Ever heard of the phrase, "never put all your eggs into one basket"?
If you are in performance and results based marketing, you should be actively exploring and investing in multiple media channels other than the commonly used ones like Facebook and Google. One of the main problems we face in advertising is finding new audiences to target. By showing your ads in various channels, you will be able to increase your chances of acquiring new customers.
In this article, we will be sharing our thoughts on TikTok advertising and how advertisers can leverage on data-driven insights to get more leads, website traffic, sales conversions, app installations, and more from the fast-growing video social platform. Mainly, we will be speaking to these two personas, and I am sure you fit into one of them:
For a start, the TikTok Ad Manager has a very similar user interface (UI) to Facebook Ad Manager. You could get started with it right away, and the dashboard is not so hard to navigate around.
(Source: TikTok Ads Manager)
To put some light into TikTok's rapid growth, here are some reasons why the platform is developing so quickly:
If you have been advertising for a long time, I’m guessing that you often leverage on audience insights and data to research your best-fit prospects online. Facebook and Google offer tools to help you find such audiences to target in your campaigns, but there is still very little information about TikTok audiences that you can research on. The advertising capability of TikTok is still developing, so you might wonder how different or impactful the audiences there are.
Update (July 2021): Facebook no longer offers their own Audience Insights dashboard. However, many advertisers have started adopting 3rd party software and tools to help them research for their best audiences for that channel.
So to get the ball rolling, here are 3 data-driven insights to help you find your best-suited targeting for your next TikTok ad campaigns.
Psychographic-based targeting helps you understand your audiences by finding out which personality traits and attitudes your consumers base their purchasing decisions on. This approach can be used to develop more segmented personas than it is possible on the basis of superficial demographic classifications.
For example, I’m known to be a Marketing and AI Expert. However, a small private circle also knows me as a video gamer. From my online activity, social media platforms will know that I’m interested in both Marketing and Video Games, but only a small portion of advertisers would bid for “Video Games” to reach me with their ads. Most advertisers will rely on their common sense to bid for “Marketing" and "Technology”, making their ads much more expensive than if they choose to bid for “Video Games”.
If it is much cheaper to reach me via the “Video Games” interest, why would you pay for reaching me through the “Marketing" and "Technology” interests, which are so much more expensive? Imagine, you could reach the same customer but through his/her other, and not very common, interest that is still very relevant to the person. In SoMin, we call this long-tail targeting.
Recently we wanted to know what are the long-tail targeting options for a fitness industry client. They wanted to run a video ad campaign to promote outdoor workout classes via TikTok. The ad aims to drive traffic to their sign up page at as little cost as possible (i.e. less than SGD $0.80 per click).
Instead of targeting the "Fitness" interest or similar, we used our platform to look for long-tail audiences. Using Computer Vision (CV) and Natural Language Processing (NLP) technologies to analyze social media textual and imagery content, cluster millions of content engagement data points into new interest categories, and build personas that most advertisers will not target in their ad campaigns.
Look at the top interests and psychological age analysis of the 'Ideal Audience' below. Not many of you will ask someone who is interested in "Business and Industry", "Fashion Accessories", "Travel", or "Beauty" to respond to an outdoor workout class ad. But, these are the interest categories that our machine learning engine classified into new personas.
(Source: SoMin AI Audience Insights and Analytics Dashboard, Top Interests and Psychological Age)
These interests were then included in the targeting for our TikTok ad campaign. As TikTok currently has limited interests to target, we find the closest corresponding matches to the list above. To make our campaign more targeted, we only choose to include the top four interests from our insights. Look at how we match them below.
(Source: TikTok Ads Manager, Ad Group)
By doing so, we are able to avoid “bidding competition” with other brands and advertisers who use fitness-related interests.
Additionally, we are able to see the top locations where our audience lives. So we added that in the location setting to further narrow our targeting.
(Source: SoMin AI Audience Insights and Analytics Dashboard, Audience Locations)
Knowing the psychological and behavioral insights of your audiences not only helps you with targeting. It also helps you to create stronger content and creatives that will be more relevant to your target customers. As creatives are part of the ad campaign, it is equally important to get this right on TikTok.
Take a look at the social snapshots of our own audiences below. You will notice that our social media and ad content engages more with people that possess the following traits:
(Source: SoMin AI Audience Insights and Analytics Dashboard, Social Snapshot)
With these insightful stats, it’ll tell you that you’ll have a higher chance of increasing your engagement rates when you publish creatives that speak to people with these ‘traits’.
Example: You can start designing and building social media or ad creatives that target the “mature, feminine people who are single at heart”
Start by asking yourself:
Now, if that’s not all, we also researched the personality traits of our audiences. Using the Myers-Briggs Type Indicator (MBTI), we are able to deduce that our audiences are mainly INFP.
(Source: SoMin AI Audience Insights and Analytics Dashboard, MBTI)
Again, we can design creative, write content, and produce videos that resonate with this persona. Ask yourself:
So, adding on to the previous example: Your content must speak to the “INFP people who are single at heart that show mature and feminine traits”
If you are able to do extensive research to gather psychographic insights like this, and apply them to your creatives and content, then you will see massive improvements in your engagement or conversion rates on TikTok.
Like Facebook and Google, the TikTok Ad Manager is kind to provide you with their own insights and data to help you optimize your campaigns.
Rule of thumb: Run an initial test campaign for at least a week to gather some data.
While there are several metrics to look at, here are five recommended ones (and how you can read and apply them).
You will be able to look at the Dayparting stats from the TikTok ad manager dashboard. Analyze the Click-through-rates (CTR) and Cost-per-click (CPC) distribution across Day and Hour. The screenshot below shows you the CTR distribution from our recent campaign.
The darker shade of blue cells corresponds to higher CTR during that hour. You can quickly tell that high click-throughs of the ads occur during the night (particularly between 8pm and 2am), and throughout the whole day on a Sunday.
(Source: TikTok Ads Manager, Dashboard: Dayparting)
Once you have the dayparting data, you can consider scheduling your ads during the hours with higher activity to achieve better results.
While you might not get the Age distribution insights directly from the dashboard, you could fetch them by creating and exporting a report. Go to ‘Reporting’ and select the Age dimension with key metrics like CPC, Clicks, and CTR. Then click on the ‘Run & Export’ button to download the report.
We tabulated our results in Google Sheets, and it looks like the screenshot below. The age ranges of 18-24 and 35-44 perform above average in terms of CTR. It helps us understand better which age groups resonate with our ads more. In this case, since the results are pretty close, we continue to show ads to our initial age setting of 18-54.
Of course, if you have outliers in the table that performs far worse than the rest, you could consider changing your age settings for future campaigns (excluding the outliers) so your ads won’t be shown to them.
(Source: SoMin Analytics)
The TikTok Ad Manager dashboard is able to show you a clear gender analysis of your campaign. We suggest looking at the number of results (clicks, conversions, app installs, etc.) to determine which gender makes a larger part of your ideal audience.
From our own ad campaign, we can immediately tell that Females make up the majority of people clicking on our ads. Since our campaign objective is Website Traffic, this is a sufficient data point to refer to. You can also look into your traffic data from Google Analytics to see if the stats tally. This way, we could optimize our campaign by creating new ad groups that only target females to see if the results improve. More often than not, it will improve.
(Source: SoMin Analytics; TikTok Ad Manager Dashboard)
Now here is the interesting part. Before the Apple iOS 14 privacy updates, most advertisers would not actively optimize their campaigns based on device OS. However, after May 2021, results from iOS devices started to slow down because many Apple users opted out of AppTrackingTransparency (ATT) policy.
In short, the ATT policy allows users who opted in to see personalized ads from media channels like Facebook and TikTok. Those who opted out chose not to see it, making them more unreachable to advertisers. Many people opted out mainly due to concerns over how their data would be processed by advertisers.
In the same dashboard on your TikTok Ad Manager, you can see the results distribution between iOS and Android devices. We can clearly see that users with Android devices make up a large chunk of our best-performing audiences. Similarly, you can consider testing another ad group targeting only Android users in order to improve your results even more.
(Source: SoMin Analytics; TikTok Ad Manager Dashboard)
Analyzing the video play stats of your ad campaign is also critical in understanding how well your video creative performs. You will be able to create custom columns in your campaign table with Video Play metrics. Here are the key metrics you could look at.
The screenshot below shows the stats from a test campaign we ran over five days using a 15-second video creative.
(Source: TikTok Ad Manager Dashboard, Campaign)
If the numbers in the image are too small and blurry, here is the summary.
You can see that 60% of viewers dropped out even before the 2-second mark, which is pretty bad. Only 2% of viewers watched the video till the end.
This could mean two key reasons why the stats are poor (though there might be other factors too):
Corresponding to the issues above, there are two quick fixes that we can implement to improve the stats:
To see impactful growth in your TikTok ad campaigns, the best way to optimize them is by using the data-driven methods. The key to your advertising success can be found in the numbers. That is why at SoMin, we employ Data Scientists and Mathematicians to help us improve the way we optimize ad campaigns and build audience insights to target more relevant people online.