In this lesson, you’ll learn the details of the different ways to run split tests in Ontraport so you can find winning versions of your marketing efforts.
Home > Ontraport for marketing > Marketing performance optimization > Split testing deep dive
Home > Marketing performance optimization > Split testing deep dive
Split testing deep dive
Course Instructor
- How to add a new version of a page to split test
- How to add a new version of an email to split test
- How to set up an automation to run the split test
- How to see which version is performing best
- Options for what to do once you’ve looked at results
- Split testing best practices
Comments
[Block//Author//First Name ##cap_first] [Block//Author//Last Name ##cap_first] on [Block//Date Added %F j, Y g:i a%+0d0h0m]
[Block//Comment]
Leave your comment
14-day free trial. No commitment.
Related resources
[Block//Resource Type]
[Block//Resource Description ##ellipsis(166)]
Transcript
One of the things new marketers learn pretty quickly is that different images and copy really can make a big difference in performance. Sometimes one offer will get you twice as many sales as another, or one ad or email will get clicked 20% more often. Improvements like that can add right up and make a big difference, and split testing is how you find them.
Split testing is when you take an asset, like an email or page, and create a second version of it. You’ll change something about it, like the design or copy. Then, you’ll split the traffic between them and compare the performance of the different versions
For example, you might have two versions of the same email with different subject lines. You’ll split test them to see which one gets more opens or clicks. This way, you’ll find out what catches people’s interest, or if they like your sense of humor in your subject lines.
When you’re automating your follow up, finding little tweaks that matter and then using the winning versions forever is a permanent improvement for your business.
There are a few ways to run split tests in Ontraport, which I’ll cover now:
To Split test pages, you’ll first go to the page you want to test.
Let’s say you’re testing a sales page in an attempt to drive more conversions.
You’ll set up split testing by clicking the “A/B” button in the top right and toggling it on.
Then click the “+Add New” button.
Every time you add a new version, you’ll be asked if you want to start fresh or start with a copy of your existing page. Typically, it’s a good idea to make a copy, since it’ll make it easier for you to test specific elements more quickly.
Once you do that, it’ll load version B of this page and you’ll be able to make edits to it.
For example, let’s say you’re a veterinarian and you want to test whether pictures of dogs or cats draw in more people. You’d make a copy of your dog version (A), and change the pictures to cat photos on your new version (B).
Toggle back and forth between the versions using that same “A/B” button in the corner.
Once you’ve made your changes, save and publish your page. This will cause all future traffic on this page to automatically be split 50/50 between your two page versions.
After your page has been published for awhile, you can load the results of your test by going back to your split tested page and clicking the “Stats” tab along the top of the screen. Here, you’ll see your different versions as well as the number of visits, unique visits and conversions.
Now let’s Split test emails.
You’ll want to start out by creating a second version of your email in “Messages.”
Just like before, you can either copy your existing message or start from scratch. Maybe you want to test a plain text version of your email versus an HTML-rich version.
To split the traffic between your two versions, head to the Automation that’s currently sending this email. Load it up and find the email element that you want to test.
Go into edit mode on your Automation and click the plus sign above your email to add a new element.
When you add a “Split element,” you’ll see new “Version A” and “Version B” paths open up.
Add the two versions of your email into the spaces provided and notice that the flow meets up again below the test.
You can edit what percentage of traffic you want to send to each version by editing the split element itself.
This is particularly useful if you’re a little concerned about testing something in a proven system. You might worry that it will perform worse than before, but this feature lets you mitigate that risk by sending only — say, 20 percent — of your traffic to the new version.
Before you launch your test, it’s important to setup a goal below the split test, after the branches come back together. This is going to be how you measure your results, and it’s one of the areas Ontraport really shines when it comes to performance optimization.
See, in most email systems, if you can test at all, the results of your test are going to be based on which email got opened or clicked the most. The problem with that is that you don’t really care about opens or clicks.. You want what comes next. Probably a sale, right? Or maybe it’s a trial or registration, but it’s not clicks.
And, what you’re going to see is that very often the email that gets the most clicks is NOT the one that gets you the most sales. This is true for all areas of marketing. Just because a bunch of people click on your facebook ad, doesn’t mean those are the right people. You want quality, not quantity.
And ontraport is the only system that shows you your test results based on the outcome you actually want. The way you do it is with goals.
So, after your split test, below where the paths reemerge, add a goal. The goal should be the real conversion that you’re hoping to achieve with your email… a sale or registration or whatever it is.
Then, after you’ve published your automation and have contacts flowing through you split test, you can go to the goal conversion rate report in performance mode, click on your goal, and you’ll see exactly how many conversions were achieved by the folks who went down either side.
Of course, you can see your opens and clicks too, by clicking on the email itself… but stay focused on your real goals, which are down here.
Now, we’ve been talking about split testing emails to keep the example simple. But you can use the same exact concept to split test entire follow up flows.
By adding a split, then building entirely different follow-up systems on each side, you can test things like types of follow-up, waits between messages and so forth.
You could, for example, test wait durations in your nurturing funnel to see if your audience responds better or worse to faster or slower communications.
You could also move around the order of your follow-up steps by placing a task earlier in the process to see if it increases conversions.
Just as with the email step, all your future contacts will be automatically split between these two paths and will rejoin at the end.
You’ll also see your results in the same way, using “Performance Mode” within the Automation itself.
For example, if you have two different paths with the same goal at the bottom, you could use the “Goal Conversion Rate” and “Goal Conversion Time” reports to get a sense of how well each version is converting — and how quickly!
Here are some Final Tips for you:
While Ontraport lets you test four versions of your pages, it may take a fairly long time to get enough visits on each of the versions to show any meaningful results. That’s why we suggest testing just two pages at a time — but it’s your call!
Similarly, it’s best to not test too many elements at the same time.
If you’re running a test where you swap out the images, headline and form fields on an order page and your new version performs better, you won’t know exactly why that is. Was it the images? The form fields? The headline? Some combination of all three?
The best thing you can do is test these elements one at a time. Test your headline first, then you’ll know if the headline was what impacted your results. Then do the same with the image, form fields, etc.
Now that you’ve learned how to set up split tests in Ontraport, you’re ready to start experimenting with your marketing and driving real, positive changes for your business!
Split testing is when you take an asset, like an email or page, and create a second version of it. You’ll change something about it, like the design or copy. Then, you’ll split the traffic between them and compare the performance of the different versions
For example, you might have two versions of the same email with different subject lines. You’ll split test them to see which one gets more opens or clicks. This way, you’ll find out what catches people’s interest, or if they like your sense of humor in your subject lines.
When you’re automating your follow up, finding little tweaks that matter and then using the winning versions forever is a permanent improvement for your business.
There are a few ways to run split tests in Ontraport, which I’ll cover now:
To Split test pages, you’ll first go to the page you want to test.
Let’s say you’re testing a sales page in an attempt to drive more conversions.
You’ll set up split testing by clicking the “A/B” button in the top right and toggling it on.
Then click the “+Add New” button.
Every time you add a new version, you’ll be asked if you want to start fresh or start with a copy of your existing page. Typically, it’s a good idea to make a copy, since it’ll make it easier for you to test specific elements more quickly.
Once you do that, it’ll load version B of this page and you’ll be able to make edits to it.
For example, let’s say you’re a veterinarian and you want to test whether pictures of dogs or cats draw in more people. You’d make a copy of your dog version (A), and change the pictures to cat photos on your new version (B).
Toggle back and forth between the versions using that same “A/B” button in the corner.
Once you’ve made your changes, save and publish your page. This will cause all future traffic on this page to automatically be split 50/50 between your two page versions.
After your page has been published for awhile, you can load the results of your test by going back to your split tested page and clicking the “Stats” tab along the top of the screen. Here, you’ll see your different versions as well as the number of visits, unique visits and conversions.
Now let’s Split test emails.
You’ll want to start out by creating a second version of your email in “Messages.”
Just like before, you can either copy your existing message or start from scratch. Maybe you want to test a plain text version of your email versus an HTML-rich version.
To split the traffic between your two versions, head to the Automation that’s currently sending this email. Load it up and find the email element that you want to test.
Go into edit mode on your Automation and click the plus sign above your email to add a new element.
When you add a “Split element,” you’ll see new “Version A” and “Version B” paths open up.
Add the two versions of your email into the spaces provided and notice that the flow meets up again below the test.
You can edit what percentage of traffic you want to send to each version by editing the split element itself.
This is particularly useful if you’re a little concerned about testing something in a proven system. You might worry that it will perform worse than before, but this feature lets you mitigate that risk by sending only — say, 20 percent — of your traffic to the new version.
Before you launch your test, it’s important to setup a goal below the split test, after the branches come back together. This is going to be how you measure your results, and it’s one of the areas Ontraport really shines when it comes to performance optimization.
See, in most email systems, if you can test at all, the results of your test are going to be based on which email got opened or clicked the most. The problem with that is that you don’t really care about opens or clicks.. You want what comes next. Probably a sale, right? Or maybe it’s a trial or registration, but it’s not clicks.
And, what you’re going to see is that very often the email that gets the most clicks is NOT the one that gets you the most sales. This is true for all areas of marketing. Just because a bunch of people click on your facebook ad, doesn’t mean those are the right people. You want quality, not quantity.
And ontraport is the only system that shows you your test results based on the outcome you actually want. The way you do it is with goals.
So, after your split test, below where the paths reemerge, add a goal. The goal should be the real conversion that you’re hoping to achieve with your email… a sale or registration or whatever it is.
Then, after you’ve published your automation and have contacts flowing through you split test, you can go to the goal conversion rate report in performance mode, click on your goal, and you’ll see exactly how many conversions were achieved by the folks who went down either side.
Of course, you can see your opens and clicks too, by clicking on the email itself… but stay focused on your real goals, which are down here.
Now, we’ve been talking about split testing emails to keep the example simple. But you can use the same exact concept to split test entire follow up flows.
By adding a split, then building entirely different follow-up systems on each side, you can test things like types of follow-up, waits between messages and so forth.
You could, for example, test wait durations in your nurturing funnel to see if your audience responds better or worse to faster or slower communications.
You could also move around the order of your follow-up steps by placing a task earlier in the process to see if it increases conversions.
Just as with the email step, all your future contacts will be automatically split between these two paths and will rejoin at the end.
You’ll also see your results in the same way, using “Performance Mode” within the Automation itself.
For example, if you have two different paths with the same goal at the bottom, you could use the “Goal Conversion Rate” and “Goal Conversion Time” reports to get a sense of how well each version is converting — and how quickly!
Here are some Final Tips for you:
While Ontraport lets you test four versions of your pages, it may take a fairly long time to get enough visits on each of the versions to show any meaningful results. That’s why we suggest testing just two pages at a time — but it’s your call!
Similarly, it’s best to not test too many elements at the same time.
If you’re running a test where you swap out the images, headline and form fields on an order page and your new version performs better, you won’t know exactly why that is. Was it the images? The form fields? The headline? Some combination of all three?
The best thing you can do is test these elements one at a time. Test your headline first, then you’ll know if the headline was what impacted your results. Then do the same with the image, form fields, etc.
Now that you’ve learned how to set up split tests in Ontraport, you’re ready to start experimenting with your marketing and driving real, positive changes for your business!
Features
© Ontraport 2024
playlist_add_check
PCI DSS, Level 1
security
Privacy Shield Certified