The Indie life: Talking with Louis Rosenfeld

Posted by

Considered one of the founders of the field of information architecture, Louis Rosenfeld along with co-author Peter Morville, literally wrote what many consider to be the book on the subject “Information Architecture for the World Wide Web.” For seven years Rosenfeld and Morville also headed up Argus Associates, one of the best-known IA consulting firms. As the “Polar Bear” book, as it’s affectionately known, goes into its second edition, Boxes and Arrows asked Rosenfeld for his thoughts on the state of IA field.

Editor’s note: We’ll also be featuring an interview with co-author Peter Morville on September 9th. (after Boxes and Arrows returns from the Labor Day weekend.)

We often see interaction design and IA compared. Let’s acknowledge once and for all that information architecture is the more difficult of the two.B&A: So, Lou, I heard that you know a little bit about IA?

Rosenfeld: I know a little about a lot of IA. Though I’m not half-bad at strategic IA, content modelling and some other IA stuff, I’m really a generalist. And that’s after doing this for the better part of a decade.

I hope that people in our field have begun to realize that there really are very many specializations within IA, ranging from thesaurus design to search engine configuration to contextual inquiry. Unfortunately, many of us seem to see IA as nothing more than blueprints and wireframes. That absolutely must change if we’re going to progress as a field.

B&A: Youโ€™re a consultant, as are many B&A readers. Do you find it easy?

Rosenfeld: It’s not the easiest job—you never know when your next project will come through. Inevitably, projects come in bunches, punctuating long periods of unsettling quiet with intense periods of insane activity. And all along, you get to pay for your own benefits and other overhead. Nope, not for everyone.

B&A: After the sad demise of Argus Associates, you now work as an independent consultant. Do you believe that working as a consultant has been more rewarding, both in a commercial sense and a personal achievement sense?

Rosenfeld: Sadly, I can’t say it’s been especially rewarding for me. Consulting is notorious for having minimal actual impact: I rarely get to see my ideas implemented, and I don’t get to be part of an incredible team as I did at Argus. The money’s fine, but that was never my motivation; if it was, I would have gone into the family business years ago.

B&A: Which would you prefer; working as a consultant, working in an agency or working in-house?

Rosenfeld: Each has its benefits and drawbacks. Personally, the most important issues are whether the job would allow me to be entrepreneurial and whether I’d get to work closely with colleagues. Being in-house or working at an agency might mean teamwork, but typically don’t make room for new ideas. On the other hand, the consultant’s life is entrepreneurial but lonely unless you go about building a company around your consulting practice.

B&A: A lot of IAs, including those who have lost their jobs, may be thinking about setting up a consultancy. Based on your experiences, what one piece of advice would you give them?

Rosenfeld: Ask yourself one question: are you entrepreneurial? Quick, do it. And answer from the gut. Don’t think about it!

If the answer isn’t immediately “yes,” then be very wary. Most information architects, and in fact, most people do not enjoy the fundamental entrepreneurial task: selling. And to succeed as a consultant, you’ll have to sell and sell and sell.

If the answer is yes, go for it: we need more people out there selling IA. That will benefit us all by raising awareness and by testing out sales, marketing, and promotional techniques that we can all learn from.

If the answer is no, don’t waste your time.

B&A: How do you “sell” IA to a potential client.

Rosenfeld: Every successful salesperson relies on a technique that fits his personality. If you’re comfortable with numbers, lead with those. If you’re a story-teller, tell anecdotes about how IA has helped others. I’m an amateur therapist, so I try hard to get prospects to articulate their “information pain.” Then I describe how IA can solve the problems they’ve already raised. So play to your strengths, though be ready with multiple techniques, as prospects respond to different approaches too.

I’ve also found it quite important to be painfully honest about what IA offers—custom consulting services that can minimize difficult problems, not a silver bullet that makes those problems magically vanish. These services take time and cost money. Good prospects will respect your integrity, and it will form the basis of a stronger working relationship. And prospects who don’t value your honesty aren’t worth your trouble anyway.

Finally, as much as I’d like to ignore the subject, you have to be prepared to make at least a soft case for the ROI (return on investment) for information architecture. That’s what most business people expect, even though much of information architecture’s benefits can’t be quantifiably measured. Try to learn what your potential client’s metrics are; these should help guide your design work, even if they don’t help you develop an airtight case for your work’s ROI.

B&A: Do you try to combine it with a general pitch that includes usability and content work?

Rosenfeld: Naturally, it depends.

Often prospects are searching for single-source solutions, so you might have to package IA with a fuller array of web design or content management services. Partnering with one or more companies that provide complementary services is a good way to do this; that’s what Argus did back in the mid-’90s.

I’m very leery of combining an IA pitch with a usability pitch. Many prospects (and, unfortunately, many information architects) don’t know the difference, and it’s important to correct this misconception. Not only is this a responsible, educational and helpful act on your part, it will also help you to distinguish yourself from usability specialists who might otherwise be perceived as competitors. Reducing the number of competitors is always a Good Thing.

B&A: What’s the biggest concern that clients have when they’re being sold on the idea of IA?

Rosenfeld: ROI. ROI is often a bigger deal for the decision-makers that your prospects are trying to sell IA to internally. And your contacts won’t do as good a job as you will at selling IA, so try your best to make the pitch yourself—in person, whenever possible.

B&A:(B&A contributor) Chad Thornton asked Jakob Nielsen what his thoughts on information architecture were. What are your thoughts on usability?

Rosenfeld: Usability is also a Good Thing. I like usability a lot. Some of my best friends are usability specialists. And I really like my information architectures to be usable.

I’m sorry to be so flip, but usability is up there with Mom and apple pie. I just get grouchy when IA and usability are mistaken for each other. Especially by practitioners. It does neither field any good, and it only confuses the market. So can we all agree to stop it already? Viva la difference!

Part of the problem is that “usability” is a term that’s easily understood, while “information architecture” is not. I’ll line up behind Peter Morville’s suggestion that information architects adopt the term “findability” as a suitable yet different bookend to “usability”. A design can be usable but not findable, and vice versa. Maybe we should start calling ourselves “findability engineers?”

B&A: Usability is finally creeping into the mainstream—especially here in the United Kingdom (where I’m from), with some of our leading web magazines (Create Online, .net) featuring regular usability columns. Do you feel that information architecture is falling behind and that it needs more public exposure?

Rosenfeld: It’s not a race by any means. Both are necessary for good design. In fact, let’s make a pact: usability specialists will henceforth lobby for information architects to be hired for their projects whenever possible. And vice versa.

And really, information architecture will benefit from the success of usability. If usability specialists are successful at pointing out design problems through testing and evaluation, designers, including information architects, will be needed to address those problems.

B&A: What can we, as professionals working in the field, do to help IA grab some of the limelight? Do we need the controversy that surrounds Nielsen’s strong viewpoints or is there something else?

Rosenfeld: Sell and sell and sell.

Woops, I already used that answer. OK, if controversy is a good publicity technique, I’ll try to make a contribution to the cause: we often see interaction design and IA compared. Let’s acknowledge once and for all that information architecture is the more difficult of the two.

Interaction design addresses a finite realm of problems. Huge, but finite, and growing in a linear fashion as new technologies present new challenges.

Besides, Jakob will undoubtedly soon come up with a small set of standard usability solutions to all interaction design problems. [grins]

But sites are being crammed with more and more content. Information is growing exponentially, and the scalar problems we’ll be struggling with in 20 years are going to make today’s IA challenges look like a casual game of tiddly-winks. Another big problem: content that’s good today will eventually go stale. A two-headed monster that’s already starting to plague many of us.

IA design conventions, if they do emerge, will quickly be rendered moot by these scalar problems. Don’t expect Google to ride to the rescue; its current combination of algorithms probably won’t hold up under the strain.

That’s why I think IA represents an increasingly larger problem space than Interaction design, and why information architects should get all the attention and the better seats at the Academy Awards.

B&A: You and Christina Wodtke have both mentioned that we should be giving away our knowledge of IA. This is a very scary thought for people who earn a living from this knowledge. Can doing this really have a positive effect?

Rosenfeld: If I publish an article in Boxes and Arrows on how to perform a card sort, who will read it? Most likely other information architects. Because of it, they may do a better job the next time they’re hired to do a card sort. Hurray! When one of us does good, we all look good.

Who else will read it? There’s a minute possibility that a potential client will, and will decide to run his own card sort exercise instead of hiring one of us. But he’s just as likely to glaze over instead, and delegate it to an information architect. Or he’ll try it, screw up, and decide to hire a professional.

As I mentioned earlier, there is an infinite number of information architecture problems. If we start to share approaches to solving the basic ones, we can all move on to the much more interesting stuff.

B&A: Do you think that IA will become less of a profession and more of a skill set? If so, where does that leave IAs—are we all destined to become teachers, educating designers and developers?

Rosenfeld: That’s a great question, and my answer is yes. All of the above. In 10 years, there probably will be people with the title “information architect,” as well as people who practice IA as part of a different role or under a different title. Not unlike our favorite analogy, the field of architecture, which includes both licensed architects and builders among its practitioners. The important thing is that much, much more IA will be practiced in the future unless some major catastrophe engulfs the planet.

B&A: So who will practice IA? Who should learn the IA skill set?

Rosenfeld: I think a better way to look at this is to ask who is practicing IA right now? Just about everyone who deals with information systems, even managing files on their own computer, knows a thing or two about information architecture. Perhaps an important way to broadly promote ourselves is to associate the frustrations of personal information management with solutions that emanate from this new field of information architecture. If everyone realizes they have IA problems, they’ll value IA experts all the more.

B&A: Is the future in consulting, or should in-house information architects be the better approach?

Rosenfeld: Again, the future is in both. IA generalists will likely predominate in-house, with specialized consultants being called in on an as-needed basis. But this will naturally vary by company size, industry, and other factors.

B&A: Do you see content management systems (CMS) as a threat to information architects? If the IA is defined by the CMS, where does the in-house IA fit in after the system has been designed and implemented?

Rosenfeld: No, CMS present a golden opportunity for information architects. True CMS (as opposed to document management systems) require schema design, not to mention content modelling and workflow analysis. They often require metadata to support searching, browsing, and administration. All this design will have to be customized to meet the needs of a particular set of users and business owners. CMS vendors’ professional services groups usually can’t handle this melange of custom design and evaluation. That’s where we come in.

B&A: When you meet someone and they ask what you do, what do you tell them? Do you have a prepared line that goes “Well, information architecture is basically…”?

Rosenfeld: I usually point out my wife, tell them to ask her, and make for the devilled eggs. Mary Jean does a much better job of explaining IA than I do. If she’s not handy, I slog through something along the lines of “I organize huge amounts of information on web sites so that users can find what they’re looking for.” Or something to that effect. I find that elevator pitches aren’t usually enough; combining them with an analogy or two helps, assuming your conversation partner is still interested, or ever was in the first place.

B&A: You’ve got a treat for us all that’s due out soon. Can you tell us a little bit about Polar Bear 2?

Rosenfeld: It will be on bookshelves at the end of August, I’m happy to report. It’ll have the same bear on the cover, though she’ll be a littler older and fatter, and hopefully wiser too.

B&A: You’ve already told me that it’s going to be much bigger than the first edition.

Rosenfeld: Yes, it’s about 500 pages, two-and-a-half times longer than the first edition.

B&A: What has changed so much since it was first published in 1997?

Rosenfeld: You, mostly. There really wasn’t an audience of self-identified information architects back in 1998. Writing for a smarter, more experienced audience meant striving for both greater breadth and depth, and lead to Peter and I rewriting almost the entire book (it’s really unfair to call it a second edition). Not an easy task, so yes, we’re relieved it’s done and we can return to leading normal lives.

B&A: What additional topics are being covered in the new book?

Rosenfeld: More of an interdisciplinary perspective, more on the relationship of IA to the broader business context, and lots more useful resources for practitioners. More sophisticated methodology, more case studies, more visuals. More things that I’m sure I’m leaving out. In general, more.

And yet, still not enough; the field is so broad now that it’s really impossible to stuff into a single volume. Disagree? Go ahead and try it yourself.

B&A: Could you have predicted the changes and additional topics five years ago?

Rosenfeld: I’m not sure we could have predicted all of these changes, but we always knew that the field would take hold, grow, and become increasingly complex.

Our book is but one snapshot of information architecture circa 2001/2002; we’re really looking forward to seeing (and reading) many more perspectives.

Paul Nattress is a former UK-based information architect who specialises in web content and has worked in-house for several UK FTSE 100 companies.


  1. So I catch the tongue in cheek aspect of it all but am still wondering about compare between IA and interaction design. Each field presents its own set of challenges and difficulties. Since we’re all supposed to be working together, it seems more productive to respect each other’s unique contributions regardless of specialty.

  2. Content will scale. True.

    The organization and ‘findability’ of such content will most likely become a more complex task. True.

    But the interface(s) that proliferate the results and launch the acquisition of such ‘findability,’ will result through a more complex design task (as opposed to a complex interface). These future widgets, allowing for the exploration/interaction of such scaling content and functionality, have not been designed yet (i.e. look at the historical development of web interface features). Will these interface components be designed by a pure LIS IA? Possibly, but most likely not. Could they they be designed by an IA/ID or ID or an engineer? Most likely.

    My point here is that no matter where new media/software development takes us in the future, IA and ID are joined at the hip. To say one will be dealing with more complex issues than the other is being a bit haughty.

  3. …or a good way to promote discussion.

    I agree; it’s highly unlikely that an LIS IA probably won’t develop the new widgets we need (though they might help).

    But do you think many such widgets really will be developed? I’ve limited my thinking on this to the familiar realm of 2D digital spaces; should we include more futuristic approaches (e.g., holographic systems, jacking directly into one’s brain) in this discussion? If we do, I’ll concede that plenty of ID innovation is on the horizon. If we stick to web browsers and similar, I still have a hard time seeing how the ID problem space will change all that much.

    And yes, IA and ID are joined at the hip, but they’re definitely NOT the same thing. I might also add that IA and UE are joined at the hip, as are IA and CM, as are IA and KM, as are IA and organizational psychology… Not trying to be flippant; just want to point out that there is a reason for all of these differences, even if there is much overlap. Part of my hope with this discussion is to remind people that those distinctions are important.

  4. This interview was great! I always love reading new articles on this site and Louis Rosenfeld material. My favorite part of this interview was, “Maybe we should start calling ourselves ‘findability engineers?'”. Although it doesn’t sound as elite or glamorous as “information architect”, it definitely gives normal people a better idea about what is being done. Everyone could relate to that because they have all been to sites that confuse them and where they can’t find things.

  5. There is plenty of ID innovation horizon within and outside of the browser environment; both now and in the future.

    But let’s focus on a specific case; take a controlled vocabulary/thesuarus for example. The development of the CV is a rich and consuming practice, and can be performed on various levels depending on requirements and the database structure in place. But the intriguing aspect of the CV is that once it has been developed and introduced to a database environment, the interface possibilities for it’s implementation are practically immeasureable.

    Are we to say that unless we’re talking about “holographic systems, (or) jacking directly into one’s brain” that innovation is not possible? Just take a look at what Plumb did with Thinkmap. And that was 4 years ago. Are we conceding convention? Are there not cutting edge interaction design solutions remaining for CV’s/search engines within the web environment? I think yes, there are.

    Reminding people that distinctions exist between professions/roles is a positive thing. But if I said, “C’mon folks, usability engineering has no weight in comparison to the complexities of interaction design” to “promote discussion,” I would be wrong in doing so.

    If I’m alone on this, so be it.

    BTW, I’m still planning on getting your book, because I do respect your knowledge as an IA. But now, I’m going to steal it from B&N.

  6. Hey, while you’re at it, take one of the books that you *don’t* steal and turn its cover to face out. (Old authors’ marketing trick.)

  7. Sure, I think I could provide that service. But if you get a sudden spike in sales in the metropolitan NYC area, I want to be handing the bear a coke on the next cover.

  8. Lou – Great interview and congratulations on the 2nd edition. I can’t wait to get my copy from O’Reilly any day now. A few years back when I read your first edition I found myself thinking, “Hey, that’s what I do.” So why didn’t anyone call it IA then? Well now everybody (or at least many more people) do and I looking forward to reading the new edition and learning about what’s been happening to me since.

  9. If you’re living in a world of content-centric Web sites, it’s true that IA is considerably more complex than ID. However, if you’re working on functionality-centric applications, just the opposite is true. As an interaction designer, information architecture seems like a pretty simple problem. However, that’s only because I work on applications with limited content and therefore only encounter simple IA problems. Structure, behavior, and presentation are all equally important and equally difficult problems although they are balanced in different ways for different types of sites and products. To say that one is more difficult than another is like saying its harder to be a pianist than a guitarist or a drummer.

    On a different tack, Lou I’m curious, if pushed for a one word description of your “occupation” what would it be? Designer, librarian, architect, other?

  10. Information architect. Or, if you were going to push me for what I did before becoming an IA: librarian.

  11. Paul, there you go with the questions again…

    “Best set of skills to do IA?” Don’t think there is one. There’s perhaps a best mix of skills, with balanced coverage of techniques that help us learn about users, content, and business context (picture those as three circles in a Venn digram).

    One may know a lot about content from a background in librarianship, but may benefit from learning about users through HCI and context through studying organizational psychology. Or one may have a background in technical communication (content), with competencies in ethnography (users) and context (management).

    This three circle approach can guide one’s efforts at building a good educational foundation to practice IA. It can also help in the recruiting of IA staff and assessment of IA vendors.

  12. I feel obligated to point out a blatant falsehood in the interview: Lou doesn’t like devilled eggs.

    Happy to serve the IA community in whatever small way I can.

  13. Hi everybody!

    Lou, I really liked reading your interview and I’m looking forward to reading what you and Peter have to say at this point in the IA timescale. As you know Lou I’m an IA turned ID, but working in ECM I am always a little of both. The question of scalability between content vs. functionality is an interesting one to me b/c well it is my universe. What I am coming across most recently is that there is scalability in functionality that has a comparison to scalability in content. Finding functionality is as much an issue as finding information is, if not more so, b/c first I have to find the functionality that allows me to find the information. ๐Ÿ˜‰ But I do think you are right, that the curves if graphically represented would not be equal. in terms of the # of objects that are part of that scale, but the # of possible solutions is definitely on par, which means the practitioner has an equal amount of sophistication.

    I recently read “Interactions” (the SIGCHI magazine) and Aaron Marcus wrote that he feels that IA is just UI Design (so is ID, and ED and usability engineering) and he feels that all these added terms are just confusing and unnecessary except as a means to describe components of UID. I was wondering what you thought about this direction.

  14. I didn’t read the article, so I really don’t know what his take is. Perhaps Aaron was just trying to ruffle a few interdisciplinary feathers like I did in this interview. Which, if it results in discussion and learning, is ultimately productive.

    I would say that IA is by no means two-dimensional. Much of it happens beneath the surface of the UI. Aaron may or may not realize this. Or, again, he may just be stirring the pot a bit so we can all have interesting conversations like this one. ๐Ÿ˜‰

Comments are closed.