Using Tik Tok recommendation algorithm for brand promotion

Speaking of the sudden emergence of Toutiao, a very important factor is its recommendation algorithm. It is precisely because of its powerful algorithms and detailed big data that Douyin can understand your preferences and personality clearly.

So let’s talk about this TikTok recommendation algorithm that has been rumored to be so magical.

For any video that passes the official review of Douyin, the system will automatically assign you an initial traffic pool. This traffic pool is not large, with only about one or two hundred video views.

No matter you are a newbie on TikTok or a top-tier internet celebrity, everyone is equal in front of the algorithm, and the initial traffic is limited.

Then, the system will look at your performance under this traffic, including the number of likes, comments, reposts, completion rate, etc. If you perform well in these data, the algorithm will judge you as high-quality content and allocate a larger traffic pool to you.

Then we will look at your data performance in the new traffic pool and decide whether to allocate a larger traffic pool to you.

In this way, with layer upon layer of user verification and layer upon layer of traffic and recommendation, as long as your content is high-quality and user feedback is good, the algorithm will give you more and more traffic, and eventually help you become a phenomenal hit video with hundreds of thousands, millions, or even tens of millions of views, and appear on the Douyin homepage of netizens across the country.

But if your video quality is average and you fail the test of the initial traffic pool, you will be finished as soon as the initial traffic ends.

This is the algorithm of Tik Tok. There are massive amounts of videos uploaded to Tik Tok every day. So what video content is good and what is bad? What kind of video content gets traffic inclination and support? It is this algorithm that Tik Tok relies on to determine the quality of video content and whether to recommend it, which ultimately ensures that good content stands out. For a content platform, the most important things are:

1. Ensure that good content is produced continuously;

2. Push better content to more people, or ensure that everyone can find the content they want to watch on the platform.

High-quality content is supported by traffic, and content producers have greater motivation to create; viewers see good content they want to watch and have more time to watch it, a positive cycle is formed, and the platform can continue to thrive.

If we look through the phenomenon to see the essence of Tik Tok's recommendation algorithm, isn't this actually an application of lean entrepreneurship thinking in content distribution?

This is the concept proposed by Eric Rise in the book "Lean Startup", the core of which is MVP verification (Minimum Viable Product).

Use a low-cost, small-batch production method to verify market reactions and obtain user feedback, thereby continuously trial and error, gaining knowledge from trial and error, and learning about the market. Ultimately, adjust the entrepreneurial path based on continuous product and cognition iteration.

There is also a term consistent with this concept called agile development, which adopts an iterative and step-by-step approach to product development, taking small steps and delivering products as early as possible. The products are then handed over to users for verification. If they are not correct, they are taken offline. If they are correct, they are maintained and modified. This way, the project is advanced in an agile way, and the user experience of the product is continuously improved and iterated.

So let's see that the essential idea of ​​TikTok algorithm is exactly the same as the above-mentioned MVP verification and agile development. First, give the video a minimum traffic volume, and then let the user verify it. If the user likes to watch it, we will give larger traffic volume and even larger traffic volume. If the user does not like to watch it, we will not give any traffic volume.

Since product development can be agile, entrepreneurship can be lean, and video distribution can add traffic, then why can't this idea be applied to branding?

Today, many companies are engaged in marketing. When shooting a video, they may spend hundreds of millions of dollars to promote it; when designing a poster, they may spread and apply it in hundreds of cities and thousands of stores across the country. However, if the video is not touching and the poster design is not right, the losses it causes to the company will be immeasurable.

Although brand creation is so important, it is the root of whether a company’s media budget is wasted. Unfortunately, the decision-making process for brand creation is still very extensive, and the evaluation criteria are extremely subjective, basically the result of a snap decision made by the client's senior management.

What message a brand should convey to consumers and what words it should use depends entirely on the preferences of the business owner. The beauty or ugliness of an advertising work is a faithful reflection of the client's aesthetic level and taste.

The tone of the advertising copy and the tonality of the brand image often depend on whether the client’s boss likes it. If the client's boss is a person born in the 1960s and likes to interfere in marketing decisions, what kind of social communication do you think this company can do?

It can even be said that many of a company's marketing decisions are made based on the preferences and tastes of the client's boss. For example, if the boss likes a certain actress, he will ask her to endorse the company's brand; if the boss likes a certain sport, the brand will sponsor the sporting event. Completely disregarding whether the tone of the sports event matches the brand, and whether corporate brand consumers like the actress.

There is no doubt that the quality of advertising works should be evaluated by consumers. Only those advertisements that consumers like and pay for are good advertisements.

But in the past, companies have conducted user testing before launching advertisements, but it is useless because consumers do not vote in a real environment, and it is difficult to measure real user feedback. Just because consumers choose “good” in the consumer questionnaire does not mean that they will actually vote with their RMB after the advertisement goes online.

So, at this time, companies should learn the TikTok algorithm when placing advertisements. For example, if a company’s marketing goal is ultimately to cover 20 million users and let 20 million users see its own advertisements. Then can I make an advertisement first and show it to 2,000 users?

If these 2,000 users say it’s good, and the play rate and conversion rate are good, then I will show this ad to 200,000 users. If the feedback from these 200,000 users is good, I will then show it to 20 million users. This way, we can avoid inaccurate advertising strategies and lack of motivation in advertising creativity to the greatest extent, and ultimately ensure the communication effect.

As more and more companies today spend their marketing budgets on targeted advertising, this strategy is entirely feasible. Companies can first purchase a small amount of media resources on a daily basis and by click volume, and then optimize advertising materials and media strategies based on user testing of reach and conversion rates before launching large-scale campaigns.

Most importantly, the core of this "agile marketing" approach is not to optimize media delivery strategies, but to optimize corporate brand strategies through user verification, allowing users to decide how a company's brand should be planned and designed, and how to implement development.

When talking about targeted delivery today, everyone likes to use a phrase: one person has one person, one solution. For the same brand’s advertisement, different people should see different advertising materials.

But in reality, the thousands of faces of each person are just different in layout and there is no essential difference. This is actually meaningless. A company does not need to have thousands of faces for thousands of people. On the contrary, a company only needs to produce three materials for advertising:

1. Product functionality-oriented advertising appeal, focusing on USP.

2. Brand image-oriented advertising appeals that emphasize emotional value.

3. Category-status advertising appeals, focusing on sales or status endorsement.

Why these three? Because in short, there are only these three standard paradigms for brand building. Looking at all the domestic brand agencies, advertising companies, and marketing consulting companies, the advertising routines they use and the brand paths they plan for companies are basically one of these three.

Foreign 4A companies usually advise companies to build their brand image and communicate emotional values ​​or values; local advertising companies usually advise companies to promote functions and selling points, which are real; and the prescriptions of marketing consulting companies are mostly to emphasize the status of the category, either saying "sales are far ahead" or "leader in XX category."

So what brand building path should companies take?

Be careful, don't be fooled. Why not make three drafts and test them?

First, spend tens of thousands of dollars to prepare three drafts and submit them together. Then, look at the user data feedback to see which appeal has a higher reach and conversion rate. This can explain a lot of problems.

These three different drafts actually represent three completely different core brand appeals and three completely different ways of brand building. Because the core appeal of a brand is at the center of the brand strategy, it determines many things about the brand, including core value extraction, brand personality, the relationship between the brand and users, brand awareness and association, etc.

Companies can use user verification to determine the direction of future brand strategies. As the entire society and media continue to move towards dataization, programmatic buying will inevitably become the absolute mainstream of advertising in the future. When programmatic buying is popular, companies first spend money to purchase a minimum traffic pool to test the core appeal of the advertisement and brand strategy. After optimizing the brand strategy based on user feedback, they then purchase a larger traffic pool for delivery.

This not only means further precision and economy in marketing expenses, but also a major optimization of the company's marketing decision-making process, allowing users to decide brand strategies and use algorithms to build brands.

Author: Empty-handed

Source: Empty-handed