7 {{ upvoteCount | shortNum }}
7 {{ upvoteCount | shortNum }}

easy way to split test sales page?

by new guy
6 replies
Hi, I am running some Adwords and want to test 2 versions of my sales page. I want to send half the visitors to one sales page and half to the other sales page. Can I do that right from adwords or is there a script I can use for that?
#easy #page #sales #split
  • Profile picture of the author JohnMcCabe
    If you're just running a quick test, an easy way to split your traffic is to double up your ads.

    For each ad, create two identical versions with two different destination urls.

    Like this:

    Ad 1 --> Page 1
    Ad 1 --> Page 2

    Ad 2 --> Page 1
    Ad 2 --> Page 2

    Set the campaign to serve the ads equally. You should get approximately 25% of your traffic to each combination. Just use Adwords' result tracking for your results.
    {{ DiscussionBoard.errors[92285].message }}
    • Profile picture of the author Justin Harrison
      Originally Posted by JohnMcCabe View Post

      If you're just running a quick test, an easy way to split your traffic is to double up your ads.

      For each ad, create two identical versions with two different destination urls.

      Like this:

      Ad 1 --> Page 1
      Ad 1 --> Page 2

      Ad 2 --> Page 1
      Ad 2 --> Page 2

      Set the campaign to serve the ads equally. You should get approximately 25% of your traffic to each combination. Just use Adwords' result tracking for your results.


      The problem with split testing this way through Adwords ad variations, is that should the visitor not take action on the first time contact, there is no way to trace or remember the exact path, and as a result your split test would therefore encounter what we in the split testing profession call "noise", which in turn would result in inconclusive results.

      You can test this way, but really it is not an accurate way to go about your testing, and I have proven with larger batch tests that you could be out by as much as 30% when drawing your final conclusions.
      {{ DiscussionBoard.errors[105079].message }}
      • Profile picture of the author JohnMcCabe
        Originally Posted by Justin Harrison View Post

        The problem with split testing this way through Adwords ad variations, is that should the visitor not take action on the first time contact, there is no way to trace or remember the exact path, and as a result your split test would therefore encounter what we in the split testing profession call “noise”, which in turn would result in inconclusive results.

        You can test this way, but really it is not an accurate way to go about your testing, and I have proven with larger batch tests that you could be out by as much as 30% when drawing your final conclusions.
        I can't argue with you there. I should have qualified my original post.

        I tend to use this method more for 'is there even a market here' testing, where I am looking for wide variations that let me draw initial conclusions in spite of the noise. In other words, in this type of test, I'm swinging for the fences, hoping for a home run or a total strikeout.

        Once you're trying to make incremental improvements, a more rigorous method is a necessity.

        Thanks, Justin...
        {{ DiscussionBoard.errors[105285].message }}
  • Profile picture of the author WEALTH EMBASSY
    this link is very interesting
    {{ DiscussionBoard.errors[105512].message }}
    • Profile picture of the author new guy
      Thanks for the tips guys. Checking into the website optimizer now.
      Signature

      Please read the sig file rules

      {{ DiscussionBoard.errors[110504].message }}

Trending Topics