jhardernet

The newest edition of A List Apart is online and there’s an interesting article by Patrick Lynch about visual design and usability.

Lynch points out that although eye tracking studies seem to show that aesthetic graphics (as opposed to information graphics) are irrelevent, that is not necessarily the case.  The brain processes these images and uses them to make inferences about an interface’s personality and usability almost instantaneously, imperceptable by the eyetracking studies. Aesthetics graphics matter in forming a gut reaction.

It’s that visceral reaction that I’m hoping to improve with the recently launched homepage test.

Wow. Quite an exciting week for me as yet another test that I designed has gone live. This one redesigns the Women’s Interest category pages. Two weeks ago a redesign of the Men’s Interest category was tested. Although I put that test together, it was in protest as I did not have any control over the design decisions being made.

Here’s the old control page, and the ugly Men’s Interest redesign (I could go on for hours about how bad of a page this is). Now, here’s my Women’s Interest redesign. This one was pretty easy, as I just tried to follow best-practice genre conventions of category pages: Easy sorting options, paginated views with product photos, and a list view without covers.

The biggest change may be removing the “Buy” buttons. On the one hand that means a user has an additional click to get through to a detail page where they can order. My theory was that someone on a category page is more interested in browsing and is still in the decision-making process, so the added clicks aren’t an interference.

Time will tell - this test has only been up for a couple hours. So far it is converting at 28% vs. 0% Go test, go!

As for the new homepage design, after 24 hours the conversion rate is slightly higher on the test page but average sale is substantially higher on the control. Also, bounce rate is up on the redesign which is a real disappointment as that will be our main indicator of success for this test. My plan is to tweak the redesigned page a bit to make the product layout a bit less overwhelming. As long as the conversion rate isn’t falling off then there’s hope.

More updates to follow.

An excerpt from the homepage test

Today at work I was given the greenlight to start testing my “radically different” homepage design.  I’d like to say this design is based on a bunch of user research, or some sort of participatory design methodology but - it’s not. It’s pretty much just me making some guesses.  At this point I’ve stared and tweaked for so long that I don’t even know what to think of it anymore. This is the point where I wish there were other designers here, or a culture of usability testing to get more feedback.

Here’s the test page and the control.  If they look the same you’ll have to clear your cookies and reload the page because it will have pulled you into the test. Check them out and then I’ll break it down.

What (I think) it gets right:

1. Fluid layout accommodates different resolutions while maximizing use of screen space. A much bigger improvement over the former table based layout.
2. Text wherever possible and graphics only when necessary make improvements in accessibility.
3. Graphics and colors have been updated and whitespace increased for a more modern look, while still in keeping with the brand image. My goal was to make it look current and friendly.
4. Pared down the navigation, getting rid of little-used buttons at the top of the old site, and decluttering the left nav bar.
5. Inclusion of Twitter stream to show that there are real people here working hard and to generate some social networking mojo.
6. Took two small search boxes and replaced them with one big, centrally located box; also, focus is on the search box so you don’t have to click into it before your start typing.

What (I think) could be better:

1. Not sure on the scrolling feature box.  Could be clunky to use and create alignment issues when re-sized to an extremely narrow width.
2. The Mags by category, combined with the scroll box, create a big wall of  products that may be a bit overwhelming.
3. Something about the design still doesn’t quite say “E-Commerce website” to me. Not sure if it’s the color scheme or what.

There it is. Traffic is getting split 50/50 between this page and the control. So far, with approximately 30 visitors each the test page has 3 sales and the control has 2. Obviously it’s too early to tell.

Hopefully this test either wins big or goes down in a blaze of glory. The worst, yet most common theme is for the conversion rates/sale to be even in which case you’ve proved nothing without learning anything.

I’ll keep you (my imaginary readers) posted on the results.

So. This is it. My new website, complete with portfolio, resume and blog.  But why? What is the raison d’etre?  Potential future employers is one audience; friends and strangers who have a reason to Google my name another.  And, if I write this blog correctly, then I may have a wider appeal to designers/researchers in general.  This entry will get the ball rolling, but in future entries I hope to reflect on triumphs and tribulations I encounter in my work.

In particular, I’m interested in figuring out how user research can influence (Web/software/UI/etc) design decisions.  How can we get the best information from users? How do we know what even is the best information? This involves drawing distinctions between things like user-centered design, and participatory design. Sometime these terms are used interchangeably, however they are different.  User-centered is a more generic term applicable to any situation where user data is involved while Participatory design means that at some point users and designers are actually in the same room working together.  Different situations call for different methodologies, but I tend to favor some face time at some point in a design process.

So that’s it for now. The ball is rolling and I’ve greeted the world.  I’ll be back in the next post hopefully discussing more specific issues.