Split | Testing
Splitting up is a good thing
Usually it’s wise to uphold the idea that, “What (God, nature) has joined together let no man separate.” But in the case of marketing, splitting up ads is more like cutting a kid’s birthday cake. Some kids want a piece with sprinkles and others want the piece with pink or blue frosting. Whenever you set the entire cake in front of all the kids at once, most of them don’t think “I want the whole thing!” (Well, some don’t…) Most kids will hone in on the piece they think looks most appealing and beg for it.
So we come to an important fact about managing a successful ad campaign: if you run multiple ads at once and allow your audience to pick their favorite one you will continually improve the response you get.
This, if you haven’t already surmised, is called split-testing, or A/B testing. (Though, I prefer split-testing for its reference to a particularly delicious combination of banana, ice cream, and other delicious accoutrements.)
Creating a successful marketing message isn’t just dumb luck. You can approach the process of creating and finding a successful ad in a scientific manner. Split-testing allows you to methodically put your ads in competition with one another, allow your audience to choose their favorite, and repeat.
Many people who create ads out of years of experience selling to their customers think, “this was a great one, it’ll definitely work!” Then they’ll pay to stick the message out there. Famous last words. Likely it will get some kind of a response, but if you didn’t just become a lucky winner, you’re just as likely to have written a dud. Until you split-test that ad that seems so well-written, you won’t ever know.
This is why split-testing is vital for you (or your marketing manager) to consistently use. If you aren’t approaching your marketing analytically, you are probably losing money.
How To Split Test
The key to effective split-testing is to test both small and big changes. Start by writing and running two (or more) ads that change only a couple of words, or the word/sentence order. After you find a winner, try to beat the winner by creating one or two more ads that are dramatically different and have them compete against the winner. When writing your marketing messaging always keep in mind the things that are important to your target audience. See this article in my blog for more help on writing ads.
Here is an example of a subtle change between two ads that were tested against one another:
Which one do you think was the winner?
Ad #1 Want To Patent Your Idea? Patent Your Business Idea Now. We Identify And Prioritize Patentable Inventions. Call For Free Consultation!
Ad #2 Patent Your Business Idea Now. Get Help From Expert Lawyers. Expert Patent Lawyers For All Types Of Inventions. Get Your Free Consultation!
The winner was Ad #1, with 5x more conversions than Ad #2, in spite of the fact that the ads were very similar to one another. Here is an example of a significant change between two ads that were tested against one another:
Which one do you think was the winner?
Ad #1: Business Attorney in (City Name). Helping Small Business Owners. Get A Free Evaluation From Trusted Attorneys To Protect Your Company's Assets.
Ad #2: Business Attorneys (City Name). Expertly Growing Your Business. Creative, Cost Effective & Timely Results. Predictable, Agreeable Flat Fees!
The winner was Ad #1, with 2x more conversions than Ad #2!
Now I’ve shared about the strategy for how to effectively split-test. But how do you actually go about doing it? I’m going to show you how in two platforms: Adwords and Facebook.
In Adwords, the process to split-test an ad could not be easier. Go into an Ad Group that you have already in place and simply click the red +Ad button to create a 2nd ad.
In Facebook’s Power Editor, adding a 2nd ad (within the same Ad Set) looks like this:
The key thing is to simply have 2-3 ads running in each Ad Group (for Adwords) or Ad Set (for Facebook). There are of course many other ad platforms out there, but they all make it easy to do this.
Once your ads have been running for enough time to get a statistically meaningful response, and your Analytics are set up correctly, you’ll see which one is bringing in more customers & revenue. How do know if your data is statistically significant enough to declare a winner? Here is one tool, and here is another. Basically, input the number of clicks and the number of conversions, and these tools will tell you if you have run the ads long enough/have enough data.
Split-Test Your Landing Pages
If you’ve read my article on Landing Pages, you already know how important they are for turning a visitor into a customer. The awesome thing is that you can split-test landing pages as well. Most, if not all landing page software lets you create an iteration of your original page, and will then automatically send an equal % of your traffic to each iteration. (Sometimes you can control what % of visitors go to each page, like in the software Unbounce.)
Here’s what that looks like in Instapage:
And here is a small section of the results page inside Instapage shows you, so you can see which iteration is performing better: