The core of traffic is to optimize key metrics such as playback volume, like volume, comment volume, and completion rate.

Author: SunnyZ

In my previous article, I mentioned that I would write about the operation and traffic strategies for TK because there have been too many projects coming to cooperate recently. To put it bluntly, the traditional old methods of Web3 are no longer sufficient (i.e., task platforms, PR collaborations, AMAs, or KOLs), and the volume exchange on Ton has already reached a bottleneck. At this time, TikTok (hereinafter referred to as TK), as the largest traffic pool, has become a battleground. Using TK KOLs effectively can sometimes yield better conversion results than Twitter. TK is still a massive traffic gap, and I personally believe this year marks the first year of the integration of Web3 and TK. If we do not seize this wave of growth in TK, we are likely to miss out on this round of new traffic. Let's look at how to operate accounts and what strategies to use to acquire traffic.

Two characteristics of traffic

Today's TK is somewhat like Kuaishou in 2019, regarding the precision of traffic. Currently, TK's video data is far from sufficient, so the hashtags for influencers are not yet fully developed. TK's streaming method is different from Douyin; it does not prioritize pushing to private pools but directly promotes to the same public pool. Therefore, for cross-border influencer searches, 'vertical influencers' are currently a misnomer. In the process of traffic conversion, we first need to understand two characteristics of TikTok traffic: precision and immediacy.

1. Precision

We know that in the conversion funnel, traffic must come before conversion. Therefore, before thinking about how to convert, two questions must be addressed: one is how to acquire traffic, and the other is how to acquire precise traffic. From the perspective of precision, traffic can be divided into precise traffic and non-precise traffic. Comparatively, acquiring traffic is not the most difficult part; obtaining precise traffic is the greatest challenge faced by the project, and only precise traffic can allow us to complete the loop. For most traffic behaviors, the value of precise traffic is thousands of times higher than that of non-precise traffic. When our project team conducts TK advertising, we must find the precise KOL in our niche; even if they have fewer followers, if the conversion effect is good, we can achieve high results with low expenditure.

2. Immediacy

From the perspective of traffic immediacy, when a video is pushed to a large number of TK users, these users will only see the video at that moment when it is pushed, and this moment is considered 'traffic passing through our video'. The likelihood of these users revisiting the video in the future is negligible, so the traffic pushed to our video only has value 'while it is passing through'. Regardless of how many followers a TK account has or how many views a short video gets, once that traffic has passed, they are merely a number.

In response to these two characteristics of TK traffic, if we want to achieve conversion in TK's vast traffic pool, we must solve it from two aspects. The first aspect is to learn how to filter traffic. Since the precision of traffic is very important, it is necessary to first learn how to filter traffic to find the precise traffic pool. This is an essential skill that every TK operator needs to master. During the TK traffic distribution process, the system will distribute content to corresponding TK users based on the different content of videos or live streams. Therefore, the core way to filter traffic is to optimize content. There are many tools and methods available, the simplest include language tools, cultural tools, music tools, etc. We can use these tools to control the direction of video traffic. For example, if the project party wants to target Thailand, then users from other regions are not precise for us. In this case, we can insert Thai text into the video content, which greatly increases the probability of filtering out traffic from other regions. The second aspect is to prepare the conversion chain in advance. Because traffic has immediacy, we need to prepare the conversion chain and the 'pool' to absorb the traffic before it arrives. So what methods can be used to absorb traffic? If we hope to complete the entire token trading loop within TK, we can ask influencers to post project party's token trading links in the pinned information or redirect to CEX or DEX, then view conversion rates through a single link. The TK platform always pursues 'future' traffic; traffic that has 'passed' is almost of no value. Regardless of how good the past data of an influencer is, that is also sunk cost. Even if an influencer shows a video with 10 million views, that traffic will not bring new project parties any traffic in the future; that number only represents traffic that has once come. Therefore, from the traffic perspective, what we need is always the next wave of traffic.

So where does traffic come from:

There are many sources of short video traffic, such as recommended traffic, follower traffic, homepage access, search access, music access, tag access, etc. Among these, the most valuable traffic in TikTok short videos is undoubtedly the system-recommended traffic, commonly referred to as 'For You' traffic. The more traffic this channel has, the better the platform assesses the video's quality, leading to further recommendations, resulting in greater playback volume. When evaluating whether the traffic source of a short video is healthy, we generally require the 'For You' traffic ratio to be greater than 30%. If it is below 30%, we need to consider whether the video or account is being restricted.

The logic of the TK short video recommendation algorithm

Before discussing the operation methods of TK accounts, we first need to understand the logic of the TK short video recommendation algorithm. As we all know, TK pushes content that you are interested in based on your behavior and preferences.

Overall, the core of traffic is to optimize key metrics such as playback volume, like volume, comment volume, and completion rate. What is the essence of optimizing these metrics? It's optimizing content. As long as the content quality is acceptable and the audience's preferences are accurately grasped, the video will naturally gain more traffic.

The TK platform's video traffic recommendation is conducted in stages. In the primary traffic pool, videos will receive 200-500 views. Specifically, after publishing a video, the platform will first push it to 200-500 viewers. The platform will then evaluate the video's popularity based on these viewers' interactions (likes, comments, shares, etc.).

If the feedback is good, the video will enter the secondary traffic pool with views reaching around 2000 times. If the video continues to perform well in the secondary pool, it will advance to the tertiary traffic pool, with views possibly reaching 5000 times or even tens of thousands. This process will continue until the video performs poorly in a certain traffic pool, at which point recommendations will stop.

It is worth noting that the primary traffic pool usually targets users in the account's location. For example, if the account's IP address and data belong to the United States, the primary traffic pool will primarily push content to American viewers. However, starting from the secondary traffic pool, geographical restrictions gradually disappear, and videos have the opportunity to gain global traffic.

So, what is the key to acquiring traffic? The answer is content. Even if the account's location and primary traffic pool are in the United States, as long as the video content aligns with the preferences of Southeast Asian audiences, the system will recommend it to users globally who are interested in Southeast Asian culture, thereby making it easier to acquire traffic from Southeast Asia.

Of course, when managing traffic, it is important to remember that traffic is essentially random. This is a common characteristic of all traffic. Our task is to continuously reduce this randomness and transform the unpredictable random game into a controllable probability game.

Data feedback is key to optimization and iteration. By deeply analyzing data, we can better understand the essence of things. In managing TK accounts, data analysis is also indispensable; it helps us gain deeper insights into the platform and clarify optimization directions.

TK KOL Cooperation

How to select KOLs, many Crypto KOLs have many followers, but their interaction data is average. Depending on the project strategy, different configurations can be made. For example, if the project has a large promotional campaign, you can configure 1-2 large KOLs and multiple smaller KOLs; if the hashtags are chosen well, it will be very easy to break out.

Of course, there is another strategy, which is to target and burst specific individuals' TK follower lists. If you know the main influencer's follower list, just select the KOLs directly.

Wherever traffic is, value is also present. Since the environment for acquiring traffic is variable and not singular, being able to stably acquire traffic amidst these changes is both an important and challenging capability.

The first step in cooperating with influencers is to find the right influencers. We need to conduct a preliminary screening of TK influencers based on characteristics such as target categories, audience demographics, and marketing objectives. The following introduces four main methods:

1. Finding KOLs through the TK App is one of the most basic and effective methods. Search for keywords related to the product on TikTok, find relevant videos, filter out well-performing content (such as significantly higher views than similar videos), and then contact the creator through their profile. This method can help find many influencers who already have a certain follower base in relevant fields.

KOL profiles usually provide contact information: some will leave a business negotiation email directly; those without direct contact information often have a landing page link on their homepage, which can be found in 'About Us' or 'Contact Us'. You can also leave a message in the video comment area and wait for the KOL to contact you actively.

2. Actively connect through the TK Influencer Plaza. This method is more direct and precise than searching directly on TK, but currently, the number of influencers settled in the Influencer Plaza is relatively small.

3. Focus on influencers collaborating with competitors.

4. Find Crypto KOL MCNs; they are often more professional and more familiar with KOLs.

Many project teams actually haven't thought things through; they haven't predetermined promotional goals or their goals are not quantifiable. Before cooperating with influencers, it is crucial to clearly set the promotional goals for this collaboration, which means understanding what effect we hope to achieve through this cooperation—whether it is increasing brand exposure, gaining followers, boosting token trading volume, or finding quality clients. Once the promotional goals are determined, it is essential to set several quantifiable metrics to evaluate whether this promotion meets expectations. If quantifiable metrics are not set, it can lead to not knowing how to optimize strategies, or results not matching expectations, such as wanting to increase sales but ending up increasing exposure, which is counterproductive.

TK Data Analysis

In TK data analysis, we mainly focus on two types of metrics: video metrics and account metrics.

Video metrics include: completion rate, like rate, comment rate, share rate, traffic source, as well as playback volume, total playback duration, average playback duration, audience distribution, and proportion.

Account metrics include: like-to-follower ratio, follower distribution, following count, follower count, like count, number of videos published, time of first video release, recent publishing frequency, and gender distribution, etc.

By analyzing these metrics, we can comprehensively assess an account's performance, applicable to both our account and other high-quality accounts.

Video Metrics

We first analyze metrics from the video dimension, examining what issues each important metric reflects, the measurement criteria for those metrics, and how to optimize videos using these metrics. However, it is important to note that different purposes, types, and categories of videos will have different data forms. For example, some videos may have high like counts, while others may have higher comment interaction rates, and some may focus on sharing and dissemination. The following recommended metric measurement standards are generally applicable across most categories, and flexibility is encouraged in reference.

Completion Rate

The completion rate refers to the proportion of viewers who complete 100% of the video viewing progress out of all viewers. The completion rate is one of the most important metrics among various video metrics and is a key factor affecting video playback volume. For project teams, the amount of time users are willing to spend on your product is a crucial metric. The completion rate on the TK platform is strongly associated with how much time the audience spends. If the video has a high completion rate, it indicates that viewers are interested in the content and are willing to spend time watching it. The platform tends to regard such videos as relatively high-quality and will further recommend them. Based on practical data summaries, the e-commerce industry generally sets the minimum benchmark for determining whether the completion rate meets standards at 30%. If the completion rate is below 30%, it is considered unacceptable, and the video must be optimized [Source: sky]. At this point, thinking from the audience's perspective, under what circumstances would viewers be unwilling to finish watching a video and swipe away? Generally, there are several reasons.

(1) The first three seconds of the video are not engaging enough. The audience's drop-off rate is often highest in the first three seconds, so we can examine whether the content in the first three seconds is eye-catching and whether there is significant room for optimization.

(2) Revealing all content in the first three seconds. Many creators understand that the first three seconds of a video must be eye-catching, so they dump the main content and highlights all at once in the first three seconds, resulting in the audience having low expectations for the following content. When viewers have no expectations, they will naturally choose to swipe away.

(3) Slow rhythm. On the TK platform, the audience's patience for each video is limited. If the video lacks a strong rhythm and fails to occupy the audience's senses sufficiently, it will lead to traffic loss. Especially for medium to long videos, it is essential that the video has a strong rhythm and high content density, making the audience feel that every second spent watching the video is worthwhile. Therefore, if the completion rate is below 30%, the above three points should mainly be referenced for optimization.

Based on these principles, the project team can observe whether the KOLs' follower counts, post data, and like counts match, which can help find influencers who can genuinely drive sales.

Like Rate

Like Rate Calculation Formula: Like Rate = Total Likes / Total Views.

Compared to completion rate, the impact of like rate on playback volume is smaller, but it is still a metric worth optimizing. When the like rate is below 4%, the video has optimization potential. From the audience's perspective, they will like a video under two circumstances: (1) The video has collectible value. Many people find good videos while browsing TK, and if they don't like or save it in time, it becomes difficult to find that video again later. Therefore, enhancing the practical value of the video can motivate viewers to like, save, and rewatch. (2) The video is highly entertaining. Interesting content can spark viewer interest, which is a key factor affecting the like rate. If the like rate is low, we can focus on enhancing the video's entertainment value. Here we can look for KOLs who create very engaging videos.

Comment Rate

Comment Rate Calculation Formula: Comment Rate = Total Comments / Total Views.

When the comment rate is below 0.4%, the video has optimization potential. Here are three methods to improve the comment rate:

Share Rate

Share Rate Calculation Formula: Share Rate = Total Shares / Total Views.

Just like Twitter, the likes, comments, and impressions need to match to be considered a good post.

Account Metrics

Like-to-Follower Ratio

At the account dimension, the first metric worth focusing on is the like-to-follower ratio, calculation formula: like-to-follower ratio = total follower count / total like count.

The like-to-follower ratio can very intuitively reflect the stickiness of an account's followers. If the like-to-follower ratio is too low, it indicates that the account's follower stickiness is low, and the content that attracts the audience only remains at the video level without rising to the influencer's persona or account level. The standard for judging the like-to-follower ratio is generally that if it's greater than 1:6, it is considered that the account has high follower stickiness and that the followers are relatively precise; if it is lower than 1:6, it is considered that the video has certain optimization space. For example, some accounts' like-to-follower ratios reach 1:15 or 1:20, which is a very typical sign of low follower stickiness.

Follower Distribution

The second metric is follower distribution. We often simply divide it into gender distribution, regional distribution, and age distribution. When choosing crypto influencers, the proportion of influencers should be configured according to the type of your project.

Acknowledgments

Many thanks to the friends who provided various information support for this article! Since I personally have not participated much in TK operations, much of the content in this article references Sky's book and the experiences of friends who actively operate TK. I also recommend that those who want to learn TK operations read and learn more. The KOL examples in this article are sourced from Wayne; if you need the TK influencer resources mentioned in the examples, feel free to contact me, and I welcome everyone to communicate with each other. My TG: SunnyZ_Crypto (please forgive the slow replies).