Comments on: 8 Common Sins of Online Conversion Testing that Organizations Let Happen

By: uberVU - social comments

uberVU - social comments — Tue, 23 Mar 2010 06:30:33 +0000

Social comments and analytics for this post…

This post was mentioned on Twitter by RexDixon: 8 Common Sins of Online Conversion Testing that Organizations Let Happen – http://bit.ly/9jQnOx #conversions #measure #abtests /via @abtests…

By: Laura Klein

Laura Klein — Mon, 22 Mar 2010 18:28:39 +0000

Great list. I’ve seen all of those! Another of my favorites is “Testing two wildly different things against each other and then assuming that every element of the one that performs better is actually better.” Sometimes you need to test very small changes against each other just to be able to narrow down what’s really making the difference. After all, if you change the text, the layout, AND the call to action, can you really say which of those is making the biggest difference in a multivariate test?

I still LOVE multivariate testing, but it is so easy to go wrong without experience and guidance.

Thanks for the post!

By: Steve P.

Steve P. — Thu, 09 Apr 2009 11:03:56 +0000

Great post. As a landing page optimization consultant I completely agree with your 8 sins especially with #6 about not running the control panel at the same time as the test panels. Often businesses that hire me tell me about why they implemented some changes and the tests are often not against the control but the control first, then the test page after and they compare results.

I’d like to add a #9 – Not all tests will have high conversion “winners” everytime.

By: Jusitn Palmer

Jusitn Palmer — Thu, 09 Apr 2009 04:02:57 +0000

#3 is a big one for me. Took a while to accept the fact that failures were basically successes, because if it weren’t for testing, I would have just guessed and caused unknown negative impact.

Great post Josh.