A/B Test to improve scroll rate on homepage

[siddev]

Hey all,

I am currently working on composing an A/B test that presents a visual change on the homepage of our site to encourage scroll rate. We have within Adobe Analytics a 25% scroll-rate metric, but are concerned that Target will look at visitors scroll behavior across the whole site rather than specifically on the homepage (where the activity exists). First, is this a legitimate worry? Secondly, how do we focus on checking whether or not homepage scrolls have changed? Would this be through setting an audience in Target (and the activity settings) or editing the Reports audience?

Thank you very much for any help or assistance on this issue. 

[alexbishop]

I'm assuming you're talking about using that 25% Scroll Metric as your Primary activity KPI; if that's the case then yes, you're correct, if you look top-level at the 25% scroll metric then it will be considering all instances of that metric.

A couple of options I can think of:
i) Use the A4T Panel in Adobe Analytics but use dimensions to filter it down to Homepage i.e., only for People in the Variant that are on the Homepage.
ii) The other option would be to create a "25% Scroll on Homepage" metric and then use that as your Primary KPI

[siddev]

Thanks for your reply! Yeah, you're correct - we are looking at 25% scrolls here. 

Would there be a way to write an audience that only factors a view on the homepage? I see that you can segment the audiences based on page/url, but wondering if this would lead to any unintended side effects. 

[alexbishop · Accepted solution]

Yes, you could just create a hit-level segment (e.g., hit where page = homepage) in Adobe Analytics and apply that to your A4T panel, then that would filter down the 25% scroll metric (and all other metrics & dimensions in the panel) to only count page views on the homepage