A Look at Web Personalization A/B Testing

Not applicable


Marketers understand the importance of A/B testing – and specifically how simple changes around copy, color, placement and headlines can make a big difference in the success of our campaigns.

Marketo’s Web Personalization (RTP) platform makes it as easy as 1-2-3 to create A/B tests.  Simply check off Split Testing, name the group for the campaigns you want to test and click Create. You can then add campaigns to the group.


RTP will take the number of campaigns in the group and split the number of impressions between the campaigns, for example:

  • If you have 2 campaigns in the group – each one will show up 50% of the time
  • If you have 3 campaigns in the group – each one will show up 33.3% of the time

How many campaigns should I test?

While you can test as many campaigns as you want, best practice recommendation is to test two to three campaigns at a time.

Does segmentation matter?

Yes.  You want to keep the audience the same so you aren’t adding an additional variable to the test. Be sure each split test group is associated with the same segment.

What should I A/B test?

In addition to what’s already been mentioned, try testing against different content pieces (e.g. Case Study A vs. Case Study B). This great Marketo blog entitled 7 Elements to A/B Test in Your Digital Ads mentions testing using time of day and day of week.

With Marketo’s winter release that is now possible! In addition to using the new web personalization scheduling feature to schedule campaigns in advance, you can now choose the time of day and days of the week your campaign should run:.


When setting campaign time zone note that the default setting runs campaigns in the web visitor's time zone. You can choose to schedule the campaign to run based on specific time zones:


How is Sampling different than A/B testing?

Sampling is where you can configure the percentage of the traffic that will see the campaign – without having second campaign version. In this example we will be showing the selected campaign to only 20% of the web visitors:


Can I A/B test and Sample?

No, both features cannot work together, meaning – if you set A/B testing and 20% sampling the sampling will be disregarded.

Auto Tune - It works for Drake, make it work for you!

Lastly, don’t forget to take advantage of the Auto Tune feature. This allows the RTP platform to automatically recognize the better performing split test campaign, and continues showing the highest converting campaign while pausing the other(s). To enable it, simply check off Auto Tune: