The TikTok algorithm creates viral moments using a different methodology than other social networks. If you use the same influencer marketing tactics on TikTok that you use on Facebook or Instagram, you are following a formula for failure.
To succeed on TikTok, you need to know what's unique about TikTok's algorithm, what influences your results, and what doesn't work.
RAD Intel has invested in learning what makes TikTok tick so we can develop high-quality, high-performing creative strategies for our clients and deliver them winning influencer marketing campaign results. Our insights into TikTok's algorithm are informed by over 600 API connections, including our unique partnership with Reddit that lets us see what authentic user discussions there are saying about TikTok content and influencers. We've analyzed over five billion data points to pinpoint what goes into making TikTok viral moments. Here we'll share what we've learned and how it can benefit your TikTok influencer marketing campaigns.
In this post, we'll explain how the TikTok algorithm works and how to put it to use to achieve effective KOL influencer campaigns. First, we'll look at what TikTok's own CEO has said about what makes the TikTok algorithm different from reigning social media models. Then we'll break down what the algorithm factors in when calculating your results, as well as what doesn't count. We'll distill three key ingredients that go into achieving viral results on TikTok. Finally, we'll share the seven-step approach RAD Intel uses to leverage TikTok's algorithm and turn data intelligence into high-performing creative campaigns that deliver measurable ROI for our clients.
TikTok's formula for serving content to users follows a different paradigm than what other social media networks use and what most influencer marketers are used to. TikTok CEO Shou Zi Chew discussed what makes the TikTok algorithm different in a recent TED Interview.
Chew explained that TikTok was inspired by reinventing the way content was displayed. When Chew and his partners first conceived the idea of TikTok about a decade ago, most social media networks displayed feeds and suggested content based on social graphs of who you knew. TikTok's designers decided to try a different approach by displaying content based on what you like.
From this initial inspiration, TikTok developed a vision of user experience centered around three key goals:
1 Providing a window for users to discover enjoyable content
2 Supplying a canvas for users to be creative
3 Building a bridge for users to connect
To achieve these goals, TikTok developed an algorithm which uses machine learning to determine and display what users like by analyzing interest signals such as views, likes, and shares. Chew explained that this represented a different approach to serving content than prior apps. Search engine apps were built to help users find content by entering keywords. Social media apps use social graphs to connect people to networks. In contrast to apps that serve content based on searches or social graphs, TikTok serves content based on a machine learning analysis of what users like.
Chew illustrated how the algorithm works with a simplified example of three different users and six different videos. The first user liked videos 1 through 4. The second user liked videos 1 through 3 and 5. The third user liked videos 1 through 3 and 6. From this, the algorithm detects that people who like videos 1 through 3 have a certain likelihood of also liking videos 4, 5, and 6.
Based on this, the algorithm will show videos 4 through 6 to all three users who liked videos 1 through 3.
TikTok uses machine learning to repeat this formula at scale by factoring in much more data for more granular personalization. The basic principle is that if two people both like certain content, there is a certain probability that each of them will like additional content liked by the other party. TikTok applies this principle at scale by using data from billions of users engaging content in real time.
Chew added that TikTok also differs from other social networks because it relies on interest signals rather than asking twenty questions about your interests and whether you like a piece of content. The content you see is determined organically by interest signals. What you watch, what you like, and what you share provides TikTok with the interest signal data it uses to analyze what you like. Based on this input and certain other variables, the algorithm uses AI-based pattern recognition to match you to content you're likely to like.
Chew's interview emphasized the importance of interest signals in influencing how TikTok content gets displayed. A more complete perspective on TikTok's algorithm factors in additional variables - let's unpack.
1 Interest signals and non-interest signals
2 Video and audio content
3 Device and account settings
1 Views
2 Completions
3 Re-watches
4 Likes
5 Shares
6 Follows of accounts
1 Videos you've skipped
2 Videos you've marked that you're not interested in
3 Filtered keywords
4 Users you've blocked
TikTok's recommendation engine uses these types of non-interest signals to exclude content from your feed. Additionally, TikTok avoids displaying duplicate content on For You Page recommendations, and it restricts recommendations of videos on certain subjects deemed problematic.
TikTok's algorithm analyzes characteristics of the content you engage. Important elements include:
1 Captions
2 Text overlays
3 Hashtags
4 Sounds
5 Effects
6 Transcripts
Trending audio plays a special role in TikTok's algorithm. The TikTok recommendation engine notes when content is attached to a trending song or sound.
The TikTok algorithm uses key user information to help it recommend content which include type of device, location settings and language preferences.
These above factors receive lower weight than interest signals and content characteristics, but they help TikTok determine the relevance of users with different devices and users from different areas and language groups.
What the TikTok recommendation engine doesn't factor in (also is important to note). As Chew emphasized, TikTok doesn't show content based on network graphs. It's not who you know, it's what you like. But TikTok is working on making your network an important factor soon.
TikTok doesn't make recommendations based on how many followers a user has. Your content can get recommended even if you only have a few followers.
TikTok doesn't recommend content based on the performance of your previous videos. Even if few people watched your last video, your next video could go viral.
What does all this imply about what it takes to achieve viral momentum on TikTok?
The platform's emphasis on what audiences like gives us a starting strategy for TikTok viral marketing that can be used across several disciplines. Let's unpack the these strategies based on knowing what your audience likes, delivering engaging content your audience will like and connecting with the KOLs your audience likes.
These steps are interconnected. Knowing what your audience likes tells you what type of content you should be posting to engage them. Knowing which KOLs your audience likes tells you which influencers you should partner with to get your content seen and by your intended viewers. Bringing all three of these ingredients together gives you a recipe for reaching the right audience with the right content through the right influencers.
But how do you get the data you need to identify your target audience's likes, your optimal content strategy, and your optimal influencers? That's where RAD Intel's data intelligence comes in.
RAD Intel puts TikTok's algorithm to use for our clients by providing the data and methodology needed to make informed decisions about audience, content, and influencer selection. We draw data from over 600 API connections, as well as our client's historical data. Armed with this data, we help our clients match measurable campaign goals to the right TikTok audiences, content, and KOLs to achieve them. Our method can be summarized as a seven-step process:
1 Defining data-driven goals and benchmarks
2 Developing AI-informed audience personas
3 Matching markets to creative strategies
4 Identifying optimal influencers
5 Onboarding influencers
6 Activating campaigns
7 Optimizing campaign performance
Here are some straightforward steps that marketing teams can take to initiate action. Starting with a data-driven strategy can be a reasonable first move, as we have observed organizations achieve positive results when implementing this approach to support their TikTok marketing efforts. Conversely, we've also witnessed mediocre outcomes when recommendations are not integrated into the company's marketing culture due to a lack of intentional processes.
Guided by client success managers, our clients begin their journey to TikTok success with a discovery process that helps them identify measurable campaign performance goals. For example, a client might seek to achieve a certain number of views, likes, or videos using the campaign hashtag.
With well-defined goals in mind, our onboarding process develops KPI-defined benchmarks, project plans, and timetables to guide our client's campaign. To select suitable benchmarks, we connect our AI-powered influencer marketing platform to your brand-owned and social media channels to draw from historical data. This steers you toward data-defined, measurable, achievable performance goals.
To help our clients reach TikTok audiences who match their campaign goals and target markets, we use our API and AI capability to help develop precisely defined audience personas. We draw from over 26 million daily active Reddit user conversations and other select API sources to discover what interests engage your target audience. We identify both dominant interests as well as secondary and overlapping interests that refine your audience persona targeting. After suggesting personas based on our research, we obtain your approval before proceeding to the next step.
With your target audience and their interests defined, we help you identify the ideal TikTok creative strategy for your campaign. Based on who your persona-defined audience is and what their interests are, we help you identify what type of content to engage them and where and when to deliver it.
After identifying your ideal creative strategy, we proceed to match you with the best key opinion leaders and influencer partners to implement your strategy. Using AI, we match your campaign objective, audience persona, and content strategy to the top 1% of influencers and KOLs meeting your criteria. With your approval, we help you select prospective influence partners.
Once we've identified influencers that match your campaign strategy, we help you connect and collaborate with them. With your approval, we secure and contract your influence partners. Through our platform, your partners can collaborate with you. We help you communicate project requirements to content creators and establish due dates. Your partners can upload content for your review and approval. We help you review content to make sure it means your brand guidelines as well as FTC regulations. Our platform makes it easy for your to review content at scale and publish approved projects through integration with TikTok as well as Facebook, Instagram, Snap, Twitter, and all paid ad managers.
Once you've approved your influencers' content, we activate your campaign. Your content goes live on your campaign's designated brand-owned and social media channels. Through our platform, you can monitor your campaign results for complete transparency on your performance and ROI.
After your campaign goes live, we help you optimize your performance to ensure you achieve your target objectives and realize a return on investment. We use a suite of AI-based insights to track KPIs such as hashtag insights, content impact by channel, sustained follower growth rate, sustained engagement rate, and correlation of your web traffic with influencer content. Based on campaign performance, we recommend mid-flight optimizations to bring your results in alignment with your goals. Our platform lets you make batch optimizations for efficient management of your campaign.
Our method delivers proven, guaranteed results. In one case study, an entertainment industry client saw a 197% lift in engagement rate, with 397 pieces of content created generating 6.8 million impressions and 754,000 engagements powered by our AI. We've delivered these types of results across multiple industries, from healthcare and ecommerce to hospitality and visual art. Our track record enables us to guarantee our results. We deliver our clients the same predictable ROI you would expect from other paid marketing channels.
When you know what makes TikTok tick, you can put TikTok's algorithm to work building your brand. TikTok's emphasis on interest signals allows you to achieve virality by delivering content matching your audience interests. RAD AI data intelligence on your audience and their interests provides the information you need to align your content perfectly with your intended audience and find the right partners to make sure your content gets seen.
View a demo or contact our team to discuss how we can help you achieve TikTok influencer marketing results.