We talk a lot about A/B testing at Fujisan because it’s really the backbone of improving your work. This makes sense — after all, it’s pretty much just the digital marketing application of scientific testing with controlled variables. You remember sixth grade science class: By testing between two versions of a situation with all variables controlled for except one, you can then see which modification to the uncontrolled variable leads to a certain outcome. Same here!
With A/B testing as a digital marketer, you can compare the results for a variety of factors. For example, creative, copy, locations, audiences, landing pages, and other variables.
Testing gives you concrete evidence, allowing you to then make decisions that are based on hard data instead of a hunch or educated guess, no matter how well-founded. From there, you can adapt to the outcome of an A/B test to improve user experience and, through extension, your own performance.
Now, A/B testing can be done in many different contexts on many different platforms — today, we’re taking a look at this application on TikTok.
A look at why A/B testing on TikTok is important
Essentially, testing is best practice — to identify best practice moving forward. If you’re running TikTok ads, this will help you determine which practices for any given ads work best.
We should note, however, that it’s important you ensure you have a large enough budget to properly test. Otherwise, with too small a budget you won’t be able to have a big enough sample size to draw valid conclusions from, leading to your results being only partially helpful at best, and actively damaging at worst.
Assuming you do have enough budget though to run statistically significant A/B tests, why is doing so on TikTok specifically important?
There are a few main traits it has that can be beneficial:
- TikTok’s trends evolve quickly and ahead of many other platforms, allowing you to get effectively the first insight into what’s relevant — and isn’t.
- Audiences vary on different platforms — what you find through A/B testing on Facebook can be drastically different than your findings from TikTok A/B tests.
- Maximum engagement — TikTok is a highly engaged platform, so once you establish what works (and what doesn’t), you can get the most out of your improved content and targeting.
What you can test on TikTok
There are two main umbrella categories you can test for: creative tests, and offer tests. These are pretty much exactly what they sound like — tests for creative are testing the contents and characteristics of the ad, whereas offer tests are testing for what the ad itself is offering.
For an offer test, there’s really only one category here — does the audience prefer this offer, or that offer?
For creative tests, there are more subcategories. These include:
- Video content: testing the focus of your video to identify what content resonates the most with your audience
- Video length: testing whether a shorter or longer-form piece works best for your audience
- Ad copy: testing to see which in-video copy, external copy, or CTAs work best
- Targeting: testing to see if people in different locations, with different interests, customer lists, and other defining TikTok audiences are more interested in your ad
Regarding the last one, TikTok also allows targeting based on user behavior, which can be useful.
Options for how to test on TikTok
There are two different ways to test: You can set up your own tests manually or set them up through TikTok’s split testing tool.
For the first option, you create your test within TikTok Ads Manager, using separate campaigns, ad groups, or individual ads depending on which you’re testing.
For TikTok’s split testing feature, it allows you to run two different versions of an ad with three different variables you can choose to test: targeting, bidding and optimizations, and creative. This tool then begins the A/B test while making sure audiences only see one version of the ad to prevent results from being potentially contaminated.
Challenges of A/B testing on TikTok
While A/B testing is useful anywhere, it’s important to be aware of some of the difficulties of doing so on TikTok.
In general, it can be difficult to create a clear-cut A/B test since, while you can control all the variables you can, there are still variables outside your control that mean you can’t necessarily accept the findings as 100% gospel.
Also, while the split testing tool makes A/B testing super easy, we typically recommend doing the first manual setup option — in doing so, that gives you as much control as possible and limits potential automated factors from split testing that could affect the validity or accuracy of a test. Although that takes a bit more work initially, when done right the manual setup tends to give more reliably trustworthy results.
The last challenge to be aware of — and we alluded to this earlier — is making sure you have enough budget and time to devote to an A/B test so you can get a significant sample size. Too small a budget or not enough time, and you’ll have results from a sample size that’s too small to tell you anything insightful and could even impact the test’s performance so drastically as to actually show you the opposite of what’s true. To that end, be sure to have a clear plan and establish what you can budget — both money- and timewise — before embarking on an A/B test.
In the end, A/B testing is a valuable tool when done right because it isolates variables to give you better insight on just one factor than you could ever get when other variables are involved. TikTok can be especially useful because of its uniqueness in many ways among social media platforms. At the same time, you have to be aware of the limitations of A/B testing and TikTok’s tools for doing so, so that you can get valid, useful results.
