12 September 2007
Use A/B Split Testing To Improve Conversions
One of my members asked me a question about how to test affiliate marketing campaigns to maximize sales conversion.
He told me that currently he was only split-testing his Yahoo Search Marketing PPC advertisements, and he wanted to know what else he could do.
In testing any marketing or advertising medium, you need to step back and look at the entire sales process or funnel.
Each point of contact you have with a prospect or customer is a stage.
And when you talk about testing, you should test variations of your ad copy at each stage.
This member happened to be testing his PPC ad conversion. This is good. It’s the first point of contact between him and his prospects.
But one idea to keep in mind is this… Just because an ad has a higher CTR, doesn’t mean that it will translate into higher sales conversions.
Think of an ad as a fish net. If you have 2 nets, and one net has large holes in it, and the other net has small holes in it, what’s going to happen when you try to catch a bunch of fish with both nets?
You’ll end up with fish in both nets, but the size of the fish will be different, because the net with the larger holes allowed the smaller fish to get out.
What do you want to see happen? Do you want bigger fish or smaller fish?
Here’s what you might want to do.
You might want to split test which ad generates the most affiliate sales and then use this ad as the control going forward.
You can easily accomplish this using the Keyword Radar software.
To do this, just add the “ad version” parameter in your landing page URL. So, when your PPC provider serves ad #1, append “&ver=1″ to your landing page URL, and when they server ad #2, append “&ver=2″ to the landing page URL.
Once you start converting some affiliate sales, you can run a Keyword Radar report on this “ad version” parameter to see which ad pulled in more sales.
Once you are confident in your ad, then you can move to the other stages in the sales funnel, such as your landing page, or autoresponder sequence, and test how variations in those stages affect sales.
Here’s how to A/B split test your landing page:
You can define a test in your Keyword Radar members area so that the software will automatically rotate variations of your ad copy, images, or any other HTML element within your landing page. Each version will be tagged with a special code so that you can later run sales reports to see which variation converted into the most sales.
As a side note, if you wanted to test which autoresponder message in your follow up sequence generated the most sales, use this parameter in your landing page url: “ar=n” and change n to the number of your email sequence.
Please comment or send me an email if you have any questions.
This looks like an excellent product and I certainly want to try it!