dreadedmonkeygod . net

Mark Hurst of Creative Good gave an excellent talk on usability testing to start off the second day of the Endeca Discover conference.

He picked two people from the audience, and had them try to pick a cable TV plan. Both participants were excellent test cases: the first had just bought a complete HDTV package, and the second barely watches TV at all, but needs fast, fat internet access.

A few key take-aways, in no particular order:

In searching the various cable TV providers' sites, neither participant used the sites' built-in search functions. Both bailed to Google when they couldn't find what they were looking for by clicking on navigation elements. (This jibes with my rule of thumb that site-specific search functions always suck.)

As one audience member observed, "There's no conversation going on." The providers' sites were very much centered around pushing customers to buy one of the pre-packaged plans provided. They put these plans front and center, with the result that both participants felt like they wanted to do more research before buying. Imagine walking into a luggage store, and having the salesman say, "These are the best three bags in the store! Pick one!"

And despite both sites featuring a three-column, side-by-side comparrison of available packages, none featured a real side-by-side that let users quickly see similarities or differences between plans. Lots of lists of features, no grids of levels/availability.

It was clear that all the presenters from Endeca (and the better presenters from Endeca partners) focused on quantifiable results, and concrete to-do lists that attendees could take home with them.

Endeca's Rob Gonzalez gave a really good talk on relevance tuning that continued that trend. The central take-aways were that relavence in search results is key to conversion, and that relavence tuning isn't a solution, but a process. I really liked that he drove home the need for adequate baseline data to work from, concrete metrics by which to measure success, and an iterative approach to reaching those goals.

He also introduced a few really nice features in Endeca 6.2 that add some human give to a search system. The kind of thing that lets you say, "You wanted a three-star restaurant under $20, but there's also a four-star restaurant at $21." Really cool.

At luch I sat at a table with Peter Cobb, founder and CEO of eBags, who spoke about reverse mentoring: they get a few interns in every year to educate their executives about what's going on in social media and talk about how to better use those media to engage with customers. Great idea.

He also talked about how eBags is getting around customers' reluctance to buy bags online. User reviews, product movies, high-resolution pictures, on-model shots, all contribute to giving customers confidence that they know what they're ordering.

And I got to talk with Jim Morris from Buzzillions.com, who had some fantastic insights into how to make reviews more trustworthy and complete (or, as I put it, "How do you filter the crazies?").

I skipped the 1-1:45 session to corner Vlad at the Endeca booth. The guy is incredibly sharp, and walked me through a simple solution to a gnarly (but subtile) problem that's been plaguing us for months. Thanks, vlad!

After that it was time for Matt Moog's talk on "How to Turn User Reviews into Social Shopping". Sounds dry, but it was great stuff, and took my earlier conversation with Jim to the next level. Moog touched on making reviews actionable, not just dead-end data that that you first read, then hit the back button.

He talked about how customers love uploading photos, proving his point with the photos that users had submitted of themselves enjoying a particular cream cheese.

And he talked about giving customers more control over what kinds of email they receive from you, and how often. This is definitely in "quick win" territory. Seems obvious now that he's said it, but I think anyone who doesn't do it radically underestimates its effectiveness.

I attended two talkes in the 3-3:45 session, and ended up bailing completely from both. The first got bogged down in the "get everyone on the same page" section, and the second got lost wandering through subtile differences vaguely articulated at length.

The closing was a nice wrap-up, "Rah-rah Endeca!" moment, and I think it was a testament to the quality of the event that all the weary attendees were still pretty enthused. And I have to give huge kudos to Endeca's Chief Strategist Paul Sonderegger for making all of his presentations really outstanding.