Home > Questions and Answers > General Questions

When should I pick a winner? [Statistically significance] (4)


04-17-2014 08:48 AM #1 phlong (Member)
When should I pick a winner? [Statistically significance]

Hi there,
I've been split testing three headlines on my landing pages and so far I could kill the bad performing.
Yet, I'm stuck with this one:

One landing page performs a bit better but if I let it run for the whole day, both convert almost equal. Should I decide the winner soon (if there is one, else I would pick the one I like if both headlines perform the same way) or let it run? I usually plan to end the split test once the landing pages have 600-700 visitors.


By the way: Is there a priority of split testing variables? Like headlines, colors, images, button. What's the hierarchy?
Thanks!


04-17-2014 09:58 AM #2 zeno (Administrator)

Firstly it would be good if the stats had column headers/labels, takes a bit more mental processing than necessary without the labels.

Over what time period did you get these stats? If two variables give very similar stats it makes sense to just leave them running - for all you know headline 1 might do better Mon/Tues/Wed and then headline 2 does better for the rest of the week.

Hierarchy wise I would test Images, then headlines/ad copy, then colors. I assume by buttons you mean button shapes/colours/wordings? I would put changes to your CTA between colors and headlines although the CTA wording is part of your ad copy.

This of course depends on the context - for banners this hierarchy is simple. For landers, you would generally test completely different designs first, then might vary the ad copy over images if you are focusing on pre-selling something.


04-17-2014 12:25 PM #3 caurmen (Administrator)

Could you paste the image again with headers for the columns? As Zeno says, it's making my brain hurt a bit. I promise to give recommendations once that's done!


04-17-2014 12:51 PM #4 phlong (Member)

Haha sure. Me idiot, forgot it.
Here we go:

Been running this for around a week now. (Daily budget $50). Just conversions today again: 2 for one LP and 1 for the other lol.
PS: One of those 2 headlines is a headline that 80% of the competition uses. So I'm testing my own headline vs competition's right now. Would be awesome if mine is the winner


Home > Questions and Answers > General Questions