When it comes to traffic, you can’t escape the BATT circle.

What is BATT?

Tencent, Alibaba, Baidu, and ByteDance.

Internet traffic will only become more and more focused, and the mainstream population will go to one fish pond. If you can't master the mainstream Internet platform, you won't be able to make a living in Internet marketing.

Traffic is like water flow. In the past, all the traffic was in small rivers and streams. You can play in any stream. As long as you can find a customer acquisition channel, you can survive in the market;

It’s different now. Traffic flows from small streams into the sea. Ten years ago, who would have thought that a content platform focusing on entertainment would have 600 million daily active users (Tik Tok).

ByteDance is an amazing company. They have been invincible since their first product, Toutiao. Today, Xiucai will talk to you about how to operate a Toutiao account.

In 2012, ByteDance began to launch the internal version of Toutiao. Compared with several other antique applications, the new Toutiao account is like a baby and not worth mentioning. But surprisingly, in the next eight years, the number of Toutiao users continued to increase and the user base continued to grow.

As of 2019, Toutiao has achieved impressive results of 700 million users and 200 million daily active users. In the Internet industry, Toutiao is a traffic platform that cannot be avoided in any way.

The writing threshold for self-media on Toutiao is much lower than that on other platforms, and you can see profits immediately. For Internet newcomers, Toutiao is a platform worth trying.

one

Since we are going to study Toutiao, we must start with the Toutiao algorithm.

Toutiao’s three algorithm mechanisms:

Recommendation mechanism: Toutiao’s recommendation mechanism is very similar to that of TikTok. The platform will first give the first wave of recommended traffic, which is about four to five hundred. Then the platform will see whether the article can be recommended to the next traffic pool based on user feedback such as article click-through rate, reading time, and forwarding. If the article has good data feedback, it will continue to be recommended to the next wave of traffic pools, and the cycle will continue until the data fails to meet the recommendation standards and the dissemination is interrupted.

It is very important to understand this algorithm, as it reveals to us the underlying logic of creating a headline account - the machine cannot recognize whether your content is high-quality. As long as your article can be read to the end and readers can interact with it, the platform will determine that it is a high-quality content, even though the content of the article may be rubbish.

In this case, as long as we can figure out the platform's recommendation mechanism and understand the platform's preferences, it is possible to write a hit article that [disregards] writing ability.

Deduplication mechanism:

Can I find an article on Baidu and move it to Toutiao? doesn't work.

After you copy a few articles, you will find that this account no longer has any recommendations.

Although the ancients said that "most articles in the world are plagiarized", from the perspective of Toutiao, I paid you to create content. If you keep sending garbage to me to make money for the platform, then I will lose a lot of money.

But then again, the intelligent algorithm is just a retard and it cannot tell the difference in content. As long as the copied articles are re-created so that the algorithm cannot recognize them, copying them is a feasible approach.

Here I will first solve a problem that you may encounter when making headlines. Why does Toutiao have no recommendations?

One reason is that the repetition is too high and there is a lot of similar content on the platform; the second reason is that the sensitive word detection mechanism is triggered. Of course, if the article contains hard advertisements, it may also be a factor that leads to the lack of reading volume of the article.

Tag mechanism:

Tags are a mechanic that runs throughout the headlines. Creators have creation tags, users have reader tags, and even every article contains different tags. Tags are an important mechanism that affects the amount of recommendations.

It’s like if you have been creating content in the emotional field before, and now you start writing about cars, the platform algorithm will be confused and will not know where to recommend your article.

On the other hand, if you write an article and the keywords of this article are accurately identified by the system and match the user's tags, then the article will definitely perform very well.

two

Now that we have finished talking about platform algorithms, let’s talk about how to create content.

The first item of content is the title. Xiucai will teach you how to write a title that will make readers want to click.

Five tips for creating titles:

1. Low repetition

Toutiao doesn’t like repeated articles. For example, everyone is following an entertainment hotspot. The first person writes "divorce + mistress" in the title and gets 1,000,000 reads. The second person writes "divorce + mistress", but the platform only gives him 200,000 reads. The number of reads is gradually reduced. By the time you write, the platform may only give you 1,000 reads.

There is a good way to avoid this: after thinking of the title, search it in the headlines search box. If there is a high degree of repetition, change the keywords.

2. Use numbers wisely

The human subconscious mind does not like content that is difficult to understand. Nowadays, readers like to scan articles in order to obtain the most information in the shortest time. Titles with numbers are more easily accepted by readers than titles without numbers.

If you don’t believe me, let’s try it. Did you not understand the sentence “The benefits of using numbers in titles” after reading it once? “3 benefits of using numbers in titles”, although the title is more attractive, the emphasis is still on “using numbers in titles”. Then look at “3 benefits of using numbers in titles”, which is simple and easy to understand, and will induce readers to click involuntarily.

3. Benefits first

Why do readers open your article? Because he has needs and desires. Putting the interest point directly in the title can make readers more interested in clicking on the article to read.

"Use these apps to make your phone photos more beautiful."

In the title, tell the readers what problem my article is solving. The more targeted the problem, the more readers will want to open and read it. There is no empty talk or hazy beauty. It is telling readers that all the dry goods you want are here, come quickly!

4. Arouse curiosity

"How do you comfort a child who's having a tantrum? The methods in these picture books are amazing."

Curiosity is human nature. Abstracting targeted questions in the title can arouse readers' curiosity and encourage them to click to read.

5. Create resonance

"What you think of as rich upbringing is nothing but unprincipled doting"

The title's viewpoint resonates with readers, making them more likely to click on the article and read it.

Still confused after reading this? You don’t need to remember all the above title-making skills. The purpose of writing them here is to let you know “what is a good article title”. In the actual operation process, especially the operation of Toutiao, what we need is assembly line production of content.

How to write a good headline?

1. Open the Yizhuan website:

https://www.yizhuan5.com/work.html#1-4

2. Click on the We-Media Library-Best Articles

3. Collect more exciting article titles.

Collect good title frameworks, replace all the keywords with your own keywords, so that you can create a production-line explosive title.

three

The title of self-media is the easiest module to use. You can create a good title by modifying one framework after another.

How to create original content?

Answer: Move the hot products from platform A, perform secondary processing, and then move them to platform B.

There are many self-media platforms nowadays: Toutiao, Baidu Baike, Baidu Xiongzhang, Jianshu, Penguin, Yidian Zixun, Sohu Public Platform, UC Dayu, NetEase, Sina Kandian, 360 Self-Media, Qutoutiao, Dongfanghao, etc.

Moreover, the content positioning of these self-media platforms is relatively similar. Content that is popular on platform A can also be popular on platform B.

Each platform is a material library for making headlines. Next, I will teach you how to perform secondary processing and produce original headline articles in an assembly line.

The secondary creation of self-media is nothing more than three techniques of addition, deletion and modification.

Addition: Replace all the pictures in the original self-media article, add materials on copyright-free websites, or use the article keywords to search for other similar articles, and use Meitu XiuXiu to slightly modify their article pictures, change the brightness, add some color, and they will be brand new pictures. You can also insert a famous quote into the original text for differentiated processing.

Delete: Delete part of the content in the article. For example, if this article quotes a story about the Tortoise and the Hare, you can search for a similar story on Baidu and insert it.

Revision: Revision from the beginning to the end of the article, and retell the original text in your own words. For example, if the original sentence in the article is [The scholar is a handsome guy], then you can use your own words to say, [The scholar is not only talented, but also has an extraordinary beauty]. You can also put the content of the comment section of the original article into the article to make it a brand new article.

An article on platform A becomes an article on platform B after the above processing. This article is original on platform B and can pass the headline de-duplication test.

Four

Article structure:

5-10 pictures and 500-1000 words will do.

The shorter the article, the higher the completion rate. If you look at the articles on QQ Kandian, they all start with a picture, the content is all made up, and they will also deliberately provoke readers to curse. But what those readers don’t know is that [the harsher the insults, the more popular the article will be, the easier it will be to spread], and the higher the creator’s income will be.

Want to make money from Toutiao? Just be a staunch title-maker. Anyway, neither the writers nor the readers believe it. It’s strange that the platform is getting stronger and stronger for no apparent reason.

When it comes to making headline content, it’s all routine. Because machines cannot identify whether an article is good or not, being able to hit the article tags is the beginning of a popular article.

The machine cannot recognize your article. It can only recognize the tags corresponding to the keywords, and the tags correspond to the user groups.

How to ensure that the article will be tagged by the machine? That is to increase the keyword density in the article.

For example, if your article is about whitening, then the keyword [whitening] should appear once in every paragraph of your article. This is an SEO optimization method. The arrangement of keywords can make it easier for Toutiao to identify the keywords of the article and recommend the content to people with the same tags.

Let me add one more point. Since we have said that whether content is spread is determined by user feedback, if you want to get article exposure, you can write controversial content. You can even use a small account to incite users' emotions in the comments at the end of the article. The small account deliberately leaves a message at the bottom of the article [Looking for someone to scold]. This is the detail of operating a headline .

The creation of templated content is actually not friendly to most people. Because the essence of machine recommendation is [user behavior], the platform will only make recommendations based on content that users like, which means that unless you actively jump out of this circle, you will not receive differentiated opinions and information. The content recommended to you by the system is all homogeneous, which is the origin of the information cocoon. If you can't get out of the information cocoon, it's easy to become a slave to information.

In your opinion, should the blame for templated content be placed on the creators or the users? Or is it this damn algorithm?

Author: Xiucai Youliao

Source: Xiucaiyouliao