Skip to content

When you purchase through referral links on our site, we earn a commission. Read our Advertiser Disclosure

Experiment #1: Multiple Split tests at once!

The digital Income ExperimentI’ve recently tested two rounds of traffic to the second funnel I mentioned in my previous post. This is the funnel where I'm testing out different prices for my front-end offer. The objective of the course is to be able to break even and or make an additional $0.70 for every hundred visitors that land on my squeeze page. In my post, I explained that I just wanted to improve the conversion of my front-end offer but in all honesty, there are many things to consider when doing this. So I went wild and decided to run several tests at once. I've set up some variations of the sales funnel and purchased traffic from the same two people whose traffic generated sales for me.
 
In the first round, my thank you page received 386 hits from approximately 800 purchased clicks coming into my funnel. I’ve checked and confirmed this as my auto-responder numbers seemed to agree with the numbers. This means the tracking software's results were verified in my Aweber account.
 
Remember that my idea is that I wanted to test out different prices for my front-end offer so that I can increase my overall profit for every traffic purchase. I created 3 different pages with three different prices. I placed all these pages in one rotator set to distribute traffic to these pages equally. For those of you who do not know what a rotator is, in short, it's a link within my tracking software that holds a number of links. The rotator is set to distribute traffic to each link equally. So by just using the rotator link, the first visitor sees one page, and the second and third visitors will each also see a different page.
 
Because I’d rather do a 50/50 split test to at least keep my control experiment described above, I decided to create a second thank you page. The second page has a welcome video where I introduce myself and share my Facebook profile. I then show some income proof by logging into one of my accounts. At the end of the video, I encourage the visitor to click on the button below the video to get access to my offer. That button links to the rotator with 3 different sales pages. I assumed this would be the best way to test out my theory.
 
So here’s how the front end of my funnel is constructed:
Traffic comes to the Squeeze page which then goes to a Thank you page. I'm running a 50/50 split test on this thank you page and will describe how below:
 
Split 1: This is a straight-up “Get The Money” Funnel meaning that the thank you page is a sales page. The objective is to get a sale as soon as a visitor becomes a subscriber. This page tracks the conversions from my squeeze page.
Split 2: This one is a pre-sell video to increase engagement, warm up the newly acquired subscriber and then send the subscriber to a sales page.  In my design, the sales page is a short page with a summary of what the subscriber should expect to receive. The call to action is a button that links to a rotator. This rotator has three different links linking to 3 different checkout pages to test out three different prices. This page also has a conversion pixel.
 
To run this test, I’ve purchased 400 clicks. Although all the traffic has yet to be delivered I’m a bit bewildered by the results. Out of 400 clicks, I’ve received 125 conversions. At first, what I noticed is that my overall conversion rate took a plunge. From 43% to 31%.  This was completely unexpected and is not good for my bottom line as the fewer subscribers I get, the fewer chances I have of making a sale. After all the sale is what I'm really after. I know the problem is not with the squeeze page because I didn't change anything there from the first run. The drop in conversions could be caused by the split test… or the traffic purchased this time is not as efficient but I doubt that's the case since I did go to the same two traffic providers that generated sales for me on my previous traffic buys.
 
Nonetheless checking my auto-responder account, I notice right away a discrepancy. Instead of 125 conversions, I see 165. Meaning my conversion rate is actually 41% which makes more sense to me. So what happened here exactly? Why did the tracker not count 40 of my conversions? Somehow the conversion pixels did not fire for those 40 subscribers, these subscribers are unique thanks to the tracker so we can't say that the drop is due to the fact that I have repeat visitors. That was not the case.
 
Another thing I noticed, is the results of the split test. This is what I was going after from the beginning. Remember my goal is to see what price is the most attractive for my subscribers. So out of the (125?) conversions, the tracking link which holds my split test of the two thank you pages only shows 76 unique hits and 85 raw hits. Raw hits occur when one specific individual clicks on the same click more than the one time we expect of him. Now this does not make sense to me as this link should read 125. I’m also puzzled as to how my conversion rate was calculated. According to the results I see from my thank you link my overall conversion rate would be 18.75% and not the 31% mentioned earlier. The plot thickens…
 
 
The very last piece of the puzzle did work fine however and that I’m pleased to see. According to the hits tracked by the rotator, the second thank you page hits match the hits of the rotator that sends traffic to my 3 other sales pages. Although this part of the puzzle worked, the question still remains, what is going on with the numbers? To answer this question I thought it would be best to contact the developers of the tracking tool. I explained the situation and you will see the reply below:
Hi there,
 
Well the first thing is that you should be using a split testing link instead of a rotator when conducting split tests.
 
I think you are testing too many things at once. Unless you are spending thousands of dollars on traffic daily then you need to test one thing at a time. 
 
Instead of testing straight to a sales page and then to a pre-sell page with 3 different salespages. You need to just test straight to a sales page vs to a pre-sell page that leads to the same salespage.
 
Once you figure out which method works better than you can test the actual salespages themselves.
 
For your conversion discrepancies please see the FAQ below as it discusses this issue in detail:
99% of the time the problem is one of the items listed in this FAQ.
Apparently, I'm doing too much at once. I should tone down my split tests and either use a different tracking link for every URL I link to or insert multiple unique pixels on the pages I would like to test out. I'm recommended to test out the conversions from just the sales page, and conversions of a pre-sell video to the exact same sales page. Although this sounds like a good test, this is not what I'm currently after. I want to test my current sales page with different prices. Yes, my previous test was all over the place but now I'm just going to focus on the one sales page with different prices.  Now in hypothesis, if I sell my front end at $37 I will make a little bit of profit for every 100 visitors I get. I'm going to try testing the following prices, $7, and $17. $27 and $37.  My control experiment does not have a buy button that displays the price. Maybe I should include an additional one containing the price. That means I could be testing up to 5 different pages. I'm eager to find out what results I'll get.

Share this post on social

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

The content on this website is for educational and informational purposes only and should not be construed as professional financial advice. We are not a financial institution and does not provide any financial products or services. We strive to provide up-to-date information but make no warranties regarding the accuracy of our information.