Implementing social media A/B testing can help turn your social media channels into high performing powerhouses. A/B testing provides valuable data into your audience’s thinking and with this information, you can create an optimized content strategy for each social channel.
A/B testing can be used for everything. For example, you can test your website page formats, email subject lines, ads headlines and blog titles. Here we will be focus specifically on social media A/B testing and how to boost your engagement with both organic and paid.
1. A/B testing vs multivariate testing: advantages and pitfalls
First and foremost, what is A/B testing? A/B testing, also called split testing, is the process of measuring two versions of something, in our case social posts, against each other to determine which one performs best.
Unlike multivariate testing, A/B testing only compares variations of one aspect of the post or landing page etc. An example would be testing the same tweet text with two different images. By keeping it simple it makes the takeaways more clear.
Meanwhile, multivariate tests involve two or more element variations. This naturally makes it the far more complicated and time-consuming option, and it requires a considerably larger amount of traffic to generate results.
Multivariate testing is also difficult to perform on organic social media due to the impossibility of segmenting audiences based on precise traits as you can with paid social advertising. This means straight off the bat that any results can never be treated as 100% accurate.
2. Implement A/B – multivariate testing for organic social media
Both types of testing can be done with the help of tracking codes. The most common and handy way to track results is with Google Analytics. Depending on the scale of your tests, adding tracking codes to all links can be a big task. Excel sheets and formulas can help streamline the process so you don’t have to enter repetitive codes. Here at Falcon.io we use our platform’s integrated tracking templates to save time and minimize errors.
"At Falcon.io, we use the integrated tracking templates to save time and minimize errors."
How does this work? By using tracking templates, any post that goes out on a certain page will automatically be generated with UTM parameters and can then be tracked. You can create several of these templates for use in A/B tests or campaign tests. The templates can be applied to any number of new posts so there’s no need to configure Campaign Tracking for each one.
In the example below, the bare link www.falcon.io is published. Once the end user or Facebook fan clicks the shortened link they are taken to a URL with these tracking tags: utm_campaign=CPH&utm_content=label&utm_source=facebook& utm_medium=Falcon.
Once you’ve set up Campaign Tracking and published your posts, you’ll want to head over to your analytics provider to monitor campaign performance.
3. What should you test?
Due to the frequency you can post, organic social media allows you to test multiple elements before you opt to back up posts or campaigns with spend. Here are just a few things to consider.
a. Test copy & CTAs
One basic but vital test you can do is with multiple variations of your messaging copy. By using both engagement and performance metrics, you will be able to quickly figure out which copy variation achieves the best results.
Of course, this depends on the goals behind the post (e.g. traffic, engagement, conversion). We recommend that you leave a few days between posts, but try to put them out at similar times.
In addition, it’s wise to keep the other elements such as the image the same in each test. And remember to adjust your tracking codes for each version.
b. Test images
Images are another good element to test. Imagery is vital – the clincher even – on social media where posts must stand out to get precious audience attention. Here too you should use differentiated tracking codes.
c. Test landing page
Last but certainly not least, landing pages are a key element we test to gauge organic and paid performance. A basic best practice here is that if you know which post image and copy perform best, you should build your landing page based on them. Then there is conversion testing.
Tools such as Optimizely or Unbounce make it easy to set up variations of your landing page. These can range from slight changes such as CTA buttons, to completely different compositions. In the latter scenario each version will be tested on one-half of your audience. Tracking parameters enable you to see user engagement with each variation – and which achieves your goals best.
Keep in mind that the more elements you add, the more post variations you will need to create and the more tracking codes you will need to monitor.
4. Implement results
We’re also A/B testing our paid campaigns with the goal of optimizing our advertising budget.
For example, while promoting a recent ebook of ours, The Essential Facebook Marketing Guide, we created ads with two copy and two image variations. On Facebook we ran two campaigns with one ad set each, with a total of eight ads, each with their own tracking codes.
After two weeks of paid promotion we were able to see which variation was performing best, and from that point focused our spend on it.
The process is similar for organic social media. Once the winning post has been found, it is best to continue publishing that variation. However, in many cases a testing round will inspire new hunches to be pursued through further testing. With patience and diligence you can build a high performing feed based on genuine insights into your audience’s behavior and interests.
Accept some things just can’t be measured
There will always be things to test. It can be very tempting to test absolutely everything. However, it is important to pace tests as well as accept that human behavior is extremely complex and you will never be able to decipher it all. Sometimes a square button will simply perform well against all expectations and for no obvious reasons. The X factor behind its success simply can’t be quantified – and you shouldn’t tear your hair out trying to figure it out.
It is best to focus on the big learnings rather than details. Optimization is an ongoing process and there’s no such thing as the perfect social media post (well, maybe the fabled Oreo one, but the stars were perfectly aligned that day).
Instead of drowning in numbers, analysis and results, be inspired by the unexpected. Ultimately, A/B testing and the learnings gained in the process will help turn you into a performance-driven marketer – and your social channels into important revenue streams.