offline.pl

Zapraszam do lektury wywiadu przeprowadzonego z Jamesem Kalbachem, który gościć będzie w na kwietniowym Polish IA Summit.

eof: I’ve seen you give talks about ethnography. What does it have to do with designing the navigation?

jk: In particular, I’ve been developing an approach for what I call “commercial ethnography,“ which situates ethnographic techniques within business contexts.

One thing I stress in my book, Designing Web Navigation, is that the amount and type of analysis you do at the beginning of a project affects the choices you make later, including navigation design. I always say, it’s like hitting a golf ball: a small deviation up front has a large effect on the final trajectory. Ethnography helps you tee up to the ball properly.

Good navigation design reflects a coordination of business goals, technology, content, and user needs. You can’t just start designing a website without understanding the broader context. That’s where ethnography comes in: it informs the direction you take from the first step, paritcularly regarding user needs. It’s the insight that fuels all subsequent decisions, I believe.

eof: All right, there’s a lot of talk about navigation layer. I feel confused though. We do need to separate library cards from the books to find the book itself, „ontology is overrated”, „leave the taxonomies for the bounded domains” and so on. Growing amount of data-about-data does not seem to take us to the semantic forever happiness. So, what about the semantic idea in the world of separated, niche markets?

jk: Well, that’s actually one of my points: just having more metadata doesn’t lead to eternal happiness. Neither does technology. Some of the biggest challenges, I believe, are in designing interfaces that leverage semantic data and technologies to bring value to both users and to businesses.

And that’s why we need IA. IA is fundamentally about the usability of information, or the human side of information. We understand that it doesn’t matter if we can make semantic relationships technically if you can’t surface them in a meaningful way that people can understand intuitively.

What’s more, understanding bounded domains and niche markets provide the focus needed to arrive at useful information solutions. Semantic connections are best exposed within the context of a given domain. The role of an IA may be to first define the boundaries of the domain. Once established, understanding the role information plays in that domain is important. How is information generated? At what frequency? What attributes give credibility to information, and what are the signs quality? Finding answers to these and similar quesitions will frame IA solutions in information-rich and metadata-rich situations. After all, there’s no right or wrong answer: it’s all relative to the context, or in this case to the niche market.

eof: Do you see any chance for a change? Where is the place for entity extraction from unstructured web content here, and I mean not just for the commercial data mining applications? Can you tell the readers how this works and think of any practical examples? Known successful implementations?

jk: Data mining and entity extraction are cool, for sure, but they’re only part of the solution. They are the machines that harvest metadata. But you still need to process this information and organize it for human consumption.

One point I make in my presentation at the IA Summit in Warsaw is that any and all types of metadata and organizations schemes may be needed in a given context. And we’ll have to use combinations of types of metadata and types of structures. So I don’t want to suggest that the future navigation layer rests on technical crunching and semantic processing. It doesn’t. Good old taxonomies will still be needed. Tagging will be useful tool. It may all be needed. But when it comes together it ultimately must help people make sense of information in ways previously not possible.

A commercial example of entity extraction I like is Silobreaker (www.silobreaker.com). It’s a news aggregation service that exposes all kinds of relationships in up-to-the-minutes news content.

Looking at something like Silobreaker, people may be tempted to say that we don’t need IA any more. But if look closer, you’ll see a ton of IA decisions that went into that product. These may not have been made by someone with the title “IA,“ but the architecture is there: categories, navigation, labels, and page flows.

eof: What about sentiment analysis? All I can think of at the moment, is another marketing data mining tool to sell us something. Is there any way this can support the semantic web development idea?

jk: Here’s where the curse of having too much metadata can actually be a blessing: sentiment analysis works best across larger collections of information, I believe. Most sentiment analysis engines can process a single document and show you which sentences treat an entity positively or negatively. But the real value of sentiment analysis, I feel, comes when you process thousands of documents to show positive and negative distributions at a higher level.

For instance, let’s say you’re a brand manager for an automobile manufacturer and you monitor that brand’s reputation in the media on a daily basis. By benchmarking both volume of mentions and the sentiment of the brand in the media, you can see trends over time. Sometimes these trends can show issues before a story before it actually apprears in mainstream media, particularly if you’re monitoring blogs. The point is, you may not even need to read the individual stories about your company at first: you can get the bigger picture across a wide selection of sources from sentiment analysis.In fact, there’s probably too much to monitor anyway.

The navigation layer plays a role here: sentiment is just another type of metadata.

eof: What do you think the social web changes in global user navigation behaviour that we should keep in mind when designing interfaces?

jk: The web has always been social. It’s always been about communication and connecting with other people. The so-called “Web 2.0“ is really Web 1.0, in my opinion. So I don’t see a big change in user behavior in particular when it comes to navigation.

That said, I do think things like RSS and mash-up dissolve the domination of the page metaphor, for better or for worse. So I think we need to fundamentally reconsider our approach to web design. We need to focus more on the core content and functionality people are trying to get to. We need to design from the inside out. At the same, we need principles of information architecture that span platforms, devices, and media.

eof: OK, now some keywords to be explained for the broader audience – please describe what information experience is and why it is worth our attention.

jk: Roughly, an information experience is all of the actions, thoughts, and feelings a person has while interacting with information. A big part of that is the way in which information presented and communicated. People like information that is rich in detail and rich in display. We also like information that tells a story. This is all part of the information experience.

In online contexts we all too often strip information from ist native context and represent it back to users in impoverished formats. For instance, my company, LexisNexis, publishes printed books and online content. Often, we make the same content available in both formats. But if you compare the same text in both places you inevitably find much more shape and detail in the printed book than online. Attention has been paid to layout, font type and size, headers, margins, page numbers, footers, and so forth. Online, that same content is reduced to plain 12-point Verdana.

Information shape, a term popularized by Andrew Dillon and Misha Vaughan, refers to both the physical properties and the semantic properties of information. The combination of these aspects has a certain form. And this determines our experience when interacting with information.

People generally don’t want to consume information from ugly databases. As this desire becomes more and more important, businesses will have to increasingly pay attention to how and when they serve up content to end users.

eof: What do you think about „rule of thumb” or „common sense” usability? Isn’t it all about folk wisdom sourced heuristics, as Dana Chisnell writes in UXMag?

jk: I’m not sure “rule of thumb“ and “common sense“ usability are the same thing, actually. So I wouldn’t lump them together. Rule of thumb suggests having some kind of predetermined guidelines. If we use that definition, I’m not a fan of that kind of usability. I explicitely wrote Designing Web Navigation to NOT be about rules of thumb. Instead, I encourage critical thinking and problem solving.

Common sense is more about design intuition, I believe. This is where ethnography plays a potential role again. You need to have the right common sense about your users before you can start designing or even testing a product. For me, user research (not product research) done up front in a projects should primarily build a certain “common sense“ for all subsequent design decisions. So the problem with common sense usability is that you need the right common sense. Often that’s what’s missing, and we end up doing things like testing good features out of a product.

eof: What’s the situation for April 2010: are the sitemaps good or bad?

jk: Definitely good.

But a sitemap doesn’t have to be a static page with an imposing array of links dominating the page, like we’re used to. I think the notion of a sitemap is changing. For one, we’re seeing a trend of including a part of the sitemap at the bottom of each page. Flickr does this, for example. Or, we’re seeing homepages include a large selection of links that resemble a partial sitemap. There may be other possibilities too. For instance, why not merge the sitemap and the main navigation somehow? The point is, I think the definition of a sitemap transforming, and that’s good.

eof: Thank You very much and see You at the conference, hoping to conduct a video interview there!

jk: You’re welcome!

Zobacz także

Wywiad Artura Kurasińskiego z Erickiem Reissem