Friday, March 8, 2013

Building a better competitor

6 out of 7 split tests fail to beat the original version, doing worse or producing no change at all.  That means 85% of the time, your split testing efforts are hurting your business.  What's worse is that no significant change takes months to determine.  If you want to win more often, you need customer feedback.

When you are about to make a change to your site, especially a major one, you should first get real feedback on the proposed changes.

It is important not to ask your customers what they want or think about a potential option before you have a finished product.  When you ask someone to think, they begin using their rational brain.  Unfortunately, the rational brain doesn't have much of an imagination, which sort of defeats the purpose of imagining a better website.

You need to create finished versions of the proposed changes, either as an image or as a working example.  Only with a finished product can you get a person's gut response, the opinion of the reptilian brain and limbic system.

You can use a service like 5 Second Test or create your own panel of visitors and customers to get feedback.

The reptilian brain and limbic system don't have the capacity for speech, so a valid response should start with like/dislike and maybe a feeling.  You know you have tapped into someone's limbic system if they have a hard time finding words.  Stick to the easy questions, "Do you like it?" "Does it feel right?"

Once you have something that your visitors and customers respond well to, you can start split testing.  It may take a few iterations to find a good competitor, but it is nothing compared to the cost of losing 85% of the time.

How do you weed out the bad competitors?