Comments on: Landing Page Testing: Testing for Impact Not Variations

By: @brandrsn

@brandrsn — Sun, 17 Jul 2011 14:52:47 +0000

Thanks Bryan. I'll let you know how I get on…

By: BryanGroks

BryanGroks — Sun, 17 Jul 2011 14:37:16 +0000

In reply to @brandrsn. You can certainly test one process against the other. However, don\'t remove fields in one versus the other, don\'t relabel them, don\'t re-order them, don\'t add or remove confidence building materials. If you remove a progress indicator you need to run another variation with a one step progress indicator. Another words make sure you have isolated and understand the variables you are testing.

By: @brandrsn

@brandrsn — Sun, 17 Jul 2011 13:43:38 +0000

Nice article Bryan. Though what if you want to test one 'type of process' against another, rather than two single pages? For example, a one-page, accordion-type registration form versus a two-page registration form.

From your post, I'd guess you would suggest only testing one 'type' of registration process against another (leaving the individual on-page elements of both variations identical, for now). Would that be correct?

By: thegrok

thegrok — Mon, 27 Jun 2011 18:06:47 +0000

In reply to Tim Duke. Even if it was just an A/B test then they should have tested the 3 variables not the dozen I had listed. Not all variables are equal.

By: Tim Duke

Tim Duke — Fri, 24 Jun 2011 17:25:26 +0000

You’re assuming this company wasn’t doing a simple A/B test. They might have such low traffic that running a full MVT test would take 108 days… But instead, they knew some of the core issues at hand, and threw in a few best practices to make things even more clear (arrows, button color, better UVP, etc).

By: Linda

Linda — Tue, 14 Jun 2011 19:23:42 +0000

It took me a while to get back to this but I’m glad I did. While I got many of the variables, I did miss a few. And I got the basic concepts as to the first (main) things to focus on as well. But it is great to have another set of eyes looking at a scenario for me! Post more of these case study exercises. It’s very interesting and helpful!

By: Bryan e Jeffrey Eisenberg - Landing page test: testare l’impatto, non le variazioni - ideawebitalia

Mon, 13 Jun 2011 04:37:51 +0000

[…] Landing Page Testing: Testing for Impact Not Variations 3 giugno 2011 […]

By: John Hyde in Nelson NZ

John Hyde in Nelson NZ — Thu, 09 Jun 2011 21:39:28 +0000

They are sweating the small stuff.

I would go really big and have a page that clearly spells out who the service is for, what the benefits are, and clearly explains the next move and how it will work.