Rich-text Reply

Combining different types of goals in to one goal

KevinMcDaniel 11-14-14
Accepted Solution

Combining different types of goals in to one goal

What I'm looking to do is essentially have an "AND" button on a goal. For example I wish a conversion to be someone clicking on an element on the landing page AND reaching the order confirmation page.


My example:

I have an element in the header of the page that is running across the entire site and I'm trying to improve its usage. Interacting with that element is my primary goal. Also though, I'm tracking order completions for each variation. The winner of my primary goal is the worst performer in terms of order completion.


I've let this test hit proper sample size and it has been running for an appropriate amount of time. I've come up with a couple ideas as to why this is the case:


  1. The experience after interacting with the element is causing people to bail. Because more people are being pushed through this funnel instead of arriving from another way it is causing the decrease in order completion.
  2. Say that element is only clicked on by 10% of the test traffic. That means that 90% of the data collected by the order completion goal is essentially noise that could be effected by whatever landing page they are on, not necessarily the improvement to the header.


Any help would be appreciated.


Khattaab 11-16-14

Re: Combining different types of goals in to one goal

@KevinMcDaniel , great attribution question. A point of clarification: Are users in the Original not able to view the header CTA or is it just less prominent?


To address your first point, if you're having success pushing more users to the next step in the funnel, but order confirmations are down due to bounce, you are likely pushing less qualified traffic through (need to do a better job of priming users with what's to come and provide them the most relevant information they need to make a purchase decision). This also presents an opportunity to start optimizing the steps between CTA click and order confirmation.

You shouldn't be concerned with noise impacting the order confirmation page conversion rate. Any user who was exposed to the new header CTA, regardless if they clicked it or not, will remain associated with the variation until the end of their session (as long as they don't clear their cookies). Consequently, the order confirmation conversion rate for users in the new CTA variation will not by influenced by users who saw the Original CTA.


The best way to apply multiple AND conditions to a user is through targeting. You could drop a cookie on those users who click the new CTA and then run a seperate A/B test that targets only users with that cookie (no need to make any variation changes to the steps that follow the CTA click).

Please refer to this related Community post on evaluating multiple conversion events.

Khattaab Khan
Director, Experience Optimization | BVAccel
Level 5