Targeting best practices to boost performance

Over the last few weeks, TikTok has rolled out big changes to the Targeting flow in TikTok Ads Manager with the aim of improving the functionality and user flow. As part of the update, we are:

1. Bringing the benefits of automatic targeting to manual targeting

2. Introducing new Smart Targeting toggles that allow you to expand beyond designated audience, interest, and behavior settings to improve ad performance

3. Removing targeting expansion and automatic targeting as standalone tooling

These changes will improve the user flow by removing the auto targeting tab, and bring those benefits to all advertisers by integrating them into manual targeting. The Smart Targeting toggles will also be more intuitive, giving you more control over what targeting variables you want to expand upon.

With these changes and new functionality on the horizon, we wanted to take a minute to remind advertisers of our latest and greatest recommendations for getting the most out of Targeting on TikTok.

1. Start with broad targeting

For most advertisers, broad targeting is the most effective solution. Advertisers can think of broad targeting as setting the minimum number of required targeting constraints. As a rule of thumb, when in doubt, do not restrict your targeting settings.

In our testing, ads in TikTok Ads Manager that reach a "fairly broad" audience—defined as including over 80% of the potential users in a given country—outperform all others by a wide margin:¹

• -15% lower cost per acquisition

• +20% higher conversion rate

Broad targeting does not mean broad delivery. It simply means that our ad models have more freedom to optimally determine the right users to see your ads. With the integration of automatic targeting into the backend, broad targeting is more effective than ever.

We recommend that all advertisers start with broad targeting. Even for advertisers using more advanced techniques, it is worth re-testing to ensure that their more advanced techniques outperform broad targeting.

2. Explore manual controls as needed

While we recommend that advertisers start with broad targeting, there are also scenarios where it makes sense for an advertiser to explore manual controls.

For example, if your business only operates in a specific geographic area, you can tailor messages and offers to specific locations.

Another example is using the high spending power dimension to target ads to users who typically spend more on purchases than the average user, which has been proven to drive higher ROAS for advertisers.

Interest and behavior targeting are also useful for brand advertisers looking to reach specific audience subsets.

3. Enable Smart Targeting

If you are targeting specific interests, behaviors or audiences, we recommend that you adopt Smart Targeting. Smart Targeting comprises two toggles: smart audience and smart interests and behavior.

With Smart Targeting, TikTok launches your ad with your exact targeting setting. If the system then observes your ad performance declining, it will expand your targeting settings based on variables that you enable. It will not expand on demographic variables like age or gender, or into excluded audiences.

On average, Smart Targeting improves CPA by 5% for web conversion advertisers.²

4. Create custom and lookalike audiences

As advertisers become more advanced, creating custom audiences can be a powerful tool to improve ad performance. Custom audiences allow an advertiser to build an audience specific to their targeting needs based on first or third-party data. For top-spending advertisers on TikTok, over 50% of ad spend uses audiences in their targeting.³

Once advertisers have established data connections with TikTok, they can leverage custom audience segments to seamlessly retarget users. Audiences can also be used for exclusions—for example, ensuring that ads do not target recent purchasers.

Connecting privacy-safe audience data from one of our many partners, such as Liveramp, is another option. With either data source, advertisers can use audiences as a seed to build a lookalike audience. Similar to broad targeting, we recommend broad lookalikes as the most effective choice for most advertisers. When using audience targeting, remember to enable the smart audience toggle to automatically keep performance stable.

5. Validate that more advanced techniques outperform broad targeting

There is no one-size-fits-all approach to successful marketing. However, broad targeting is the best approach for allowing TikTok's machine learning systems to optimize your ad delivery.

We recommend that advertisers progress through the above steps to get the most out of targeting. However, they should test and validate each optimization, always ensuring it outperforms broad targeting.

Using this test-and-learn approach, advertisers can easily determine the best targeting settings for their business.

Sources

1. TikTok back-end testing, comparing overall web conversion ads performance in Q2 2023

2. Strict Test v Control of 2,000+ Web Conversion Ads in Q2 2023

3. TikTok internal data 2023, based on 100 highest-spending advertisers