UX Success

The Power of Three is symbolized by a "triquetra", a Celtic pattern that shows the center of three connected circles

Three is a magic number, for sure.

1. Where am I ?
2. What can I do here?
3. What should I do next?

Three is key to optimizing landing pages — that and delivering on the promise the search result makes. It used to be people asked how to optimize a HOME page, but thank goodness folks are realizing that Google Search is the real home page these days. ( More on that some other time).

For awhile now, in lab-based usability tests, I”ve been doing the famous “5 second test” to see if people can grok the gist of the site from the page they find themselves on:

I show a page out of context  for a grand total of five seconds and ask the test participant to write down what they remember…..

Now, there is this site for doing quick tests like this. “Five Second Test” I found it while I was searching around for tools and stuff. This is a really cool little app where one can take a random 5 second test or set up their own.

I did just that.

Here is how:

* got a screenshot of my client”s home page;

* went to http://fivesecondtest.com * filled in the form;

* uploaded the screenshot

* posted the generated link at Twitter and asked tweeps participate.

The participants were asked to look at the picture for 5 seconds and list the  5 things they remembered.

The results:    I got some responses ( like 22). The top mentions:

* Orange

* Spanish

* Nurse / mention of the person in the picture

* Cancer

* Cherries

This was a slight concern to me because I couldn’t answer the really important questions.

(Here is where the POWER of THREE comes in).  I ask what I deem to be the three most important questions after asking participants to list their five things.

Adding these three powerful questions to the mix really helps me start to understand if that landing page is contextual, has the value proposition articulated, and has a clear path to a good outcome.  Try it!

Check out the tool  Five Second Test

OK. Let me just get this out of the way. This clip of Hammy only very loosely relates to A/B testing ….. but it is so doggone cute that I just had find a way to post and talk about it.

My husband sent it to me because he knows I cannot resist the cuteness of some little critter doing cute things. Interestingly, the comments about this clip on YouTube, are mostly: “saaaaaay, this test method isn”t valid…. doesn”t prove a thing……method is flawed…not enough subjects…bad experiment…etc.”  Those comments are what got me thinking about live web A/B and multivariate testing vs. say, user testing in the lab.

A/B and multivariate testing is an old-school marketing experiment technique. Instead of trying to figure out what creative treatment or messaging or button placement on your campaign landing page will be the best at driving conversion, do an experiment by launching several versions of the page that randomly serve to users and find out for sure. I am writing this because many people lately have been asking me to test user preference in in the lab. 

 Trust me, I love a good usability lab test.  I live to conduct them. Watching a software developer download a product evaluation is much more fun even then watching Hammy cutely eat the organic broccoli….. Sadly, lab testing isn”t the best way to determine preference *sigh*.Both usability lab testing and A/B web testing have their places.  Lab testing has many benefits, primarily revealing behavioral user issues that are difficult to figure out with web analytics.  Lab testing allows us to observe users (like Hammy) interacting. And, allows us to ask them “why”.  This is an invaluable technique I use to grok how a user will behave with a complex interactive process. 

A/B and multivariate testing, on the other hand are quantitative methods that measure user preference by putting randomly assigned people into two or more “controlled and different” experiences on the LIVE website and seeing which experience performs better against preset goals. A/B and multivariate testing is not perfect but we get good quantitative data, and a good analyst can often draw qualitative conclusions from that data….. or, form a hypothesis that leads to a lab test.

A/B testing is not perfect but a good way to spend your money BEFORE committing to the expense of a lab test.  If you are interested in a FREE way to get your A/B experiments started,  check out this tutorial to find out how to use Google Website Optimizer.