Is TikTok worth buying without the algorithm? Technology experts: It will be difficult for buyers to replicate its magic

Is TikTok worth buying without the algorithm? Technology experts: It will be difficult for buyers to replicate its magic

This article is reprinted with permission from AI new media Quantum Bit (public account ID: QbitAI). Please contact the source for reprinting.

“If TikTok’s recommendation algorithm is not included in the deal, it will be difficult for the company that buys TikTok to replicate its magic.”

Several foreign technology experts said this in an interview with US media Business Insider.

[[341240]]

One of the important reasons is not the algorithm itself, but the data.

According to Sensor Tower data, Douyin and TikTok's global downloads exceeded 2 billion times in May this year. In the United States alone, TikTok has 100 million users.

Based on such a huge amount of user data, TikTok has trained a mature recommendation model. The data set composed of these real data is really difficult for other companies to imitate.

How TikTok recommends videos

Here, let’s briefly talk about how TikTok uses algorithms to achieve “viral” spread.

The most important thing, of course, is the recommendation system.

Personalized recommendations are based on some behaviors of users during actual use. According to TikTok official information, they mainly include:

  • User interactions: Such as videos that users like or share, accounts they follow, comments they make, and short videos that they shoot themselves.
  • Video information. May include details such as title, sound, and tags.
  • Device and account settings. Such as language preference, country setting, and device type.

All of these factors are processed by the recommendation system and weighted according to their value to the user. For example, if a user watches a long video from beginning to end, this metric will be given more weight.

The development and maintenance of the recommendation system is a continuous process. TikTok will continue to adjust the model, improve accuracy, and re-evaluate the factors and weights that contribute to the recommendation based on user feedback, research, and data.

[[341241]]

Data is the key

According to Business Insider, Bondy Valdovinos Kaye, a researcher at Queensland University of Technology, said:

I personally think that without this algorithm, TikTok would not be TikTok.

Sabba Keynejad, founder of video editing app Veed, said that she has not seen any company that can really catch up with TikTok in terms of recommendation systems.

However, purely from an algorithmic perspective, the technology used by TikTok is based on an industry-standard architecture.

The difference is that "each click reveals the user's preferences, and to some extent, provides TikTok with more useful information about user preferences," and TikTok has a huge user base, said Nikita Aggarwal of the Oxford Internet Institute.

[[341242]]

Moreover, TikTok can test user interest on a scale unimaginable for long-form video platforms like YouTube.

TikTok can provide users with hundreds of short videos of no more than one minute every hour. Eugene Wei, former product manager of Amazon, believes:

It only takes a few seconds of feedback to register how I feel about a video, and since TikTok is a full-screen app, it can assume that anything I do is a reflection of how I feel about the video.

Previously, Sensor Tower mobile strategist Craig Chapple also commented:

TikTok's continued rise gives it a larger foundation not only to succeed in the short term, but also to keep existing and potential competitors out in the long term.

Therefore, if the algorithm is not included in the TikTok transaction, then even if they provide all the user data and video data, they will have to start all over again to reproduce the "magic" of its viral spread - that is, it is actually quite difficult for the buyer to replicate the success of the TikTok algorithm.

Nikita Aggarwal also pointed out that historical training is clearly of great value.

Is TikTok still worth buying without the algorithm?

On August 28, the Ministry of Commerce and the Ministry of Science and Technology adjusted and released the "Catalogue of Technologies Prohibited and Restricted Export from China".

The clause "personalized information push service technology based on data analysis" was added to the catalog, which was generally interpreted by the outside world as directly targeting TikTok's algorithm technology.

After the new regulations were introduced, the Wall Street Journal quoted sources as saying that TikTok's acquisition negotiations had reached a deadlock.

According to the report, a source said that for bidders, "TikTok without algorithms is like a luxury car with a cheap engine" because a large part of TikTok's value lies in its core algorithm. If the transaction does not include the algorithm, it will completely change their vision of acquiring TikTok.

Another person close to the bidder said that if the algorithm cannot be acquired, the deal is likely to fail.

But it was also revealed that one of the bidders could still buy TikTok without the algorithm and then write a new algorithm for it.

In this regard, US industry insiders predict that the deal to sell TikTok may be postponed until after the US election in November.

Currently, there are two major groups participating in the bidding for TikTok. One is an alliance formed by Microsoft and Walmart; on the other hand, Oracle has cooperated with investors including Sequoia Capital and General Atlantic.

<<:  Say goodbye to data cables! Transfer files between mobile and computer faster

>>:  Research shows that developers are using user moods to get more five-star reviews

Recommend

Study on debugging Android Native C++ memory leaks

[[433799]] This article is reprinted from the WeC...

A must-have user growth map for new product developers

User growth is an eternal topic in the Internet c...

Christmas and New Year's Day marketing plan (attached with download)

Christmas and New Year's Day are important fe...

up to date! Data rankings of 56 information flow advertising media platforms!

The following is the latest traffic rankings of 5...

How an ice crystal can teach us more about our environment and climate

When a fluffy, soft cloud appears on the horizon,...

Foreign media: Apple's performance in 2016 was really not very good

According to foreign media reports, after a one o...

The method of increasing users by “old customers bringing in new customers”!

Before talking about the method of “old employees...

A shark "mummy" was found in an abandoned aquarium!

In the past half month, we have collected these i...