always, excellent
questions!
Thank you Peter! Thank you Lou!
|