Creative testing for TikTok and Meta Ads

Creative is the biggest lever we have in performance marketing. It's no longer a secret. But does this mean open targeting, creating a few creatives and the Meta and TikTok algorithm already handles the rest, which is the golden path? It is not quite as simple as that. That's exactly why we're introducing you to Creative Testing in this article.

What is creative testing?

Creative testing is fundamentally about finding out which version of your ads is best received by your target audience and thus brings you the most revenue.

It all works like this: You create different versions of your ads with different images, videos, texts, and headlines. Then you let them compete against each other to see which one works better.

The idea is that by testing, you can find out which parts of your ad are doing well and which aren't. This information will then help you improve your ad and get the best results.

Why should I do creative testing?

1. Improve the performance of your ads: Creative testing is the best way (and maybe the only one) to find out what your audience really likes. By continuously testing and optimizing creatives, marketers can achieve better results.
2. Avoid ad fatigue: Over time, the same creatives can become boring and lose their effect. By testing and switching between different creatives, ad fatigue can be reduced and potential customers can be addressed in a more appealing way.

How can a creative test run?

Basically, the start of creative testing is a hypothesis that you have about your product or service. WHEN I communicate X in my creative, I expect Y BECAUSE Z.

But first, let's name the complete creative testing process:

1. Setting up a hypothesis
2. Test your hypothesis measurably
3. Evaluation of creative testing
4. Creative testing iteration

Now let's look at the individual components:

Setting up a hypothesis

There are several ways to create a hypothesis for creative testing.

Based on existing data

You have already gained experience in your Facebook or TikTok advertising account and therefore know what has worked well and less well so far.

Based on the existing data, you can make new hypotheses. For example, you've recognized that creatives that highlight a single product achieve more sales than creatives that promote the entire range of products. From this, you could form the following hypothesis:

WHEN I promote individual products in a video, I get more sales BECAUSE the previous data shows that.

Based on ideas

If you're running ads for your or a company for the first time, then of course you don't have any ad data in your advertising account yet. Therefore, you must make assumptions or brainstorm ideas on how you can ideally present your product or service to your target group in the best possible way.

It is best to start the thought process from the customer: What problem does your product/service solve and how do I show this added value to the person while they are scrolling through the feed without suspecting anything? It's best to go by yourself. What would you pay attention to when scrolling?

If you need additional design input, the meta as well TikTok Ad Library an ideal place to be inspired by competitors or other advertisers.

Based on your brainstorming and researching the competition, you can then make initial hypotheses.

For example:

IF my ads have the same layout as the competition, then we perform well BECAUSE the competition knows what they're doing.

Once you've made a hypothesis or you've defined different creatives, you can move on to the next step.

Test your hypothesis measurably

In order to be able to measure in principle, a clean implementation of the Meta and/or TikTok pixel is of course crucial. Before you invest a lot of money in creative testing, it's therefore worthwhile to check the tracking setup again.

So that we can also analyze the creative tests separately from the other active campaigns, it makes sense to set up a dedicated testing campaign. Personally, I use a testing subscription (Ad Set Budget Optimization) campaign to give each creative a targeted budget. Of course, you could also set up a CBO (Campaign Budget Optimization) campaign and also define several target groups at ad set level. For example, Interest+ Open Targeting. But since we want to focus 100% on the creative in this testing, I would recommend a ABO set up.

Within your ad sets or ad groups, you are now testing different creative concepts. In the best case scenario, it is important that you only change one variable. Otherwise, it will be difficult to make a meaningful decision as to why this creative has now performed better than another. This could be, for example, a different hook or a different communication style.

In our contribution to the Naming Convention We have already explained why this is so important. A clean naming convention is particularly beneficial in creative testing. It not only helps you understand more quickly which settings you have made, but it also gives you information about which creative elements and concepts you are testing at the ad level.

Through a targeted naming convention, you can therefore identify your winning creative combinations. In other words, which products work best in which way of presentation. Is a voice over about an individual product that highlights the advantages the best or is it more of a classic picture of the product range selection with an appealing headline that describes the problem? Questions like this can be easily answered via a clean naming convention be answered.

What should the creative tests look like?

Depending on the available advertising budget per month and creative run, various ads should be tested. With a smaller advertising budget (<10k per month), rough concepts such as UGC vs. high-quality videos should be tested against each other. With larger accounts, “minor tests” such as different headlines can also be carried out. In general, it should be said that you should move more and more from Macro Level to Micro Level. In other words, the larger patterns such as content art (video vs. image) and communication style (pain vs. value) should be tested first. Before you include minor topics such as background colors in testing.

Evaluation of creative testing

A question that comes up again and again in creative testing is: How long does the test have to run and how much budget do I have to spend on it? Due to the delay of the data displayed in Meta Ads Manager due to the many data protection measures alone, it is recommended that creative testing run for at least 3 days. This also excludes individual daily trends. Of course, a creative test that is played at €100 per day is more meaningful than a test with €5. That is why it is definitely recommended to generate 5,000 - 10,000 impressions during a creative test. In the best case, over several days. This can vary quite a bit from industry to industry and even account to account, as this is 100% tied to the CPM.

In addition, it is of course advisable to pay attention to soft metrics such as CTR and thumbstop ratio and to validate how these metrics turn out after the first 3,000 - 5,000 impressions. As a result, early testing can be scaled or stopped.

Which metrics should be looked at in general?

At the start of every creative test, a “success metric” should be defined. Say a metric on which success depends. In most cases, these are probably general conversion metrics such as Purchase, Lead, Add to Carts, etc. However, you don't just want to trigger direct sales or leads with every ad. Especially for products with a longer customer journey, people are already happy about good click rates and certain interactions on the website. Metrics such as CTR, thumbstop ratio, soft conversions such as view content and add to carts should therefore also be included in an assessment of the creative test.

When was a test successful?

This is of course a very individual question and depends on the individual case, of course. However, as a marketer, you are happy about every increase in results. Even a 5-10% increase in CTR can be a real “win.”

With these many small tests, the big picture should never be forgotten. For example, if the CTR has improved immensely and a sales uplift could be identified even though the tracked conversions were lower, this may be due to the test. Especially when looking hard at conversion metrics, the entire attribution and tracking issue should not be neglected, as this can falsify the results.

What happens to the “Winner Creatives”?

Even if the test has been completed and a winner has been identified, it would not be recommended to pause the campaign or ad set. Personally, I keep running an ad set until I'm no longer satisfied with the results achieved. In addition, the “Winner Creative” can of course be duplicated into a Scaling or ASC (Advantage+ Shopping Campaign) campaign.

Identify patterns

If you use a clean naming convention and have always maintained it diligently, certain patterns can usually be identified after a variety of creative tests. You can use these patterns using tools such as DatAds, easily identify games. When you connect your ad account, DatAds scans your ad level naming convention. If terms come up more frequently, DatAds suggests them to you as groups. As a result, all creatives who have a certain thing in common end up in these groups. For example, it could be that all creatives were created by the same creator or the same hook was used.

As a result, comprehensive analyses can be carried out more quickly and patterns can therefore be identified more easily.

Here you can test DatAds free of charge for 14 days.

Creative testing iteration

Even if you've identified a winner, the topic of creative testing doesn't stop. New hypotheses should be developed based on the learnings in order to get new tests back on track as quickly as possible. Existing winners can therefore be further optimized and ad fatigue can thus also be actively counteracted.

It is important to say that creativity should not be lost when it comes to iteration. In the best case scenario, only one component is changed in Creative Test, but new concepts that take a completely different approach should also be tested. These tests can also be documented in a testing set up.

Conclusion

Creative testing is a way to bring targeted structure and scaling potential to an ad account. Through creative testing, patterns can be identified that are beneficial to your advertising goals. It is important to define a clear success metric and test it over a longer period of time. Even though best practice requires changing only one variable per test, one should not forget to intersperse new creatives that have a completely different style than previous top performers over and over again. Most of the time, even these creative outliers are the ads that take the account to the next level.

‍