Unlock Reports:
Industry-wide studies on consumer behavior from the Cohley Insights team.
Social content testing is when you experiment with different content styles in order to figure out what connects with your audience. We spend so much time in a digital social space. Finding content that resonates and actually connects, is so important.
In this article, we’ll cover:
Why is social media testing important? It provides the data you need to make posts that reach your goals. You develop a few pieces of content that you think will work and then test them out to see which drive the best results. Then, you look at what your top-performing content has in common and make more of it.
Social media is a platform that changes quickly, and granularly. At the end of the day, you need content that will resonate with a specific set of people, at a specific time, on the right platform, and achieve your goal (ie. more followers, comments, sales, or click-throughs to your website). There are a lot of moving pieces there, so sometimes the easiest way to navigate it is to get out there and try a few things out.
There are a lot of ways to test social media content. You probably already do some of your own testing on your personal accounts. For example, like most of us, you probably want more likes on Instagram. (Not to say that everyone is just out for likes. But getting likes releases dopamine, so a lot of us go for those likes without even realizing that’s what we’re doing.)
You’ve maybe heard that Instragram photos of people’s faces get more likes. You might then start posting more photos of yourself and your family. You also might notice your travel photos get a lot of likes, so you post a lot of travel throwback photos. This is a type of social testing. You’re tailoring your content to what gets the most engagement.
But when we’re running a high-profile account, that’s tied to a business, brand, or influencer, we want to be a little more intentional about our social media testing. Luckily for us, there are a few best practices already out there.
Social content testing is adjusting what you are posting to fit your audience and reach your goals. Having a goal, and tailoring your content to that goal is super important. Why? Because content is the main thing your audience is consuming and expecting on a platform. It’s how you reach and connect with your audience in a relevant way.
Most social media platforms are built around content models. So, understandably, your audience on those platforms is going there to find the content that the platform delivers. Your job then is to create content that makes sense with the platform and meets your audience where they’re at.
On Instagram, your audience sees and engages primarily with visual images. Image content is what the app is built around, so it’s what your audience is going there to find. On Youtube, your audience engages with video content. Video then, is the primary delivery method for your message. That message, and the content format it’s in, then have to somehow achieve your goal.
Lucky for us, we have a few of those goals and their top content already mapped out.
Testing your copy means you are not fundamentally changing the content you are posting, per se, but you change how it’s presented. Social post copy includes things like:
Testing your social post copy is important because it’s what shapes your audience’s first impression, and also what a social platform’s algorithm will use to decide if your post is relevant.
Post timing is all about knowing your audience, and when they’re using the platform you’re on. Testing out which time is best is important because you want to meet your customer where they’re at as seamlessly as possible, and reach them when they want to interact with your brand.
A/B testing is the process of trying out two different versions of a post on two different halves of your audience. Meaning half of your audience will see version A, while the other half will see version B. To be successful, you want to make sure that group A and B are selected randomly, and are even. Then you can see which post performed the best with your audience. A/B testing can be built-in natively on platforms like Facebook Ads Manager, or created externally for an email campaign, for example.
Multivariate testing uses the same concept of A/B testing, where you change one thing between each post to see what resonates more. The difference is that multivariate testing uses more than just 2 options, and changes more than just one thing.
Running a social media test can look so many different ways. But fear not! There are a few social testing best practices to keep you on the right track.
What are you hoping to achieve? Are you going for likes on a specific post? Or likes each week? Do you want to see how much traffic you can drive to your website? These are all different goals and will shape what you post and how you test it.
In our work with Clorox owned Kingsford Charcoal, our goal was to expand to new audiences, by increasing impressions.
We cannot emphasize this enough. Your audience is who you are trying to reach and connect with. You are essentially forming a relationship, that you hope will last so that they keep coming back for more. Figure out what they like and dislike, what their interests are when they’re on social, and what platforms. Then, use your testing process to get to know them even better.
With Kingsford Charcoal, we knew the audience was working class, and family-focused. Kingsford Charcoal did a little digging, and their data found that this particular demographic loves to see grilled hot dogs with perfect char marks. Knowing that audience, changes the type of content we are going to test.
What is the status of your goal’s metric right now? You need to know where you started to see how far you went, and how well your content is performing. It makes total sense, but it is super easy to forget if you’re not careful.
With Kingsford Charcoal, we were able to see how many impressions they were getting on different channels and platforms, and build off of our initial data.
Check-in on how your different content is doing. For shorter-cycle content like Instagram or Facebook, check-in much more regularly than longer-cycle content, like blog posts, or even TikTok videos, that circulate for a longer time.
Cohley and Kingsford Charcoal ultimately decided to Kingsford generate 229 unique assets via 11 highly talented content creators. Each one of these talented content creators did a mini photoshoot for the Clorox brand. (And yes, there were some hot dog photos!) They then were able to post those results, and see how each image was performing.
Test when your audience will respond best. It does not make sense to test a holiday post in July, even if you want to know which post will be better for that December. It doesn’t make sense to test how many people will click on a sleep meditation app at 11 am, when you’re targeting night-sleepers.
King Charcoal’s content push was super timely for Clorox, since we could see that King Charcoal wasn’t ranking as highly as it’s competitors on Google. And hello, that needed to change! We also were able to post summer images in the summer months, when we knew interest in grilling would be high.
A hypothesis is your starting theory or explanation, based on the limited information you already have. It’s the starting point for your investigative testing. In the context of social, take posting an Instagram photo as an example. Your hypothesis might be that more hashtags will get you more impressions.
To go back to a real world example, with King Charcoal, our hypothesis was that the more content we were able to generate, the more impressions we would get.
Your variable is what you are testing, and the one thing you are changing across tests. Using the previous example of using more hashtags to get more impressions, the variable you are changing is the number of hashtags.
For our work with King Charcoal, the variable was the number of impressions.
The control variable is what you are not changing across tests. It is the unchanging standard. In the previous example, you are not changing the photo you are posting, just the hashtags. So the photo is the controlled variable.
The controlled variable for our King Charcoal content would be the target audience. We know we are trying to reach family-focused, working class folks, so we won’t change that across tests. We’d run another test with different content for different audience groups.
Metric is how you are measuring your goal. Again using the above example, if your goal is getting more photos from your brand on a users’ Instagram feeds, the metric you are likely measuring is number of impressions.
Impressions are the actual metric we are measuring for King Charcoal. Although keep in mind that most campaigns have multiple metrics. We might also want to measure new followers, or new website users from social.
Statistical significance can get really math-y real fast, but essentially, it’s figuring out if the difference between your tests and their results are large enough to matter. This our Instagram photo example, this would be taking the number of impressions from each hashtag test, and calculating to see if the difference between impressions is random, or actually due to the number of hashtags.
The best way to run a social test is to test as much content as possible! To know your audience and what makes them tick (or rather, click!) takes time. They need you to post enough content that they like for you to be able to see a clear pattern.
Cohley makes content marketing easier. We partner with brands to shape their content strategy across platforms, no matter what your goals and needs. We generate content ideas, find ways to expand, and build content partnerships.
Apply to Become a Content Creator