Divine Translation Bureau is a compilation team under 36Kr, focusing on science and technology, business, workplace, life and other fields, focusing on introducing new technologies, new ideas and new trends from abroad.
Editor's note: Why are there so many chaos in Western social media such as Facebook and Twitter? What lessons should the next wave of social networking start-ups learn? Most of the most famous western social applications choose to combine social graphs with dynamic messages. A product manager believes that using social maps to generate interest maps is the original sin of western social media. The article comes from compilation and space, and we publish it in three parts. This is the second part.
Image source: Pexels
Focus on:
China's way of building social infrastructure is more logical.
Western social applications are also more dependent on advertising revenue.
Social media interferes with the formation of users' social maps.
The harm of map design errors is that such errors will only appear after the app reaches a certain product market matching degree.
Most applications do not choose to fundamentally solve the problem of map design, but choose tinkering.
Tik Tok is an internal map built as an internal map.
This is the fundamental problem of western social networks: to die, not to die.
But again, this is a unique problem in western social media design.
When we combine social graph and interest graph, we introduce an unnecessary content matching problem.
. I don't feel uneasy because my friends don't follow me on Tik Tok or Reddit, or I think this is purely an interest and/or entertainment network. Those products are very clear, and everyone should pay attention to their own interests.
At least in this respect,
China's way of building social infrastructure is more logical.
. Wechat has a dominant social graph and serves as the underlying social infrastructure for other parts of the Internet in China. Other apps don't have to copy everyone's social map owned by WeChat, so they can focus on what they are best at, and may or may not need to back up the map.
Western social applications are also more dependent on advertising revenue.
The lifeblood of their income statement is the flow of dynamic news.
. This means that the relevance of dynamic messages is the most important. After an individual's social graph deviates from his personal interests anywhere, boring content will invade dynamic news. The signal-to-noise ratio shifts in the wrong direction. Instead of trimming and adjusting their social charts to fix their dynamic news, most users did the simplest thing: walk away.
As a product manager or designer of social applications, you may disagree with this view. Pay attention to who is the user's own choice and who is the choice of other users. You can't control this. But this statement ignores this point:
The App puts its hand on the scale and tries its best to push every user to a specific type of map.
Take the initial user registration process as an example. Every week on a new social application, Testflight, I seem to encounter such a modal dialog box:
What I want to find is where this requirement appears and how this app is considered. In most cases, before users know what an application does, they will be asked to authorize access to their address book to pay attention to any matching users. Urging others to copy their address books, these apps obviously choose to do things according to everyone's social map in the real world.
Not surprisingly, social applications listed this license as a key part of the registration process. IOS address book is now the only "open source social map", and new apps can use it to quickly promote their own apps. In the article "Network is the Key", I think the network itself provides the largest share of value for social networks. As for what kind of content is allowed to enter dynamic messages and how to determine the format of these contents, these are not so important. In a short time, large social maps such as Facebook or Twitter allow third-party applications to use these maps, and even allow batch copying.
As we all know, Instagram has made a good start by drawing on Twitter's social graph to build its own social graph. But it didn't take long for these companies to realize that they were providing ammunition to future competitors. So they
Began to strictly restrict access to their own maps.
. You can also provide Facebook or Twitter authentication for your application as an authentication option, but if you want your own social graph, the mobile phone address book is the easiest to get started.
There is another way for App to really influence its social graph, that is, to recommend the attention list. These often appear when users browse the app for the first time, or interspersed in dynamic messages, and sometimes appear next to dynamic messages.
Early Twitter users were very lucky. Their names appeared on the recommended follow-up list of the first edition of Twitter, which made them have hundreds of thousands or even millions of followers today, because they were displayed in front of every new user.
This is a huge social capital subsidy, but I find many options in this list puzzling. A few years ago, a friend first set up his Twitter account. He showed me the list of accounts recommended by Twitter when he registered. There's Donald Trump in it. No matter what your political inclination is, this choice is suspicious. Let's push every new user to the political orientation of Twitter, and this is the worst Twitter.
For some people, such as those who often go to fight clubs on weekends, politicized Twitter may be a perfect dopamine injection, but when users just sign up and Twitter knows nothing about them, it is an incredible gamble.
For years, people have been surprised by Facebook's "suggest friends" gadget. Wow, how did they know I knew that man? Yes, of course I will make friends with them. However, as mentioned above, considering the way of constructing dynamic messages, this may be a map design error.
On the other hand, it is also important to help users get the right type of followers. At this point, the influence of fanaticism is two-way. Cult leaders use their charm to develop fans, and then these fans will in turn shape cult leaders. This is a feedback loop, but it is not necessarily healthy.
In addition to one-way door errors, map design errors are also very harmful.
Because this kind of error will only appear after the app reaches a certain degree of product market matching.
. But by that time, not only the social graph that has been formed is difficult to cancel, but it will also violate the expectations of users who have accepted the app. By this time, you have fallen into a dilemma. You have done something to die, not something to die. After an app reaches a certain degree of product market fit, even if it is partially maximized, it needs real courage to recover.
But this does not stop social applications from trying to solve problems. For many social applications, reducing the traffic of dynamic messages is a matter of life and death. however
Most applications do not choose to fundamentally solve the problem of map design, but choose tinkering.
. The most popular method is to switch to providing dynamic messages by algorithm rather than in chronological order. The task of this algorithm is to filter the content of the account you choose to pay attention to. I want to recover the signal from the noise. In order to determine what to keep and what to discard, dynamic message algorithms will check various signals, but basically, they all want to guess what will attract you.
Nevertheless, this is still an expedient measure for upstream mistakes. Look at Facebook, they
Every few years, you will swing between news content and more personalized content from people you know.
. Unless they admit that the fundamental problem is to find dynamic news in the huge social graph, they will never really solve the problem of customer churn. However, if they give up the basic architecture of dynamic messaging, it will be the boldest decision they have made in a long history. This is not only because almost all their income comes from dynamic news, but also because such a single map may be their most powerful structural defense against antitrust actions.
Unlike the two-way dating mechanism dominated by Facebook,
Twitter is built on a map of one-way attention.
. Theoretically, this should reduce the influence of atlas design problems. However, like any interest map based on social map, this method has the same defects. You may be interested in some of someone's interests, but not others. Twitter prefers a single Twitter account that focuses on niche markets. But most people don't choose to maintain multiple Twitter accounts in order to completely separate their favorite tweet topics.
As far as finding system defects is concerned, seeing who wants to destroy it is one of my favorite heuristic methods. Advanced social media users have long tried to avoid the problem of map design. Some users want to create a fake Instragram account or another Twitter account, in part to create an alternative map that is more suitable for a specific purpose. You can imagine a social architecture that does not require users to create multiple accounts to implement these strategies. But in this world where each social media account can only be associated with one identity, users have been locked in the map of one account.
App may have a clever way to help solve the problem of map design, that is
Eliminate the burden of users paying attention to accounts that they are no longer interested in.
. Just as our social graph will change in our lifetime, so will our online social graph. The friends we make in kindergarten are often different from those we make in primary school, high school, university and later.
Over time,
Social products with higher fidelity will automatically match our social maps when observing our interaction patterns.
. Imagine what would happen if Twitter or Instagram silently ignored accounts that you haven't interacted with for a while, accounts that have been dormant, and so on? Twitter and Facebook provide mute and other means to reduce what we see from others without canceling friends or attention, but it is a heavy workload. Frankly speaking, anyone who uses one of these methods looks like a coward.
Chat application, because it focuses on direct communication between two people or different groups, can naturally achieve this by pushing the thread with the latest news to the top of the application window. People who live far away from us just fall to the bottom of the screen.
LIFO method has always been a universal and quite effective correlation heuristic method.
There is another possible solution to the problem of atlas design, that is
Separating user's dynamic information from social graph.
. In my three series of articles about Tik Tok, I talked about the fundamental architectural differences between this application and most western social media. Tik Tok provides dynamic news and doesn't pay attention to any account. Instead, it did two things.
First, it will show you everything, then observe the reaction and try to understand what you are interested in. It tries to understand your taste and does it well.
Tik Tok constructed an interest map as an interest map.
Second, Tik Tok uses a two-stage screening process to check each candidate video. First, it uses the worst screening tool known to mankind to check these videos: a group of users with hundreds of people, mainly Generation Z, to check them. Well, this statement is really incorrect. Anyone can act as a test audience for the video. However, as it happens, Tik Tok's user base is relatively young, so most people in this group will be Generation Z. In addition, as we all know, in the animal kingdom, after hyenas and killer bumblebees, a group of Generation Z users who constantly complain to the old guy are the most terrible hunters. If these test audiences don't show any interest, the video will be thrown into the trash can in Tik Tok, and unless someone finds it directly on someone's profile, it will be buried from now on.
Then, it will use an algorithm to determine whether the video will interest each user according to their taste. Even if you don't pay attention to the creator of the video, if Tik Tok's algorithm thinks you will like this video, you will see it on your "recommendation page".
This is the fundamental problem of western social networks: to die, not to die.
Translator: Bossi