Split testing is extremely important when you’re doing any kind of facebook advertising, because this is how you continually improve the results of your campaigns. I’ve had campaigns that failed when they were first launched, but then after split testing of the different e elements of the sales process, they quickly became profitable. You can do the same thing. So don’t worry if something isn’t working right away, you need to analyze each portion and then start testing different elements to improve. It really becomes a math game at that point.
We are going to talk about:
What is Facebook split testing.
What to split test to maximize Facebook advertising profits.
What Is split Testing?
Split testing is simultaneously presenting two or more different versions of a single Item of a marketing funnel. When I say simultaneously, I mean running your ads at the same time. For example, you’re running an ad for an affiliate offer, but you don’t think you have the ability to run ads simultaneously. One week you’re going to run the ad with one headline, the next week you’re going to run the same ad with a different headline and then compare the two after the fact.
As an exaggerated example, imagine you’re running that campaign the week before Christmas and the week after. The week before is the first edition you want to test and the week after is the second edition you want to test. Obviously the results from those two different time periods are going to be dramatically different, but not because of the element you’re testing, not because of the headline. It’s just because of what’s going on in the World on those specific dates.
Those kinds of things can be happening in a much smaller manner all the time. If you’re running ads on Facebook, one week you could have no competitors and the next week you could have a bunch of competitors and those can really skew your results. You’re not really getting good data to know what’s working with your split test. This is ideal, but that’s not always possible given the technology that we have, but ideally, the tests are running at the same, really splitting per visitor: one visitor gets version A the next Visitor gets version B.
You need to have two or more different versions. Most of the split tests I run are just
two different versions. For example, I may run a split test that has two different sales videos and I don’t try to complicate it with too many versions. Each additional version you add is going to take you longer to find out which one is working.
Say for example that you’re only getting a hundred visitors a day and your split testing two options. That’s 50 visitors per day, per option. But if you test 10 different versions, that’s only 10 visitors per option. If your conversion rate is, say, 1%, it’s going to take you forever to figure out which version is actually working.
Typically, I run two different versions of an element, for example, two different versions of a sales video and two different versions of a headline or different elements that I’m going to explain in a minute. You can run multiple tests. I’ve had it before where I was testing three or four different landing pages at a time. You just need to know that it will take longer to figure which one is working.
When I say a single item, you can run what’s called multivariate tests, meaning you’re testing a headline and a different sales video and then maybe a different kind of call to action button on the video, for example.
Those elements will work differently together, so you’d have tons of different versions you’d be testin all at once. The problem with that is, again, it takes you longer to get data and you really have to have the technology to be able to pull out which combination of those elements is working the best.
What to split test
Here is what to split test. It isn’t everything, but I will give you a good idea to start. This is really going to open your eyes, knowing that no matter what results you get out of the gate, you can get somewhere much better by iteratively testing all of these elements.
This sales funnel here is a Facebook ad going to a page post, going to a landing page, going to an order form, going to the upsells and ending on a thank you page. If you’re promoting a product, an affiliate product or a product on Amazon.com for example, anything like that is going to change this funnel slightly, but the ideas and concepts remain the same.
So, first off, your ad. In your ad you can test your headline, your description and your image. For example, I can tell you that there is not much that is good about this ad in particular.
The image is pretty horrible; you can’t even read what those words are or what the image even is. It does have good colors; black and yellow tend to stand out well, you’ll see it on a lot of billboards. These are the different elements you could be testing.
With Facebook ads, you want to be running a maximum of two ads per campaign. For example, one ad could have one image and the next ad a different image. You’re going to want to start splitting up per campaign. You can, and always should, be testing the image, headline and description.
If you’re driving people to a page post you can be testing different images to find out which one of those works better. A place where I’ve found massive improvements is In the text and call to action for that page post.
Really writing that hard-hitting sales copy and including multiple strong call to action can really improve results. Getting people to that page is the first thing, but getting from that page into your actual landing page or website, that’s where a lot people fall short.
3. Landing Page
Next we have the landing page. This is going to depend on what kind of landing or pre-sale page you’re using, but basic elements include testing your headline, testing your sales presentation method and testing the additional sales elements.
The additional sales elements in the example are a text below the video and a countdown timer below the video. You can also test how long the buy button is delayed if you’re using one. Just about anything else you can think of adding benefits, features, testimonials, any of that kind of stuff can and should be tested along the way.
4. Order Form
Next is the order form. If you’re using your own you can always test different things or if you’re using a ClickBank form for example, you can test out things in the header area.
So, the first thing you could test is the header area. You can create your own graphic that makes people feel like they’re still on the same site. You should have all the different elements at the bottom of the page: 60-day money back guarantee, secure order, instant access worldwide.
You could test the different trust symbols by running it with one and running it with 10 different symbols to see which gets better results.
You can and should always test price and price options. For example, rather than selling their product for $49.95, it could be tested at $69.95 or $39.95, whatever price you want that could possibly make sense.
Also test price options. It doesn’t necessarily apply to digital products, but if you are selling a physical product, don’t just give them the option to buy one, give them the option to buy 3 or 6 or 10. Any of those things can dramatically improve your results and should be tested.
You can add testimonials, features, benefits and images of the product. These things can and should be tested on the order form. If you’ve done a good job With your advertising, with your targeting, with your landing page and your video sales letter, a lot of times you can get a massive boost in results and profitability by improving the order form.
At this point people are ready to pull out their credit card, they just need to feel comfortable and reassured that they should do so and their credit card information isn’t going to get stolen. Whatever you can do to make them feel comfortable buying from you, can reap huge rewards in your business.
In upsell s you can test which upsells to offer. For example, you are selling a weight loss eBook, you could test adding a workout eBook as your upsell and you could test adding a recipe book. One may outperform the other, so you could test what types of upsells to offer. You could test upsell pricing; whether people are sold those upsells for $17, a low-end or $97, a high-end, anything like that is good to be testing.
You can test the upsell sales method. You want to present them with more videos. We see this a lot online, where someone is selling upfront via Video, they typically sell their upsells via video as well. But that doesn’t mean it’s the right way to do it for your market. You could test a video sales letter vs. a text sales letter for the same upsell and see which one performs better.
5. Offer Testing
We all know your Offer is critically important to whether or not people want to buy at all. You could test what your guarantee is. Are you giving a 30-day or 60- day guarantee or you could try giving a 10-year guarantee or a lifetime guarantee, test those different elements.
You should also test your scarcity. Are there are limited number of units available? Is there a limited time period within which they have to offer or within which the price is going to go up afterwards?
What bonuses are you offering? Maybe some of those upsells would help your offer perform better if they were thrown in as bonuses in the initial upfront offer.
Test whether your purchase method makes you more money sending people to your own shopping cart or sending them to Amazon.com or any other third-party site. There’s a big trust-factor built into a site like Amazon, but you have more flexibility with your own shopping cart. Neither one is the right answer; it’s just all about the numbers, which one makes you more money?
So, this is a lot of stuff you could be testing, and it’s not everything, but like I said, it may take you a year to get all of those elements testing. My rule is you always want to have one test running for each component of your funnel.
So, one test running for your ads. One test running if you’re using a page post. One test running for your landing page, like the headline or video. One test running for your order form, like different trust symbols and pricing. One test running for your upsell sequence and possibly one test running for your offer if it works in with these other tests.
Doing this’s going help you constantly improve the results of your marketing performance.
Split testing is extremely important because it is how you improve the results of your campaigns. Remember to test single items in your sales funnel, but be sure to test multiple elements to get the best results. There are several different things you can test in each portion of your sales funnel don’t neglect any of them.
Please share your thoughts and comment below. I would love to hear from you any questions you may have.