A term frequently used in the digital marketing industry and a necessary process in order to get the best results possible for your marketing efforts and campaigns.
After all, How do you know what works best until you have tried multiple solutions?
When the term “Split-Testing” is brought up in conversation, there are usually three responses you will get.
- “What on earth do I need to be split-testing?”
- “Split-Testing? Completed it, I know what I’m doing.”
- “I’m split-testing multiple aspects of my marketing campaigns but how do I take action from it?”
This is because common mistakes consist of ignorance to metrics, thinking you know more than algorithms and not being able to take valuable insights from the testing you have implemented.
The most important element to split-test in your digital marketing campaign is the offer/product or service. If you’re confident you have a proven, compelling offer/product or service, then keep on reading...otherwise see our post on successfully split-testing offers here.
For anyone scratching their head when it comes to high-impact testing ideas, we are going to reveal Elements Of Your Facebook Campaign You Need To Be Split-Testing
Below are five of our favourite variables to test for improving performance and profitability.
- Ad Copy Length
- Creative Type
- Creative Style
- It’s a common mistake speaking to multiple audience avatars in a single piece of ad copy.
In reality, the saying “if you’ve got one arse, don’t try and sit on two horses” is correct. It’s better to focus on a single customer avatar in a given piece of copy, as they usually have unique wants, desires, challenges, and pain points.
This enables you to really zero in on exactly that will truly speak to them, rather than trying to appeal to everybody and ultimately, appeal to nobody.
We call this the “angle” of the copy. Taking your ad copy and breaking it down into two or three individual angles that you really go to town. When creating your angles, think about the main pain points of your avatar as well as the key benefits to your business.
For example, testimonials are a great asset to utilise in ad copy. Social proof holds great weight when it comes to influencing buyer decisions so if you can take genuine reviews and feature them within your ad copy, you can highlight benefits of your product/service without sounding biased.
If your product aids back pain (or any pain point), make this a focal point in one of the ad copy variants. Use research and stats to back up your points and address the daily niggles of someone with back pain faces. Use this to pitch your product as the ultimate solution to their problem.
Additionally, in retargeting campaigns, promote features such as payment plans and money-back guarantees which can help the consumer with their decision and might be the reason they don’t go to a competitor.
We find that different angles provide some of the biggest differences in cost per acquisition across our campaigns. Get to know your audience and what drives their buying decision!
Ad Copy Length
-The length of your copy can radically impact your results. Most advertisers run short copy, and there’s nothing wrong with that, but it can often lead to higher CTRs and lower conversion rates.
This applies for small-medium businesses in general. The likes of Nike can retarget you with a carousel ad featuring just the words “Shop Now” and you’re still just as likely to purchase that new pair of trainers.
By running nothing but short copy, you’re only catering for the super enthusiastic consumers who click immediately. There are plenty of potential buyers who need more information and reassurance before completing their checkout process.
Long copy enables you to increase purchase intent and give personality to your brand so the destination page doesn’t have to do all the work and the consumer feels more confident in their decision, which usually results in repeat purchases/continued business.
For more tips on how you can nail down on ad copy see this post...click here
- Your audience will resonate with different types of media differently. It almost goes without saying in this day and age, but you need to be testing both images and videos.
We prefer to present these in the single image/video format versus carousels. But have seen the latter work in some cases when used creatively. Typically we will see videos get much better CTRs than images, with the conversion rate influenced by the quality of the content.
One of the reasons you need both is, both creative types offer something different. Ad images (when done right) can perform well because one picture tells all and you don’t have to rely on engaging someone for them to consume all the content (as a video would).
Alternatively, videos have the benefit of bringing your brand to life, building a rapport with your audience and visually capturing their attention. Most advertisers start off with images which is fine, but if you are looking for mass appeal here’s a basic formula to attract interest at large:
Demonstrate mass problem + Agitate that problem + present your product as the only solution = Sale made
It’s super simple, but you can get pretty creative inside of these guidelines.
Let us give you a video example:
0:00 - 0:03 = Person is struggling with issue your product solves
0:03 - 0:06 = Agitate the problem by introducing subproblem
0:06 - 0:10 = He/she holds your product up to the camera
0:10 - 0:15 = They are using the product
0:15 - 0:20 = Showing how the product helped
0:20 - 0:25 = Person is telling story about how the product changed their life
- If you’ve split tested your images extensively, have you considered the patterns among your best and worst performers? We tend to focus first on “natural” images versus “professional” images.
We loosely use the term “natural” and “professional” images. The obvious difference is a creative clearly created by a graphic designer and an image a business owner has taken in selfie mode on their iPhone. A common assumption is the latter should never be used in advertising campaigns which is a huge mistake as personable content can fly.
Natural images will tend to have higher CTRS, but professional images enable you to objection handle much more effectively assisting conversion rates.
If you strip Facebook down to it’s core, people are on there to interact with their friends and join the latest Facebook group to give them some accountability. In short, people don’t like being advertised to. This is often why brand videos and professional content underperform. With that being said, professional content has its perks. Especially when it comes to branding, and reputation.
With all of that in mind, getting granular with these creative types can make the difference between someone clicking or someone purchasing. So what can you split-test…?
- Images taken on your iPhone (Business owners smiling with clients, at family events or on holiday)
- Professionally designed graphic images (highlighting product features and benefits)
- User generated content - (Customer reviews, clips of customers using the product or talking about it)
- 6 second clips GIFS and videos to increase the engagement of an image
- Instagram stories reminding people to complete their purchase
You can go into much more detail - The model male or female, what angle are they shown at, what is the background colour - we could go on. Look for patterns to form hypotheses, and test on that basis.
Do the same for videos - is it long or short? Is it a “talking head” video, or featuring the product silently? Consider the patterns across your videos and let the performance inform your decisions. And don’t forget to check the video engagement stats - you might find cutting a slow intro or re-ordering an existing video is what’s required, rather than creating totally new content every time.
- When launching any campaign, we will always split-test interest-based, lookalikes, and broad audiences. We like to keep interest audiences as broad as they can be before creating smaller, targeted interest-based audiences. This way we can ensure they will be early performers.
Before going crazy with creative testing, you’ll need to know which audiences are most receptive to your offer. You could have the most engaging video or eye-catching, benefit driven image but if it’s placed in front of someone who wouldn’t try your product/service for free, it’s a great gone to waste.
There are a variety of cold audiences you can create and test on Facebook particularly. Interest-based audiences give you control by being able to target people who are interested in Brands and Facebook Pages that you select. You can compile a list of similar products or companies to form a niche audience.
Lookalikes take away the research task for you. Facebook will find similar people that have taken action on pages that have your Facebook pixel on them, whether that be purchase, add to cart, initiate checkout, lead or even view content. You can also decide what percentage of these audiences you want to target. For example you could target an audience that is similar to the top 1% of people who buy from you.
Finally, don’t discount a broad, non targeted audience. Let Facebook do the heavy lifting and see if it can outperform your audience creation efforts.
From spending millions of pounds on Facebook, here are the insights that we find to be true...but this still means you should split-test to find out for yourself!
- Pixel conversion lookalike audiences are high-converting and consistent
- Customer lookalike audiences are great for increasing average order value
- Do not underestimate broad targeting and always test non-targeted audiences
- Utilise competitors and similar brands as interests in interest-based audiences
- Use Facebook’s “Audience Insights” to inspire new audience ideas
SO there you have it.
5 elements of your Facebook campaign you need to be split-testing. Now the common questions tend to be “how long do I split-test for?” Or “how do I determine top performers?”...
Well it depends on price point, amount spent and key performance indicators. As a general note if you’re spending more than £350 daily, you should be able to see results within a week or two. Determining top performers depends on your objective. It will more than likely be cost per acquisition. Set a target CPA and compare the results between each element after two weeks of testing!