Douyin marketing plan and advertising number selection strategy

In the past three years, I have managed the entire Douyin marketing cases for hundreds of brands, including new consumer brands such as Huaxizi, Ushiyan, and Neiwai, as well as traditional domestic brands such as OSM, Chando, and Mary Kay, as well as international brand groups such as Unilever, L'Oreal, and Mondelez, and other Fortune 500 companies.

Whether it is new consumption or international brands, whether we are in charge or in-house, everyone has paid a very high tuition fee for the promotion of products through Douyin KOLs. The KOL market changes too fast, platform rules change too fast, and traffic gameplay emerges in endlessly. Without an effective methodology, "content-driven growth" will only be empty talk.

Half a year ago, I wrote an article about my experience in selecting KOL numbers on Douyin. Since then, the Douyin Star Map system has undergone many iterations and many new data indicators have been added. Some time ago, we upgraded the KOL number selection methodology, namely the "CAFE Scientific Number Selection" methodology and algorithm engine system. I share this with you again, hoping that you can avoid detours and pitfalls.

01Content drives growth

In the past era of brand media marketing, brands chose the right track, did a good job of positioning, bought out big media, and repeatedly brainwashed people with TVCs and slogans. The battle at the marketing level was basically over. Big media, big channels, and big distribution were the main themes of past marketing. Ye Maozhong, Hua Yuhua and others created countless classic cases. That was the golden age of top-down communication, which is basically over today.

Today we have entered the content Internet era. Consumers learn about a new brand or a new product more often through KOLs on social media, a product recommendation note, or a short video recommending a good product. Through content-based "KOL recommendations", consumers discover a treasure, ignite their "desire to buy", and then place an order to start trying. From social media promotion to product packaging, copywriting, functions, and user experience, you can feel the brand’s “attitude” through these “contents”.

Content creates consumer mindsets, completes category positioning, and conveys brand attitudes. For many new consumer brands today, content construction is roughly equivalent to brand building. Content drives brand growth and amplifies brand potential. This is a brand opportunity in the new consumer era.

In the 2021 Tmall Double 11 category champion list, 700 new brands topped the category champion list, an increase of nearly 50% over the Double 11 new brands in 2020. There were also 275 new brands whose sales growth on Tmall doubled for three consecutive years.

Let’s take a look at these category champions:

Huaxizi (makeup), Kelaqi (lip gloss), Youshiyan (eye cream), Ubras (lingerie), Sandunban (coffee), Shisanyu (Hanfu), Tineco (floor scrubber), Ulike (hair removal device), BOP (oral care), Moody (contact lenses), Spes (shampoo), Wang Xiaolu (chicken feet), Daily Dark Chocolate (chocolate), Wonderlab (probiotics), etc.

You will find that they have an obvious commonality, that is, they are almost all "content brands", they all focus heavily on "TikTok marketing", create strong "category explosive products", and form a certain "brand potential".

We have entered the era of content Internet, and consumers' perceptions have changed. Brands that can win the future or survive must be content brands or evolve into content brands. Brands without content capabilities will inevitably be eliminated in the future.

02 Seven deadly sins of Douyin's failure to promote products

The first step for a brand to enter Douyin is often to start with the promotion of KOLs. Whether this step is taken well may determine the brand's perception of Douyin. Various failures in the first step often lead to many brands' misjudgment or confusion. The fastest brands take three months to adjust and implement, while the slowest brands still have not formed a correct perception after three years.

I summarize the seven deadly sins of brand promotion failure on Douyin:

The first sin: The founder failed to personally participate in the game

The founders or decision-makers lack knowledge about TikTok and have no awareness of content-driven growth. There are often many misjudgments in the company's highest decision-making, and there are no clear and effective goals.

This year, I will still meet brands with a scale of over 1 billion, and their founders are asking me whether they want to enter Douyin and why. I think we should ask HOW instead of WHY.

The second sin: ROI-only KPIs

The legend of KOL’s high ROI lasted until the first half of 2019. We also invested in many cases with ROI 1.5+, but that era is over. After that, it was more about intensive cultivation, strategic combination, and ROI orientation, which made many brands miss the best entry period, resulting in higher costs later.

Last year, when I was serving a client, the KPI requirement for the grass-roots product was ROI greater than 1. Therefore, the budget could not be spent every month. The brand reduced its investment, momentum declined, and sales declined. Therefore, the budget was controlled even more, forming a vicious cycle.

The third sin: using Dabo as a weed

There is a misunderstanding about the differences between planting grass and Dabo. Planting grass is a boost to brand assets, which is "saving money", while Dabo is a consumption of brand assets, which is "withdrawing money". Dabo is reaping the benefits of the brand and will only reap less and less.

In the early years, there was a very well-known brand. It reaped the benefits of live streaming crazily. The price cuts in live streaming damaged the channel ecology. When the benefits faded, the brand sales went from bad to worse. If it didn't broadcast, it would die. Broadcasting is like drinking poison to quench thirst. In my opinion, the brand is dead.

The fourth sin: wavering between in-house and agency

Some brands believe that content capabilities must be firmly in their own hands. Once they make up their minds, they and their in-house team will have to go into the game and accept trial and error and losses. Some brands believe that professional matters should be handled by professionals, so they should work with the most professional companies and teams. Both of these brands have developed well in recent years, but brands that waver in between have wasted a lot of budget and precious time.

The highest pursuit of in-house is Hua Xizi, at least it is the direction that the founder is determined to work towards. There is another possibility. A billion-plus brand has been working with us for three years, and the brand team has only more than 30 people. They pursue professional division of labor and ultimate human efficiency.

The fifth sin: too many unscrupulous MCNs and service agencies

When the brand first entered the market, no one could avoid this big pit. MCN has incubated a large number of standardized accounts for the purpose of commercialization. In order to meet agency data requirements, ensure CPM and CPE, falsify KOL data in various ways, and even establish a complete sewer system to ensure ROI.

This year, I helped a consulting client to do a KOL review, and the results were shocking. They found that 80% of the accounts they invested in were fake accounts, and at least 70% of the data was false. I asked about KPI and agency charging model. KPI is to guarantee CPM and CPE, so fake data has to be generated. Agency does not charge service fees, so it can only promote water accounts with a rebate of 20-40%. However, many of the high-quality KOLs we invested in today do not receive any rebate at all. This result is inversely proportional to the customer's requirement, saving 10% in service fees and increasing 90% in investment losses.

Sin 6: Not Amplifying Content Traffic

I chose the right account, prepared the content, and invested in Dou+, thinking that this was effective promotion. I was totally wrong! The cost of KOL is so high that if the traffic of high-quality content cannot be amplified by 50-100 times, it would be a huge waste of content.

In many of the cases we have invested in in the past, content traffic (Dou+ category/content services/expert bidding) accounted for 1 to multiple times of the KOL content budget, and the ceiling for a single video was 50 million views. This is why there is a higher ROI behind popular videos, because the content traffic cost is low and the ROI is high.

The seventh deadly sin: core team members leave

If core team members leave, the first six sins may happen all over again.

This year, the founder of a new consumer brand approached me. We had worked together two years ago and then the in-house team took over. We did a good job during the bonus period, but the environment has been changing and the brand has encountered difficulties. The most painful thing is the departure of core members, which made the situation worse. My advice is still to stay in-house and don’t give up, but you can use external brains to make up for your shortcomings and you can also introduce proxy horse racing in the process.

Among the seven deadly sins listed above, I would like to ask friends from the brand side, how many of them have you committed?

03CAFE Scientific Number Selection Methodology

If there is no scientific Douyin content system, every brand will have to go through the same detours as its predecessors. The first step to promote Douyin is to start with "scientific account selection".

My personal background is as an Internet product manager. In the field of TikTok marketing growth, the logic has always been technology and data driven. A long time ago, we were studying the possibility of using the Tik Tok KOL algorithm to select numbers, so we came up with today's "CAFE" methodology and technical algorithms.

C: Communication

A: Commercial Power – Advertising

F: Fanspower

E: Expansion

CAFE is an "algorithm engine" for Douyin KOLs that is implemented through technical means based on the experience of a large number of Douyin KOLs in selecting accounts. It breaks down KOL data into four dimensions: communication, business, fans, and growth to calculate the value, and obtains the "KOL RANK" guided by the ability to plant grass (bring goods). At the same time, it quantifies the core data indicators of KOLs to form an effective and scientific account selection system, avoid water accounts and fake accounts, improve the "efficiency" of account selection, and improve the "effect" of delivery.

The CAFE algorithm rules cannot be made public, but according to the four dimensions of the algorithm breakdown and the core indicators of KOL quantification, as long as you follow the logic, you can still avoid more pitfalls and at least improve the accuracy of KOL number selection by 50%+.

04CAFE: Communication

Communication

From the perspective of video content data, to identify the fundamentals of the "communication side", we generally analyze it based on several core indicators such as interaction rate, completion rate, median playback volume, CPM, account activity, etc. Because the communication power reflects the basic content capabilities of the KOL, data comparison and analysis between personal videos and business order (Star Chart) videos must be done to avoid the tragedy of having excellent personal videos but failing once they take on advertising.

Interaction rate: number of interactive behaviors (likes + comments + shares) / number of plays

This data is an indicator with likes as the large base. We often compare the interaction rate with the like rate. Vertical bloggers often have high like rates and low interaction rates, while the data for non-vertical bloggers (such as plot accounts) are just the opposite. This indicator has average weight for vertical talk show hosts and is for reference only.

If the KOL data is falsified, the interaction rate will often be very high. If it exceeds 10%, you should be alert. There is a high probability that the data is abnormal. It is either fake data or Dou+ maintaining the data. In short, it is not a good thing.

Completion rate: number of complete plays/number of plays

The completion rate to a certain extent represents the KOL's ability to create content and fan stickiness, but this is not absolute. First, it depends on the length of the video. Luo Wangyu's average business order is 3 minutes, and the completion rate is less than 1%, but it does not affect the ability to bring goods. Second, it depends on the type of content. Some organizations are good at riding on hot topics, but that does not mean that the content is commercially viable.

The completion rate of vertical bloggers' commercial videos is lower than that of their personal videos. However, if the difference is too large, there will be a trend to a certain extent. Even if the individual is excellent but the advertisement fails, it will be useless if no one buys it.

Median playback volume: playback volume of Star Map works (organic playback volume)

This is a new indicator that is constantly iterating on the Star Map backend. Both 30-day and 90-day indicators are good references, but it also reflects a very cruel fact. For mid- and low-end vertical categories, the median number of KOLs quoted by Star Map in the 50,000-100,000 range may be only 300,000 in real terms. The elasticity above and below shows the cost-effectiveness.

Based on this number of views, the ROI may be disappointing, but is it really meaningless? Obviously not. Investing in vertical KOLs is for the "content model" to ultimately achieve "amplification of the grass-planting effect" and "ROI enhancement" through content traffic.

CPM: Cost per thousand views

Corresponding to the CPM translated from the median playback volume, in the strong scene non-vertical grass-planting strategy, the content strength indicators, especially the CPM indicators, are highly valued. If it is a non-vertical KOL, there is a high probability that there is no content investment and traffic support (the content model is not suitable for investment and traffic amplification). If the CPM is lower than expected, that is, it is impossible to achieve broad exposure, that is, the CPM is low enough, and it is impossible to achieve deep planting, that is, the content model has a high conversion rate, then give up quickly.

Account activity: KOL's monthly video update frequency

If a KOL only updates a few videos every month, what is the status of this account? It is possible that the KOL has given up or is in poor condition recently. The update frequency cannot determine the quality of the content, but low-frequency updates are fatal on short video platforms. If a KOL updates frequently every month, but more than 80% of their posts are commercial works, what does this mean? This account may be created for the purpose of taking orders, or it may be too commercialized, neither of which is a good thing.

A stable update frequency (10+ updates per month) and a balance between personal videos and business videos are all manifestations of KOLs’ responsibility to their fans. If the foundation of content and the support of fans are lost, there will be no value in cooperating with them.

05CAFE：Advertising

Commercial Power – Advertising

From the perspective of KOL's recommendation and sales data, to identify the fundamentals of the "conversion side", we usually analyze it from several core indicators such as the recommendation index, comment and like rate, shopping cart click-through rate, GPM, category preference, etc. In the grass-planting strategy of vertical KOLs, commercial power indicators are highly valued.

Like rate: (comment + share) / like

The core indicator based on comments and shares is a strong stickiness indicator. If a short video has a large number of comments and shares, there may be three situations:

The first type is high-quality content that resonates with people and generates a lot of discussion;

The second type is that bloggers have strong fan stickiness and high fan engagement;

The third type is that the content is valuable and is shared by users with others.

No matter which of these situations is a very good phenomenon, it will be fully verified from the perspective of bringing goods. The ideal rate of vertical KOL likes is 5%+.

I have recently been studying Juliang Cloud Map (TikTok version of Data Bank), the A3 population, that is, the potential purchasing population. I have broken down the weight of acquiring the A3 population. I will not go into details about the process, but the conclusion is that planting grass (commenting, sharing, searching) is the most efficient strategy to acquire the A3 population, without a doubt. The like rate will become the compass of the A3 population strategy.

Shopping cart click rate: shopping cart clicks/shopping cart plays

From how many people watched the video, that is, the number of views, to how many people clicked on the shared yellow cart, that is, the number of shopping cart clicks. Shopping cart clicks represent potential store-entry behavior and represent store-entry efficiency. If this data reaches the Tmall link, there will probably be a loss of more than 50%, but there will be no loss in the closed loop of a small store.

The shopping cart click-through rates vary greatly among different categories. As more and more small stores join in, Douyin already supports differentiating shopping cart click-through rates by category. Of course, the higher this value, the better.

Because the shopping cart click-through rate data is based on the shopping cart formation playback volume rather than the video playback volume, this means that, first, we can check whether the KOL is brushing fake data; second, we can determine the reasons why some KOLs’ shopping cart click-through rates are inflated.

GPM: Thousands of views

Once the KOL has added a shopping cart to the Douyin store, the data related to the sales will be counted. Currently, the amount of data samples is still relatively limited, but it is already of reference value. There are several other indicators related to GPM, such as the shopping cart click-through rate and the sales range reference. Although they are all range values, the sales range can be estimated by using the playback volume corresponding to the upper and lower limits of GPM (about 2 times), which can be reduced by about 50-70% compared to the range given by the star map, and the accuracy will be higher.

In addition, we use the calculation principle of the Dou+ model and deduce it with GPM. We can roughly estimate the KOL's content investment model, that is, the "content model" mentioned earlier, and use it as a reference for pre-selection of numbers, and even allocate content investment budget in advance.

Category Preference

The same KOL may show completely different results when recommending different categories and products, and the ability to bring goods may even differ by 10 times. In fact, KOLs themselves have obvious "category preferences". It may be because of their personal positioning, or it may be based on certain types of content. Douyin has given KOLs obvious labels.

In the past, we would pay more attention to the "explosive articles" on the front end to make judgments. Today, Star Map's backend product analysis can be checked again with the front-end data to provide double insurance for category preferences. Therefore, in the process of selecting numbers, KOLs who can match category preferences will show higher efficiency in their ability to bring goods.

06CAFE: Fanspower

Fanspower

From the perspective of KOL fan portrait, to identify the fundamentals of the "fan side" of fan purchasing power, we generally analyze it based on several core indicators such as heavy fan activity, 24-30 age index, iPhone index, proportion of female fans, and proportion of first-tier cities. The upgraded algorithm will compare the video viewing portrait with the fan portrait again. If there is a large difference, there is a high probability that the fan portrait is suspected of being fake.

Activity Index:

Generally, we look at the ratio of "heavy fan activity", which is closely related to the quality of fans. The higher the heavy activity, the better the quality of fans. Generally, the ideal ratio is 70%+. Because this indicator is too explicit and will greatly affect the acceptance of orders from water accounts, this indicator has been taken offline on the Star Map front desk, but there are still ways to obtain it.

iPhone Index:

The proportion of iPhone users among fans is also an index of white, rich and beautiful. Although many domestic mobile phones are becoming more and more high-end today, the proportion of iPhone is still like a weather vane, becoming the standard for the high proportion of those top KOLs who bring products. Generally the ideal range is 40%+.

24-30 Index:

People in the age group of 24-30 are the backbone of consumption. The proportion of this group of users can reflect the consumption quality of the influencer’s fans. Generally, the ideal range is 20%+, the more the better. In addition, in terms of age index, if the KOL fan data is falsified, there will often be huge anomalies in the age ratio when comparing the audience portrait with the fan portrait, which means that the population behind the KOL's promotion is distorted. No matter how the content is created, it is highly likely that it will not be effective.

First-tier city index:

Star Map has increased the proportion of people in first-tier, new first-tier, and second-tier cities. We analyzed the city dimension and found that KOLs with the ability to promote products have a higher proportion in first-tier and new first-tier cities. On the one hand, this is related to the purchasing power of urban people. On the other hand, it also shows that the dissemination starts from the top population and then gradually penetrates into the sinking market. Of course, this index can also be used in reverse. If you focus on the sinking market, you can focus on the proportion of people in third-tier cities and below. The higher the better.

Female fan index:

Most of the grass-planting products are still dominated by female users. The proportion of female fans is of great significance, generally at least 70%+. In addition, when comparing the audience portrait and the fan portrait, if the KOL is fake or has fake fans, the comparison will also reveal a large difference rate.

07CAFE：Expansion

Growth

From the perspective of the speed at which KOLs gain followers, to identify the fundamentals of the "growth side", we generally analyze it based on several core indicators, such as the follower growth index, the number of followers gained in 30 days, and the number of followers gained in 90 days.

30-day/90-day increase in followers:

Based on the data captured by this indicator, we can still find KOLs with very good "growth potential". The most cost-effective stage for most KOLs is the growth stage from the tail to the middle. The most obvious change is that the speed of fan growth in 30 days and 90 days is very fast. Investment in this stage can reap the dividends of the KOL growth period.

This year, Douyin carried out a KOL fan cleaning operation, eliminating many KOLs with abnormal fans. As a result, you will see that many water accounts have lost extremely severe fans, often losing tens of thousands of fans a day. This means that when they were increasing their fans, they were increasing their fans by tens of thousands a day. If you encounter such accounts, please choose with caution.

08Super Content System

Is planting grass equal to the content system? Obviously not entirely. Selecting the number is the first step and the most important link. Then comes KOL content creation, and then the potential amplification effect of content traffic.

What is the logic behind KOL’s recommendation? I think it's taking advantage of the situation.

Today we say that "brand power" is quite vague and its timeliness is difficult to measure, while "brand potential" represents whether the brand is popular recently, which consumers can feel more clearly. Consumers' perception is not affected by the brand's will, but they ultimately vote with their feet and make the best choice.

What is the ultimate goal of brands promoting products on Douyin? I think it is to create a "hot product effect" and exert "brand potential".

In order to effectively promote brands on Douyin, a complete “super content system” is needed.

The first step is to start with "scientific number selection", starting from the KOL data fundamentals, breaking down the core data dimensions, scientifically measuring data indicators, and judging the KOL's grass-planting value.

The second step is to move on to refined "content creation". KOLs' popular videos all have mature content frameworks and routines. In essence, we are pursuing a "content model" that is grass-planting, and ultimately amplifies the content effect through content traffic.

The third step is the "brand-effect delivery" of the content amplification effect. With the support of content traffic (Dou+/content services/expert bidding), the traffic of KOL's high-quality content can be amplified 10-100 times, and finally the content and traffic can be penetrated. This does not refer to AD or UD commercial traffic, but only the amplification of native content traffic.

09 Conclusion

A few days ago, I was invited to a roundtable forum at the Knife Growth Summit. The host asked me: Can the TikTok marketing methodology be replicated? My answer is yes.

We entered the field of Douyin's marketing growth in 2019, from our first client Hua Xizi to serving hundreds of brands in three years. We have witnessed new consumer brands taking the lead and achieving sales of 5 billion in three years, witnessed the successful transformation of traditional brand Douyin and the success of international brands in creating popular products by copying cutting-edge strategies.

Behind the growth of these brands is the underlying logic of "content-driven growth". Almost all of the strategies can be copied, and Douyin no longer has any secrets.

The ultimate outcome of a brand is a competition of understanding of the underlying logic of growth, a challenge to the founder’s cognition, and the application of methodology and refined operations, and a competition of team execution.

Author: Growth on the road

Source: Growth on the way