Article author: Cube Labs
Source: Cube Labs
In my last article, I mentioned that I would write about TK's operation and traffic investment because there are too many projects that have come to cooperate with TK recently. To put it bluntly, the traditional three things of Web3 are no longer enough [that is, task platform, cooperative PR, AMA or KOL], and the exchange of ton has entered a bottleneck period. At this time, Tik ToK [hereinafter referred to as TK], as the largest traffic pool, has become a battleground. How to make good use of TK's KOL, the conversion effect is sometimes much better than Twitter. TK is still a huge traffic depression. I personally think that this year is the first year of the combination of Web3 and TK. If you don't grasp this wave of TK growth, you will most likely miss this round of new traffic. Let's take a look at how to operate the account and what kind of tactics to use to obtain traffic.
Two characteristics of traffic
The current TK is equivalent to Kuaishou in 2019. Regarding the accuracy of traffic, TK’s current video data is far from enough, so the hashtags for influencers are not yet perfect. The current TK’s push method is different from Douyin. It does not prioritize pushing to the private domain pool, but directly pushes to the same public domain pool for promotion. So for our current cross-border search for influencers, the same is true. "Vertical influencers" are currently a false proposition in TK. In the process of traffic conversion, we must first understand the two characteristics of TikTok traffic: accuracy and immediacy.
1. Accuracy We know that in the conversion funnel, there must be traffic before there can be conversion. Therefore, before thinking about how to convert, two problems must be solved. One is how to obtain traffic, and the other is how to obtain accurate traffic. Traffic can be divided into accurate traffic and non-accurate traffic from the perspective of accuracy**. In comparison, obtaining traffic is not the most difficult, but obtaining accurate traffic is the biggest challenge facing the project party, and only accurate traffic can allow us to run a closed loop. ** For most investment behaviors, accurate traffic is thousands of times more valuable than non-accurate traffic. When our project party is doing TK investment, we must find the accurate KOL of our own track. Even if there are not so many fans, but if the conversion effect is good, we can "spend little money to pretend to be big."
2. Immediacy From the perspective of traffic immediacy, after the video is pushed to a large number of TK users, these users will only see the video at the moment of being pushed, and this moment is regarded as "the traffic is passing through our video". The probability of these users watching the video again in the future is negligible, so the traffic pushed by the system for our video is only valuable "when it passes". No matter how many fans a TK account has or how many views a short video has, once the traffic has passed, it is just a number.
In view of these two characteristics of TK traffic, if you want to complete the conversion in TK's huge traffic pool, you also need to solve it from two aspects. The first aspect is to learn to filter traffic. Since the accuracy of traffic is very important, you first need to learn how to filter traffic in order to find an accurate traffic pool. This is also a necessary skill that every TK operator needs to master. In the process of TK traffic distribution, the system will distribute it to the corresponding TK users according to different videos or live broadcasts. Therefore, the core way to filter traffic is to optimize content. There are also many tools and methods. The simplest ones are 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 launch in Thailand, then users in other regions are not accurate for us. In that case, Thai copywriting can be inserted into the video content, so that there is a high probability of filtering out traffic from other regions. The second aspect is to prepare the conversion link in advance. Since traffic is instantaneous, we must prepare the conversion link and the "pool" to undertake it before the traffic arrives. So what are the ways to undertake traffic? If you want to complete the closed loop of the entire token chain within TK, you can let the influencer post the token trading link of the project party in the top information, or jump to cex or dex, and then check the conversion rate through a single link. The TK platform always pursues "future" traffic, and the "past" traffic is almost worthless. No matter how good the influencer's past data is, those are silent costs. Even if a certain video of the influencer is shown to have 10 million views, this traffic will not bring traffic to the new project party in the future. This number only means that the traffic has come. Therefore, from the traffic side, what we need is always the next wave of traffic.
So where does the traffic come from?
There are many sources of traffic for short videos, such as referral traffic, fan traffic, home page entry, search entry, music entry, tag entry, etc. In TikTok short videos, the most valuable traffic is undoubtedly the traffic recommended by the system, which is what we often call "For You" traffic. The more traffic from this channel, the better the platform determines the video quality, and then the recommendation is added, the greater the imagination in the number of views. When we measure whether the traffic source of a short video is healthy, we generally require that the proportion of "For You" traffic is greater than 30%. If it is less than 30%, we need to consider whether the video or account is restricted. [Data source: Sky]
The logic of TK short video recommendation algorithm
Before we talk about how TK accounts operate, we first need to understand the logic of TK's short video recommendation algorithm. Everyone knows that TK will push content that you are interested in to you based on your behavior and preferences.
In general, the core of traffic investment is to optimize key indicators such as playback volume, likes, comments, and completion rate. What is the essence of optimizing these indicators? It is to optimize the content. As long as the content quality is good and the audience's preferences are accurately grasped, the video will naturally get more traffic.
The video traffic recommendation of the TK platform is carried out step by step. In the first-level traffic pool, the video will receive 200 to 500 views. Specifically, after the video is released, the platform will first push the video to 200 to 500 viewers. The platform then evaluates the popularity of the video based on the interactive behavior of these viewers (likes, comments, forwarding, etc.).
If the feedback is good, the video will enter the second-level traffic pool, and the number of views can reach about 2,000. If the video continues to perform well in the second-level pool, it will be promoted to the third-level traffic pool, and the number of views may reach 5,000 or even tens of thousands. This process will continue until the video performs poorly in a certain traffic pool, and the recommendation will stop.
It is worth noting that the first-level traffic pool is usually for users in the location of the account. For example, if the account's IP address and data location (i.e. the country/region to which the TikTok account belongs) are in the United States, the first-level traffic pool will mainly push to American audiences. However, starting from the second-level traffic pool, the geographical restrictions will gradually disappear, and the video will have the opportunity to obtain global traffic.
So, what is the key to gaining traffic? The answer is content. Even if the account is located in the United States and the primary traffic pool is in the United States, as long as the video content caters to the preferences of Southeast Asian audiences, the system will recommend it to users around the world who are interested in Southeast Asian culture, making it easier to gain traffic in Southeast Asia.
Of course, when operating traffic, we must keep in mind that traffic is essentially random. This is a common feature of all traffic. Our task is to continuously reduce this randomness and transform the uncertain random game into a controllable probability game.
Data feedback is the key to optimization and iteration. By deeply analyzing the data, we can better understand the essence of things. When operating a TK account, data analysis is also indispensable. It can help us understand the platform more deeply and clarify the direction of optimization.
TK Talent Cooperation
How to select KOLs? Many crpto KOLs have a lot of fans, but their reposts, comments and likes are average. Different configurations can be made according to different project strategies. For example, if the project has a relatively large event promotion, 1-2 big KOLs can be configured, and then multiple small KOLs can be configured. If the hashtag is well selected, it will be very easy to get out of the circle.
Of course, there is another strategy, which is to targetedly blast someone’s TK follow list. If you know the follow list of Yijie, you can just choose those KOLs.
**Where the traffic is, there is the value. **Since the traffic acquisition environment is changeable rather than single, being able to stably acquire traffic in this change is an important and challenging ability.
The first step in influencer collaboration is to find influencers. We need to conduct a preliminary screening of TK influencers based on characteristics such as target categories, audience groups, and marketing purposes. The following are four main methods:
1. Find through TK App. Finding influencers directly through 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 the content with better performance (such as significantly higher playback volume than similar videos), and then go to the creator's homepage to contact them. This method can find a large number of influencers who already have a certain number of fans in related fields.
The influencer's homepage usually provides contact information: some will directly leave an email address for business negotiations; if there is no direct contact information, a landing page link will often be posted on the homepage, and the contact information can be found in "About Us" or "Contact Us". You can also leave a message in the video comment area and wait for the influencer to contact you.
2. Actively establish connections through TK Talent Plaza. This method is more direct and accurate than searching directly on TK, but there are relatively few celebrities in the Talent Plaza.
3. Pay attention to influencers who collaborate with competitors
4. Find Crypto KOL MCN, they are often more professional and more familiar with KOLs
Many project parties have not thought clearly before cooperating, and have not determined the promotion goals in advance or the goals are not quantifiable. Before cooperating with influencers, you must first determine the promotion goals of this cooperation, that is, what kind of effect you hope to achieve through this cooperation, whether it is to increase brand exposure, increase the number of fans, increase token trading volume, or find high-quality customers, etc. After determining the promotion goal, you must set several quantifiable indicators for this goal to measure whether this promotion has achieved expectations. If you do not set quantifiable indicators, you will eventually not know how to optimize the strategy, or the results will not meet expectations. For example, you originally wanted to increase sales, but the result was to increase exposure, which would be counterproductive.
TK Data Analysis
In TK data analysis, we mainly focus on two types of indicators: video indicators and account indicators.
Video indicators include: completion rate, like rate, comment rate, forwarding rate, traffic source, as well as playback volume, total playback time, average playback time, audience distribution and proportion.
Account indicators include: fan-to-like ratio, fan distribution, number of followers, number of fans, number of likes, number of videos posted, time of first post, recent posting frequency and gender distribution.
By analyzing these indicators, we can comprehensively evaluate the performance of an account, whether it is our own account or other high-quality accounts.
In TK data analysis, we mainly focus on two types of indicators: video indicators and account indicators.
Video indicators include: completion rate, like rate, comment rate, forwarding rate, traffic source, as well as playback volume, total playback time, average playback time, audience distribution and proportion.
Account indicators include: fan-to-like ratio, fan distribution, number of followers, number of fans, number of likes, number of videos posted, time of first post, recent posting frequency and gender distribution.
By analyzing these indicators, we can comprehensively evaluate the performance of an account, whether it is our own account or other high-quality accounts.
Video Metrics
We first analyze the indicators of the video dimension, and analyze one by one what problems each important indicator reflects, what the indicator measurement criteria are, and how to optimize the video through these indicators. However, it should be noted in advance that the data forms of videos with different purposes, types, and categories will be different. For example, some videos have a high number of likes, some videos have a high comment interaction rate, and some videos focus on sharing and dissemination. The indicator measurement standards recommended below are common measurement standards for most categories. Remember to be flexible when referring to them.
Completion rate
The completion rate refers to the proportion of viewers who have completed 100% of the video viewing progress to all viewers. The completion rate is one of the most important indicators among many video indicators and is a key factor affecting the video playback volume. For project parties, how long users are willing to spend on your product is a very critical indicator. The video completion rate on the TK platform has a very strong correlation with the time spent by the audience. If the video completion rate is high, it means that the audience is more interested in the content of the video and is willing to spend time watching it. The platform often considers such videos to be relatively high-quality videos and will further recommend the video.
Through summarizing practical data, the e-commerce industry generally sets the minimum baseline for measuring whether the completion rate meets the standard at 30%. If the completion rate is lower than 30%, it is considered unacceptable and the video must be optimized [Source: sky]. At this time, think from the perspective of the audience, under what circumstances do the audience not want to watch the video to the end and directly swipe away? Generally speaking, there are several reasons. (1) The first three seconds of the video are not attractive enough. The audience churn rate is often the highest in the first three seconds, so we can examine whether the content in the first three seconds of the video is eye-catching and whether there is a large room for optimization. (2) The first three seconds reveal all the content. Many creators know that the first three seconds of the video must be eye-catching, so they pile the main content and exciting content of the video in the first three seconds, resulting in the audience not having too high expectations for the following content. Without any expectations, the audience will naturally choose to swipe away.
(3) The rhythm is slow. On the TK platform, the audience's patience for each video is limited. If the video does not have a strong sense of rhythm and cannot fully occupy the audience's senses, it will lead to traffic loss. Especially for medium and long videos, the main video needs to have a strong sense of rhythm and a high content density, so that the audience feels that every second he spends on this video is worth it. Therefore, if the completion rate is less than 30%, you can mainly refer to the above three points for optimization.
Based on these principles, the project party can observe whether the number of fans, post data and likes of the influencer match, so as to find the influencer who can really call orders.
Like rate
The formula for calculating the like rate is: like rate = total likes/total playbacks. Compared with the completion rate, the like rate has a smaller impact on the playback volume, but it is still an indicator worth optimizing. When the like rate is lower than 4%, there is room for optimization of the video. From the audience's perspective, they will like the video in two situations: (1) The video has collection value. Many people see good videos while browsing TK. If they do not like and collect them in time, it will be difficult to find the video again later. Therefore, improving the practical value of the video can motivate the audience to like, collect and watch it repeatedly. (2) The video is interesting. Interesting content can stimulate the interest of the audience, which is a key factor affecting the like rate. If the like rate is low, you can focus on improving the fun of the video. Here we can look for those experts who are good at shooting videos.
Comment rate
Comment rate calculation formula: Comment rate = total number of comments / total number of views. When the comment rate is lower than 0.4%, there is room for improvement. Here are three ways to improve the comment rate:
Forwarding rate
Formula for calculating forwarding rate: Forwarding rate = total forwarding volume/total playback volume.
In fact, just like Twitter, a post is considered good only if the number of retweets, comments, likes and impressions match.
Account Indicator: Like Ratio
In the account dimension, the first indicator that deserves special attention is the fan-to-like ratio, which is calculated as follows: fan-to-like ratio = total number of fans/total number of likes. The fan-to-like ratio can very intuitively reflect the fan stickiness of an account. If the fan-to-like ratio is too low, it means that the fan stickiness of the account is low, and the content that attracts the audience only stays at the video level and has not risen to the level of the person's personality or account. The judgment standard for the fan-to-like ratio is generally that if it is greater than 1:6, it is considered that the account has high fan stickiness and more accurate fans; if it is lower than 1:6, it is considered that the video has a certain amount of room for optimization. For example, some accounts have a fan-to-like ratio of 1:15 or 1:20, which is a very typical manifestation of low fan stickiness.
Fans Distribution
The second indicator is the fan distribution. We often simply divide it into gender distribution, regional distribution and age distribution. When choosing crypto experts, you should configure the expert ratio according to the type of your project.
Acknowledgements
I am very grateful to all the friends who provided various information support for this article! I am especially grateful for the strong support of MetaEra. Because I have not participated in many TK operations, a lot of content in this article refers to Sky's book and the experience of my friend Mengmeng who actually does TK operations. I also recommend that friends who want to learn TK operations can read and learn more. The KOL case in the article comes from Wayne. If you need the TK expert resources in the case, you can contact me. You are also welcome to communicate with each other. My TG: SunnyZ_Crypto [Please forgive me for the slow reply]