How to Split Test Banner Ads Fast – Case Study

Written by Zac Johnson
social social social

Whether you are promoting your own web site, or a simple affiliate campaign, it’s important to know what creatives are pulling the best click through rates and conversions. What converts on one web site, may not convert on another, same with gender and other demographics.

Today I was playing around with some banners that HostGator sent over to me, and I figured I would test them out to see what type of click through rates they would get. I had around 50 different banners to choose from, but I focused on the 125×125 banner sizes which also promoted their low $3.96 web hosting packages.

Split Testing on Niche Web Sites

A great way to target to web site owners and buy advertising across different niche web site is through BuySellAds, and that’s who I’m going to use for this case study. I purchased two banner ad spots; one on a graphics tutorial site (SITE 1), and another on a blog template site (SITE 2). Both web sites attract an audience that would be interested in web hosting and site design. Now keep in mind, I only setup this case study a few hours ago, but the impressions and clicks are already coming in, and they are already quite interesting. You can see the stats from today below.

Site 1 Banner Stats:

Site 2 Banner Stats:

Sure, it’s not the most data in the world, but it should make for an interesting case study after the final numbers are in. It’s still extremely early in the reporting, but it’s quite interesting to see such a vast difference in the initial stats. SITE 1 makes Banner #2 look like a winner, but SITE2 has the complete opposite stats.

I have a few other campaigns running that I will share with you as they come to a close. I will report full stats from this campaign next month when this campaign closes. How much of a difference do you think we will see in click through rates after 250,000+ impressions are served?

Additional Creative Split Testing

Split testing ad campaigns is nothing new, it’s just a matter of finding different creatives to test and how to track it. With BuySellAds you can buy advertising on a monthly or cpm basis, and it’s all focused on banner advertising. It’s also extremely easy to add, remove and change ad campaigns as they are running. Advertising on web sites listed in the BuySellAds market place is a great option if you are looking to launch a product and want to get a good idea on what banners people are actually clicking on. However, BuySellAds isn’t the only option out there… there’s actually hundred of different sites to buy advertising and help your quickly spend a few hundred dollars to test what people are clicking on. Another great example is ads.pof.com.

Ads.pof.com is the self serve advertising platform for Plenty of Fish, which can be very beneficial for demographic and specific audience ad copy testing. You can really push through a lot of volume on Plenty of Fish fast, so be sure to watch your daily spend amounts and cost per click rates. Ads.pof.com is just like advertising through Facebook, but you get to use regular banner ads (768×90, 300×250), which are what most web sites and affiliate programs use for their affiliate marketing.

In the example above you can see two banner ads for a weight loss offer, both being promoted to the same audience. While they are a very similar, the bottom better gets around a 25% better click through rate. This may not seem like a lot, but if you end up serving a million impressions (which isn’t hard), using the banner with the better ctr would result in almost 100 more clicks. It’s all a numbers game and testing what works.

What are your favorite methods and sources for split testing ad creatives? Have you used PlentyOfFish, FacebookAds or BuySellAds yet?

(Visited 34 times, 1 visits today)

10 Replies to “How to Split Test Banner Ads Fast – Case Study”

  1. hi Zac
    Its always great to see major changes in CTR by just changing the position of Banner Ds.
    Thanks for this.

  2. The only thing I have ever learned from split testing is that I can NEVER predict the results. When I think that difference between my splits is so minor that there is no possible way it could have an impact, I usually find the most surprising results.
    My recent post Day 18

  3. Overall this is very good option to promote a new coming product for any company because now days it's very popular. Earlier source of ads was not enough. Company was promoted any product through sing boards beside the road and in transport buses and in train also . But now this is the very unique and very great idea to ads of product.
    My recent post Summer Holiday Packages

    1. Ad networks make it very easy to pinpoint and select what demographic you would like to reach so easily now. Think back how different it was 5-10 years ago… just buying flat traffic on a cpm basis!

Comments are closed.