Sunday, March 16, 2008

Integrating user feedback...

Today we launched our new and improved TickerHound design!

Well, the design isn't very "new", it's just been updated a bit based on all the feedback we received from the TickerHound community over the last 10 weeks.

From the get go we told ourselves that we wanted to build a site that our users absolutely LOVED to visit. Our goal was to create the iPod of financial education sites - and no, I don't mean it would play music when you visited, I mean that we wanted the user experience to be so damned compelling and so pleasurable that every visitor felt like the site was built just for them.

So in order to do that we knew we'd have to be in constant communication with our members. We knew we'd have to actively solicit feedback from and actively LISTEN to what they were telling us.

Now, when requesting feedback, it's very easy to take the stuff you want to hear and throw out the rest. I believe it's what psychologists call the "confirmation bias" - you tend to only pay attention to the data that supports your predetermined hypotheses. Being that this predisposition is hardwired into our brains, we knew we had to be OVERLY conscious of it and integrate it into our decision making process on a regular basis.

So as we worked with our design team over at nclud, LLC. (great guys, super talented, I highly recommend them) we would constantly revisit older feedback surveys. We would reexamine data we originally tossed out and we'd revisit features we decided not to add - some of them stayed in the trash can, others ended up on the site.

The point being - while you can't take every single suggestion from every single member, you are doomed to failure if you don't actively examine and reexamine what your users are telling you. Those who fail to ACTIVELY LISTEN will die!

Ok fine, I'm being a little over dramatic here but the point is still the same - listen to your members and they will help you succeed.

The other thing we had to remember was that sometimes our members will tell us more through their actions than their words. For instance, when I saw that we were getting a higher percentage of user activity per day on older questions (questions not on the first page of the questions list), I knew we had to add more numbers to our pagination scheme (it used to be 3, now it's 9).

We also relied HEAVILY on our Google Analytics data - yes, we're cheap, we use Google Analytics. Here's a perfect example:

One of the goals with the redesign was to increase the amount of time visitors spent on the site. It's not particularly low or anything but it'd be great to get it higher (obviously). So I examined the pages that have gotten the highest exit rates over the last 8 weeks. Then I used Google Analytics to track the visitors' paths through the site until they hit the high-exit pages.

It turns out that the pages they exited on tended to only be the 2nd Question page they visited (these are the pages that contain the full question and the associated answers - if any). And they would tend to exit more often than not if it was a question page that didn't have an answer yet (seems intuitive but it just didn't occur to us until we examined the data).

So now we reworked the code for displaying the "Related Questions" in the right-column of the Question pages. Now, the top 2 "Related Questions" will be questions that have at least 1 answer.

While we don't have a lot of data yet, it's clear from early tests that our strategy is paying off in terms of "page views per visit" which I feel will ultimately translate into more time spent on the site.

We'll be adding more features throughout March and we've also made it a company mission to add 2 new features (that have been requested by our users of course) each and every month.

We might not be the sharpest tools in the shed, but we want to be the best LISTENERS on Wall Street!

blog comments powered by Disqus