Monday, November 26, 2007

Multi-variate Browsing for Booze

How multivariate "smart browsing" can dramatically enhance the user experience.

I've commented in a previous post about the conceptual tug of war between search and browse on the Web.

Obviously, they both have their place. What I've been noticing lately is that differences in how a site lets you browse through information can have a dramatic impact on the user experience.

Let's take a couple of sites as an example.

Corporate Wines

Yes, it's true that these aren't university Web sites, and at Chico State, booze is a sensitive subject. But in my other life, I'm a wine columnist, so I spend a lot of time shopping for (but not necessarily buying) wine.

Let's say that you had a bottle of 2002 Rodney Strong Symmetry and want to buy another. On both sites, you can get to it pretty quickly using the search box (though admittedly, the search function at is much, much better).

But what if you're looking for a 90+ point Napa Cabernet Sauvignon for under $40?

That sort of multi-variate search can be very difficult to handle via a search box. It's very easy to end up with more noise than signal in the search results. A high functioning browse structure ("smart browsing") can be a better way to go. Let's see how the sites handle it...

First of all, the Corporate Wine site doesn't show wine ratings, so right off the bat you're at a disadvantage. Second, the only thing they let you browse by is region - "California". Once there, you have to slog through 889 wines. At this point, I've already given up.

On the other hand, the site lets you refine your browsing criteria on the fly by price, type, region and rating. And the type and region filters have multiple levels, so once you select California, you can further refine your browse to only Napa Valley.

In just six clicks, I've narrowed my browse down to Napa Valley Cabs between $20-$40, and sorted them with top rated wines at the top.

I really like the ability to refine my browsing criteria on the fly, and I think does just about the best job of this of any wine Web site (though Vinfolio is just as good - and has better wines). I only wish that they had a decent wine selection to make it worth visiting their site. They should also add vintage to the list of variables that you can use to refine your browse.

OK, so how does this apply to university Web sites? I'm really not sure. This approach obviously has serious advantages when doing a multi-variate browse (e.g., price, type, region, rating, vintage), but probably has little, if any, advantage on a single-variate browse.

It might have some value where students are trying to browse through program listings. Let's say you're looking for Master's programs in the College of Engineering... in that case, maybe it would be helpful, but I'm not sure that students would be randomly looking for Master's programs.

Regardless, it's worth giving consideration to exactly how browsing your site works, and its worth realizing that browsing can depend on more than just your information architecture. For me, I'm going to stick this in the back of my mind, so that when a situation that would benefit from this approach appears, I'll be able to go, "Ah ha!"

Tuesday, November 20, 2007

The Carousel of Mystery - Information Visibility on the University of Notre Dame Home Page

Disturbing trends in information visibility and why I don't like the Flash carousel on the Notre Dame home page.

The University of Notre Dame home page was redesigned last August and was featured last month on the eduStyle Web site.

Notre Dame Home Page

A lot of people like this design, particularly the Flash "carousel" of news. Though overall I think the design is fine, the Flash carousel is exactly the thing about the page I don't like.

It's not that it's particularly badly designed, but it demonstrates a continuing and growing trend in Web development and design toward lower and lower information visibility. In some ways, this is inevitable, as more and more content and links are crammed onto pages, but my personal opinion is that whenever you hide something, you hurt yourself.

So what's my beef with the carousel?
  • How many features are there?
  • How do I navigate them?
  • Are there any that are interesting to me?
  • Is there any way to know without going through them all?
  • I'm busy, I don't have time to poke through all these to find something interesting. Why are you making me do this?
  • In fact, I'm not going to do this. I'm going to ignore it and look for the link I want.
My concern here is that I don't know what information is hidden in the carousel, and I'm not really given sufficient motivation to take the time to explore. I'm not sure that anything is served by this form of presentation. Saving space? Maybe. Enhancing the user experience? Probably not.

You could argue that it's an improvement over the random static feature used on a lot of sites, by providing a number of partially visible features, but I wonder if user testing would show to be much of an improvement.

Better, I think, to have one feature prominent, but have all the other features visible, something like the way Ohio State does it.

That way, I can see all of the features. If the most prominent one doesn't interest me, maybe one of the others will. They certainly won't if I can't see them.

The bottom line to me is: If you hide information, it might as well not exist.

Wednesday, November 14, 2007

The Future of the University Web Site

Focusing the university home page to act as a portal to audience-specific "sub-sites", instead of a catch all everything-to-everyone page.

Users' expectations of Web sites has evolved rapidly over the past few years - and continues to evolve. The concept that a "single" Web site can meet the needs of the wide variety of audiences that come to a university Web site is no longer valid. Users expect a tailored experience, designed specifically to meet their needs, address their concerns, and to fit their mental model of the site. No one "all purpose" university Web site can hope to do this - except to do it poorly.

In fact, a university's Web site is not a single site, but instead a collection of a wide variety of sites, each serving specific audiences and specific needs, and each providing a different facet of what the University is. Accounting Operations, the College of Business, and University Housing are each part of the university and the university's Web presence, but each serves a different purpose, addresses a different audience, and meets a different need.

The top level pages in the university Web site (including, of course, the home page) are no different, and should no longer be viewed as a monolithic set of "one size fits all" pages attempting to meet the needs of all users. Instead, the focus should be to provide audience-focused experiences, custom-designed to meet the needs of the main user groups that are likely to visit a university's site. These include the usual suspects of prospective students, current students, faculty, staff, alumni, family, and community members, as well as other groups.

There is no reason that we cannot provide a tailored experience for each of these groups, at least at the top level of pages within the site. And there is every reason why we should this level of experience if the university is to compete for students in the coming years.

How do we do this? By providing a home page designed to allow users to quickly self-identify and self-select themselves. Once a user identifies themselves (by clicking on "Prospective Students" or by clicking on the "Admissions" link, for example), they are taken to a sub-site specifically designed to meet their needs. Of course, it's impossible to fully guess the needs of all users, but prospective students have a very similar set of needs when visiting a university site that can be met by designing a section of the university's Web site to focus specifically on prospective students and de-emphasizes content and links irrelevant to those users.

And I'm not talking about the "clearinghouse" pages that so many universities use where each of the audience pages is simply a list of links. That's the short, easy, thoughtless and bad way out. By doing that, you're not providing each audience with a custom-designed experience; you're telling them that their experience isn't important enough for you to waste time trying to design it. "Here's a bunch of links. Knock yourself out," is about all you're saying.

Users' expectations of the level of experience they should have has increased dramatically over the past few years, and to attract those users to our campus, and to retain them, we have no choice but to meet their expectations.

In the end, we will end up with not one university site at the top level, but a series of sub-sites, each designed for a specific audience with specific common needs and perceptions.

Who is doing this? I'm sure there are a large number of universities (other than Chico State) doing this out there, but I'm just going to list the few I can find that are really investing the effort: