Product analysis of soul, Tantan and Jimu social apps!

In 1943, Abraham Maslow proposed the famous Maslow's hierarchy of needs theory, the basic content of which is to divide human needs into five types, from high to low: self-actualization needs, respect needs, social needs, safety needs and physiological needs.

We understand that social interaction is a basic need of every person and is more important than safety needs and physiological needs. After entering the Internet era, the advantage of efficient information transmission has expanded the way of making friends. QQ is the earliest and largest stranger social product in China. Then in the mobile Internet era, WeChat as a social tool has become a new way of life for people. In this process, challengers continue to emerge who want to get a share of the social dividends, but all ended in failure. WeChat's position is temporarily determined to be unshakable.

In recent years, various companies have begun to look for breakthroughs in the social segment. There are countless community social products based on topics and interests to appearance-based social products that directly match people based on their looks. (According to statistics, there were more than 150 social products on the domestic market in 2019)

So, if social interaction is a basic human need, are online stranger social products necessarily necessities? Do stranger social products directly change people’s lives, or do they improve people’s efficiency in making friends? Or is it just a demand created unilaterally based on Internet thinking?

The picture is provided exclusively by @pattlexue and cannot be used without permission

01 Overview

This article mainly analyzes three products, which can be summarized into two factions:

1. Introverted

Led by Soul , it advocates "soul social networking". Its core business is that users can express their true selves through the UGC community. The product's value proposition is to reduce people's sense of loneliness. The innovation is that it combines personality analysis with AI artificial intelligence algorithms. It is a social product that recommends people and things of interest to users in a "mind matching" way.

2. Extroverts

Led by Tantan and Jimu, they advocate "social networking based on appearance". Their core business is based on LBS location service, actively pushing other users' information to users on the platform (displaying user photos and other basic information in the form of cards). Users can choose to swipe right to indicate "like" or swipe left to indicate "dislike". When both users make the "like" choice at the same time, the two parties are "successfully matched" and can directly start a conversation. It is a social product based on "face matching".

Below, we will analyze the business logic, differentiation, and business model of the main functions of these three products, hoping to give you a better understanding of the three most popular stranger social apps on the market.

02 The underlying core business logic

This section will analyze how the three apps obtain the user's most basic portrait based on the normal process after registration:

1. soul

Step 1: Collect basic user profiles for the underlying logic of the matching algorithm to provide a more accurate matching experience, including:

Required information for registration

• Birthday, gender, region
• Primary personality test (6 questions)

(※ Figure ① is the "Soul Test" page, Figure ② is the basic personality test questions, Figure ③ is the matching results with other users, and Figure ④/⑤ are the analysis of the matching and mismatching results)

Optional information after registration

• Intermediate personality test questions (26 questions), advanced personality test questions (32 questions) and other interesting test questions (several types);
• Choose your own "attraction signature": Attraction signature means that users can label themselves, thereby attracting people with the same label as themselves, which helps improve the accuracy of the matching algorithm;
• Attraction signature types: including occupation, things you are doing, type of music you like to listen to, dramas you like to watch or are watching, games you are playing, type of books you are reading, and favorite authors;
• The latest version centrally manages gravity signatures, making it easier to locate gravity signature types. The content in Gravity Signature also has highlights. Among them, there are some relatively "new" words in "doing", such as singing and dancing, some relatively "fun" words, such as bank money counter, and some relatively "popular" words, such as straight men, straight women, programmers, and product dogs, which can better approach users and resonate with users. The benefits of centralized tag management are very obvious, but the disadvantages are that the granularity of tags is sometimes too coarse and sometimes too fine, which can only be applied to some users. In the future, the tag library should be continuously improved to meet the needs of a wider range of users.
• Optimization points: Add gravity signature search function, or optimize tag classification to meet the needs of fast search and screening. There are as many as 50 attractive signatures under the "Doing" type. It takes 5 clicks to switch the list by "Change Batch" to browse all of them, and the content is in disorder and not sorted alphabetically. Since the label content is updated infrequently and the frequency of user settings and updates is not high, we consider implementing it through the simplest interactive method, but this will also affect the user experience.

(※ Figures ① to ⑥ are the content of the "Doing", "Loving", "Listening", "Watching", "Playing" and "Reading" attraction signature types respectively)

Step 2: Recommend content based on user information and user portrait data generated by system analysis. The main forms of recommended content are in two dimensions:

• Content dimension: Users can pull out dynamic information of other users recommended by the system based on their preferences in the "square" (see Figure ① below)
• Human dimension: Users can actively trigger the "soul matching" function, and the system will instantly push online and highly matched users, without adding friends to start a conversation directly (see the following figure ②, ③, ④, ⑤)

Note:

• Figure ① is "Square": the name of the location that carries the UGC information flow in the Soul application;
• Figure 2 is "Planet": the first tab of the Soul App, which provides multiple touchpoint paths;
• Figure ③ is the "Soul Matching" interface, which is matching users;
• Figure 4 is the user chat interface. The user matched through "soul matching" will directly pop up the chat window. The main content of the interface includes the matching degree with the user, the other party's attraction signature, the same attraction signature of both parties, common points, such as age or city, and reference topics for icebreakers, such as "Talk to him about his favorite food", "Do you have any favorite sports? Tell him why you like them", but the actual use is not obvious and it is a bit abrupt in the first conversation between strangers;
• Figure ⑤ is the personal homepage, which mainly displays the number of days the user has registered, the number of dynamics posted, and the attractive signatures and dynamic content.

Step 3: Mining, screening, classification, analysis and calibration will also be carried out in subsequent user behaviors to improve a more comprehensive user profile.

2. Tantan

Step 1: Compared with Soul, Tantan’s approach is relatively primitive and relies on active user input. After registration, it will guide users to complete their personal information to facilitate optimized matching (most of the information is optional and skipping it will not affect the user’s use of the main functions).

(※ Figures ①, ②, and ③ are the Tantan profile editing interface)

(※The above picture is the information structure diagram of Tantan’s personal profile)

Step 2: Make personalized recommendations based on the underlying user portrait information data.

(※ Figure 1 is the homepage of Tantan, showing the user's card summary information; Figures 2, 3, and 4 are expanded to show more information about the user, which will clearly mark the common points between the two parties; Figure 5 is the user's personal homepage)

3. Jimu

Step 1: Compared with the above two methods, JiMu's approach can be described as simple and crude. It requires you to select one interest (and only one) when registering, which will serve as the basis for subsequent system recommendations.

(※The above picture is the structure of Jimu interest information)

Step 2: Discovery Filter: Select the distance range and interest range (all interests or the same interests as yours) for filtering.

( ※ Figures ① and ② are the Jimu discovery and filtering pages; Figure ③ is the homepage, showing the user's card information; Figure ④ is the user's personal homepage)

4. Summary

In summary, if the above three applications are ranked by the complexity of the business logic of their core functions, it is Soul>TanTan>JiMu, and the directness of the purpose of making friends is inversely proportional to the complexity of the business logic of their core functions, which is JiMu>TanTan>Soul. Part of the reason for this is related to the target user groups of the products themselves:

soul

Soul mainly targets young users in first- and second-tier cities, with 84% of them being 25 years old and below. Forty percent of users are students and prefer to communicate and share online. The gender ratio is relatively balanced (rare in dating apps). In the early days, Soul also targeted the market of overseas student users and understood their pain points - they needed a community that could relieve loneliness in a foreign country and share and communicate with domestic netizens. Within less than a year of the product launch, it supported registration and login for overseas users, and placed advertisements on YouTube and Instagram to attract overseas users. In China, it used Douban and Zhihu channels to attract young and energetic community enthusiasts.

It is also worth mentioning that the installation ratio (overlap) of Tantan by soul users is only 25.5%, which shows that soul has accurately found this segmented user group - they are more willing to express, share and communicate and are not in a hurry to find a partner through quick means.

Tantan

Tantan initially imitated Tinder, using simple interactive methods to improve the efficiency of users browsing other users' information. It gained a huge advantage in the early stage and seized market share. According to the official data provided by Tantan before it was acquired by Momo, (in June 2018), Tantan had 90 million registered mobile users, of which more than 75% were users born after 1990, with 6 million daily active users and a next-day retention rate of 75%. It has completed more than 3 billion user matches, and the ratio of male to female users is 6:4. By the end of 2019, the number of registered users exceeded 300 million, with young users aged 18-25 accounting for the highest proportion. In the App Store social and lifestyle categories, it ranked first in both download rankings and revenue rankings.

Nearly 42% of Tantan users are aged between 25 and 34, most of whom are white-collar workers with high incomes, and they attach more importance to the efficiency of direct matching.

Jimu

Jimu is a social platform for young people with a "trendy" and "cool" style. It advocates expanding social circles based on interests. Its main users are concentrated in first- and second-tier cities. Nearly 80% of them are born after 1995. They are the Z generation group who pursue music, art, and trendy culture. This can be confirmed by the following aspects of the product:

1) The interest items provided are more youthful, individual, alternative and niche, such as LiveHouse, modified cars, music production, release, quirks, letters, etc.;

2) The App UI is mainly in black and white, which is consistent with the young and fashionable visual concept conveyed by the official website brand.

(※The picture above is the official website of Tantan, which has not been updated for several years. The overall style and design are relatively "retro")

(※The above picture is the official website of Jimu. Click the picture to open a new page. The official website is in the form of a full video background. The different scenes in the video content highlight the core values ​​of the product "interest", "making friends" and "dating")

3) The average quality of user profile pictures is relatively good, close to "Internet pictures". If you scroll through them quickly, you will have a feeling that "they all look similar". The Internet pictures here are not derogatory, but refer to people who keep up with the trend, are good at dressing up and taking selfies.

If you try both Tantan and Jimu at the same time, the difference will be very obvious, and you may even wonder if there is really a "shill". The main reason is that the attributes of the user groups they are targeting are different. Secondly, the small user base of the platform is also a factor. However, this also raises another issue worthy of attention - the authenticity of user avatars. What should we do if there are too many "online pictures"?

03 Discuss issues regarding real people and the importance of certification

The importance of real users is ranked from high to low: Tantan>Jimu>Soul

1. soul

Since Soul focuses on spiritual social interaction, user avatars use a cartoon avatar mode similar to the early Tencent QQ, and do not support customized uploaded pictures. On the surface, it basically blocks the appearance factor, so the product logic does not emphasize the issue of real-name authentication, and it can only be determined by users' independent judgment.

Tantan and Jimu focus on social networking based on appearance, and the authenticity of the avatars or photos helps to improve the efficiency of matching.

2. Tantan

The solution adopted by Tantan is "face recognition technology comparison". The prerequisite for authentication is to upload a main profile picture (up to 6 profile pictures can be uploaded). The main profile picture must be a clear face photo for easy authentication.

During the function testing process, it was found that even if the authentication is unsuccessful, Tantan will retain the facial data of the person at the time of authentication. If a new avatar is added later, the system will automatically initiate a comparison to reduce user operations. (Since only the main avatar photo is compared, the avatars in the other five positions are not affected)

3. Jimu

Jimu requires users to manually fill in their real names and ID numbers, and then perform "facial feature biometrics" to complete real-name authentication. Its main purpose is to complete "real-name authentication", while Tantan's "real profile picture authentication" can prove that the profile picture uploaded by the user is his or her own photo.

4. Summary

In contrast, the positioning of Jimu’s “real-name authentication” function is unclear. It seems more like a function added for the sake of adding it, and it does not play much of a role. If it is to restrict user behavior, it is not a mandatory requirement.

The "real avatar" function of Tantan is more in line with the product positioning, which is designed to increase user exposure and matching efficiency. After passing the authentication, the user will have an obvious authentication mark on his body, which can provide effective evidence for other users to make judgments.

A more accurate statement is that Tantan's "real profile picture" authentication can only mean that there is a high probability or 99% probability that it is the real person, because stealing online pictures is very common in stranger dating apps and is often used in black industries such as fraud and diversion.

There are a large number of related advertising posts in the Tantan Baidu Post Bar, such as selling accounts (mainly accounts disguised as women, the more people "like" them, the more expensive they are), providing Moments materials (daily real life photos, voice packages and videos of the same person, which are more authentic), and the rampant real avatar "proxy authentication" services on the market. They will mark the key points of the faces in the photos, and then generate the blinking, opening the mouth, shaking the head and other action information required for positive live real person authentication. They can even perform 3D facial modeling for more demanding facial recognition. Therefore, for apps like Tantan, "real people" are particularly important. They must continuously optimize the authentication accuracy or increase the difficulty of authentication to raise the threshold for "proxy authentication", continue to crack down on related malicious users, check abnormal accounts, and maintain the "health" of the platform.

04 About the difference between communities

The importance of community building is ranked from high to low: Soul>Jimu>Tantan

1. soul

Community is the core functional business of Soul. What methods can be used to create a tree hole and spiritual garden for Soul users? This is what Soul has been pursuing and emphasizing. Let's see how they do it:

1) Identify the target user group

In the early days of advertising, we focused on Zhihu and Douban, two relatively literary and young communities. The advertising copy we produced was adapted to local conditions and in line with the culture of the corresponding community, and also in line with the product's consistent tone - "Follow your soul to find you". It conveyed to users in an emotional way and empathized with them. The users we attracted also recognized this value and hoped to share their joys, sorrows, anger and happiness here, share content that they dare not post on WeChat Moments, and even find their soulmate.

(※The above picture is the advertisement released by Soul on Douban)

2) Refined operations

Through reasonable operation means, users are encouraged to produce high-quality UGC content and active user atmosphere, thus improving the activity and stickiness of community users:

1. Through the operation background monitoring and screening of high-quality posts with high popularity in a short period of time, manually pin them to the top;
2. If the content is widely disseminated, after manual screening, an invitation for "high-quality content recommendation" will be sent to its users. After the user passes the official authorized invitation, he or she can get more traffic exposure and resource support;
3. Periodically release official topic activities (most users are well aware of traffic, and can get it the fastest by riding on the popularity. They quickly respond to official topic activities and publish dynamics with hot tags on a daily basis to gain exposure. These are the results of long-term education)

(※ Figure ① is the activity page released by the Soul official assistant; Figures ②, ③, and ④ are the topic details pages of the corresponding tag labels)

3) AI big data matching algorithm to distribute content

Different from Weibo, which mainly recommends top KOLs, Soul has thoroughly implemented the positioning of "personal content sharing community". Through a reasonable data distribution model, it makes user content distribution more equal and effective. Just like the brand slogan "Follow the soul to find you", Soul also does this. Users can easily find content and people that interest them in the "Square". This is the charm of the "personality analysis + AI" matching algorithm created by Soul.

4) Implement the official incubation plan to encourage users to continue to create high-quality content

The "SSR" incubation plan (Super Soul Real) was launched on November 25, 2019, aiming to encourage users to create more high-quality content, support users who have continuous creation in a certain vertical field, and create top KOLs in the user stratification. The support rewards include traffic, resource support, dedicated operation and data analysis support, and cash rewards.

Its main purpose is to enrich the content form of the community and plan the content production in six vertical fields to maximize and more comprehensively meet the needs of users with different interest groups.

(※The above picture shows the six vertical fields and the contents included in the "SSR" plan)

5) Create a good community environment to match and stabilize the overall tone of the product

By forming a content review team, we strictly control the content quality, and conduct 24-hour manual review on content (abnormalities) that cannot be determined by machine review to ensure content compliance.

But it also brings a problem. When the machine recognition is not particularly accurate, the user's post cannot be sent out in time. For products that focus on UGC content communities, the resulting bad experience is fatal. If the content posted by the user does not receive effective exposure to obtain clicks (reads) and feedback (likes and comments) within a short period of time, it will miss the content exposure point and will soon sink to the bottom of the sea, greatly undermining the user's desire and confidence to create.

Especially during the peak of the Spring Festival epidemic, actual experience shows that the review speed of posts involving pictures, audio and video is obviously too slow (even when posting ordinary landscape photos), which makes people wonder whether the system has performed machine judgment on the images, videos and audio voices in the content, or whether the number of posts is large during peak periods, the machine processing speed is slow, the judgment accuracy is low, and the content is piled up in the review queue, resulting in untimely manual review and long cycle. The product itself does not provide the post publishing status. Even if the review fails, the user cannot be informed of the status and reasons in time.

However, there are signs of such strict control over content:

From April to June 2019, the Cyberspace Administration of China launched a special rectification campaign on the content of instant messaging tools. Platforms that spread historical nihilism and obscene and pornographic content were punished to varying degrees, including interviews, removal from shelves, and shutdowns. Many platforms involving stranger social gameplay were also hit to varying degrees.

Soul is no exception. Two months after being removed from the shelves, it began to internally check all possible risks and formulated a strict content review and monitoring plan. While increasing its control over content, it also affected the experience of some users.

After being re-listed in August and September last year, a large-scale promotion was carried out to attract new users. Due to the surge in the number of registered users, the average quality of the content declined, which led to the loss of some old users. They generally felt that "Soul is not like the previous one", which undoubtedly put forward greater requirements for the content operation of the community:

1) Reduce obvious online image tendencies, meaningless content (mainly manifested in vanity, attention-seeking content), and "inexplicably" popular content; (For example, leg photos. Girls posting such content on Soul are very popular with boys, and the number of likes will be inexplicably high. Other users will follow suit, leading to a vicious cycle, and users who create content seriously will be disappointed)

(Special note: The pictures posted by users in the following cases are the original effects presented when they were posted by the users. I did not do any secondary coding. However, in order to protect the relevant personnel, special processing was done on the avatars, names and some texts to ensure that they cannot be found through search on the platform to avoid harassment or harm to them.

*The following pictures are only used for product case analysis. If they cause you any inconvenience, please contact me to delete them. Thank you for your understanding. ）

Figure 1:

This is one of the "inexplicable" popular contents. After analyzing the published pictures and texts, it is difficult to understand why the number of likes can reach 300+. The original photo does not show the face. After communicating with the user, it was confirmed that it was not the paintings exhibited on the wall that resonated with the audience. Finally, it was speculated that it might be due to the large number of followers of the user.

Figure 2:

It falls into the category of "meaningless content". The main reasons for its popularity are: JK elements and exposed legs.

In the early stage of using the App, you can often see popular content related to "showing legs" and "showing body shape". The main feature is that users can attract a lot of clicks without showing their facial information (it can be ruled out that it is due to user preference. During the experience, there is no behavioral data on following or liking and commenting on related types of tag topics and dynamics)

Figure ③ and the following figures ②③④:

It also falls into the category of "meaningless content". It can gain a lot of attention and clicks in a short period of time through obviously tempting pictures or texts. Even the selfie picture in Figure ④ below has been "hashed" beyond recognition by the original user. (23 likes in 2 minutes)

Figure ①:

is a typical example. A simple "flirting" sentence and a picture that can prove that the user is a female user can attract nearly 900 users to like it, which is basically accompanied by "private message" bombardment. The situation of 99+ logos on the chat icon is something that most female users of Soul can encounter more or less, whether intentionally or unintentionally.

Why is it emphasized that any picture that shows that the user is a female can attract attention?

A male user once shared an experience where after posting a photo that had a small chance of being mistaken for a girl’s, he received private messages from some male users. The screenshots are as follows. We will analyze this issue at the end.

(※Prerequisite: The male user’s dynamics contain photos that clearly prove his gender; and it can be clearly seen from the above chat records that some men’s initiative does not prove that they are interested in you after getting to know you, but simply because you are a rare "woman")

2) To enhance the accuracy and precision of the recommendation algorithm, users need to be stratified in more and finer dimensions, with the aim of reducing the communication between users at different levels, so that they coexist but are not mutually exclusive. This is also the main reason why some users feel that "Soul seems to have changed";

3) Enhance the experience of the "not interested" function , analyze the main reasons for users' disinterest, and trace the root causes of the reasons, mainly the content, text, style, tags, topics, or the author of the picture. If we only use the labels given by the system to the content, it is easy to be inconsistent with the user's subjective feelings. We can reduce and control the sample size of relevant labels, and then make preference judgments and calibrations on users' subsequent behaviors;

Guide users to train the machine so that it can better understand users and strengthen the popularization of recommendation or "not interested" functions. Popularization requires that the function be easy to use, clear and accurate in interaction.

2. Tantan

Since Tantan’s community module has undergone a major update recently, the following will briefly list the previous and next versions for comparative analysis:

Before the revision:

Before the revision, Tantan had not only matching and chatting functions, but also a "Friends Circle" function. The function scope is as the name suggests, which is roughly the same as WeChat Friends Circle and supports commenting and liking the dynamics of successfully matched friends.

A relatively small difference is that the dynamic "private comment" function is deliberately emphasized. This is done to reduce users' vigilance, meaning "only visible to the other party". After commenting, the comment content and summary or picture will be forwarded to the other party's chat window. This is different from WeChat. Tantan hopes that the final process will return to the closed loop of instant chat to increase the willingness of both parties to communicate.

(※ Figures ① and ② are the Tantan “Friends Circle” before the revision; Figure ③ is the chat window interface)

After the revision:

The updated points after the revision are as follows:

• New "Nearby" module: You can view the dynamics recommended by people nearby (also known as "nearby circle of friends")
• New "Personal Homepage" module: You can learn about the user's posted updates, fans, followers, likes and other information through the user's personal homepage
• New "Follow" module: In addition to viewing the dynamics of successfully matched friends, you can also track the dynamic messages of other users by following them one-way. If both parties follow each other, they can also reach a matching relationship and have a conversation

(※ Figure 1 is the "Follow" page of Tantan after the revision; Figure 2 is the user's personal homepage; Figure 3 is the "Nearby" page)

The community attributes have been enhanced after the revision. The newly added "Nearby" module and "Follow" function have also increased the paths and segmented scenarios of contact between users, which can promote users' desire to produce content, increase personal dynamic exposure and extend user usage time. However, the quality of content in the early stage was uneven. Through observation, it can be concluded that the recommendation mechanism is relatively random, and the matching algorithm needs to be strengthened and optimized to make recommendations in the direction of user preference. In addition, there is a rather strange problem after the update. The "private comments" emphasized in the original version have become public comments, and deletion is not supported. The logic before and after the iterative function contradicts each other.

3. Jimu

Jimu Community is mainly divided into four small modules, namely:

• Follow: Display the dynamic content of the users you have followed (including yourself)
• Recommendation: Display user dynamic content recommended by the system
• Friends: Display dynamic content of friends who have "successfully matched"
• Region: You can switch regions at will to display user dynamic content recommended by different regions

The content recommended by Jimu Community is obviously more inclusive and "open", and there will be many contents that tend to skirt the line (such as explicit copywriting, overly revealing pictures and suggestive photo poses) on the hot recommendations. It may attract bold, open-minded and showy young user groups, but it is also more likely to face possible supervision and is not a sustainable, virtuous content community.

Before analyzing how Jimu formed its current product tonality, let’s first understand the entrepreneurial journey of Jimu’s founder Cai Di:

• In May 2015, he launched his first entrepreneurial project, SeSe, in Shenzhen. The project mainly solved the problem of social isolation among strangers in nightclubs. By implanting LBS strong positioning devices between tables in nightclubs, users could obtain social information of strangers in nearby seats by scanning QR codes, thus improving the efficiency of ice-breaking matching between tables and establishing effective interactive relationships. Later, the project was not continued due to funding problems.
• After the failure of the first project, Cai Di participated in and researched a large number of offline music festivals, LiveHouse and other youth cultural activities. After in-depth contact with users, he directed the core target group of the second entrepreneurial project to the youngest and most trendy group.
• In August 2016, Jimu iOS official version was launched. The goal set was to create a social app in the field of youth culture and establish an O2O social service system. In the early days, users were matched online based on common interests. After a successful match, the system recommended related activities, and then completed the common goal offline as a closed loop, in order to strengthen social relationships and retain users. The offline part is achieved through cooperation with brands, hosting trendy parties or youth activities, for example, establishing cooperative relationships with well-known domestic music labels (such as Strawberry Music Festival), and then expanding to other event organizers. At most, we hold 2,000 to 3,000 events a year and capture a lot of targeted users offline.

Therefore, it is not difficult to understand why Jimu has formed a temperament that is completely different from other products, as can be seen from the fact that it lists various niche interests as interest items. In the early days, we used very vertical media publicity, held trendy culture and street events to attract users with similar values, connected industry resources and interest circles, and then self-propagated among users to precipitate mid- and top-quality seed users.

(※ Figure ① is the community "recommendation" page; Figure ②③ are the community "region" page; Figure ③ is the dynamic details page; Figure ④ is the personal homepage)

However, there is an obvious problem in the communities of Tantan and Jimu: the community interactivity is poor. The top recommended contents in the communities (especially Jimu) are mostly those focused on gaining recognition from others and satisfying personal vanity, desire to show off, and attracting attention. Most of the comments are one-way and lack interactivity. In addition, the recommended contents are basically irrelevant to the interests of users, and the recommendation mechanism has more random elements.

Since being acquired by Inke in October last year, there has been no obvious improvement in the relevant matching or recommendation functions, strategies and algorithms.

If a community is merely used as a way to gain additional user exposure but lacks key interactions, it will not be enough to establish an effective closed loop and form an independent functional product on the platform. It is difficult to maintain a long-term relationship between strangers if they rely solely on their ability to "find topics/have awkward conversations", which results in a lot of "one-time" waste . A good community should be more like a place that can carry users' interests, life dynamics, and personal business cards (multi-dimensional display), providing more guiding topics for users to break the ice. (For example, I matched with a user through Soul's "Soul Matching", and after clicking on her homepage, I found that she had recently watched the movie The Tuner, which I also happened to have watched. We then had a common topic and could start a conversation about it)

4. Summary

Social networking among strangers cannot escape the vicious circle of vulgarity. Under the temptation of traffic and commercial interests, it is easy to play tricks. Especially in 2019, many apps were removed from the shelves or interviewed due to sensitive issues of community management.

Recently, a vicious incident was revealed in the industry. In order to crack down on the competitor uki, Soul's operating partner Li deliberately set up a scheme to create illegal content, and used this to report and suppress his peers, resulting in uki being removed from the shelves for nearly four months, the number of registered users of the product plummeted, and the company's operations were suspended.

It can be seen that the stranger social industry is highly competitive and everyone is trying their best to seize market share. According to relevant reports, a former employee of Soul revealed that the idea of ​​maliciously reporting competitors actually came into being in 2018, and this vicious incident occurred after Soul was removed from the shelves in June 2019. The company actually implemented it because they were worried that competitors would take the opportunity to catch up and snatch traffic. It also highlights the anxiety of various companies in the stranger social track. The only primary goal is to "survive". Sometimes some commercial value has to be sacrificed, and products must adhere to compliance to continue to create value for users and profits for the company.

05 Comparison on Commercialization

At the current stage, the three products are mainly monetized through membership subscriptions and the sale of value-added services such as virtual currency.

Tantan announced in March this year that it was conducting internal testing of its live streaming function, and will open it for large-scale testing soon (April 15), and it will not be long before the full launch. In addition, Jimu also reported news related to live streaming in early April. This is not just a competition between Tantan and Jimu, but also a contest between Momo and Inke.

Below is a brief comparison of the paid features and prices of the three products:

1. soul

【Super Star】Main privileges include checking who has visited your homepage, accelerating chat level, publishing 5-minute audio and video updates, etc.

（※soul-Super Star Price List）

[Virtual Currency] It can be used to purchase privileges for certain functions and virtual goods. It has “Gifts” comparable to QQ, which can be used to give to friends. It has “Super Cute Face Pinching” comparable to QQ Show. In the future, it will probably expand in the direction of QQ’s “Centimeter Show”.

At present, there is no obvious sign of other types of monetization models for Soul. The only attempts being explored are some extensions of the current model. For example, the flagship "Soul Matching" function is undergoing a feasibility test of charging a fee. By gradually reducing the daily match limit, improving matching accuracy and efficiency, and then slowly transitioning to payment, this method is more suitable for the transition from a long-term free function to a paid function. The final replacement speed still depends on user response data. Users will only pay for it if the experience is significantly different from that in the free stage.

(※Transition process from free to paid)

2. Tantan

[VIP] Mainly used for regretting wrong swipes, changing location at will, unlimited "likes" (120 likes per day for ordinary users)

（※Tantan-VIP Member Price List）

[See who likes me] After activation, you can see which users "like" you

(※Tantan-"See who likes me" privilege function price list)

[Flash Chat Matching] There are 3 masked online real-time chat opportunities every day. You can unlock the other party's avatar after 20 rounds of conversation, or you can unlock it directly by purchasing the "Sneak Peek Privilege".

Online real-time matching chat makes up for the low timeliness and high latency of "card matching". Flash chat can only be conducted with users who have activated this function. Although the user range is narrowed, the willingness of both parties to "match" is highly similar.

A masking mechanism is set up for the first 20 rounds after the start of flash chat (meaning users cannot see the other party's real profile picture), which means "soul socializing" first, and completing the first round of matching test by chatting about world views and interests. Unfortunately, the reality is very bleak, and it is easy to unlock by swiping each other's screens. In the end, everyone is looking for "what does the person you are chatting with look like?" Although the flash chat function seems to be more valuable in the sliding card, it is out of place in the functional matrix of the product if it has to put on a "mask", which is a bit like going backwards.

I couldn't help but wonder whether there was a charge point for masking the face first or a demand point for masking the face first.

(※Tantan-"Flash Chat Pairing" function price list)

[Super Exposure] This function can increase exposure within 30 minutes (10 times the number of people are preferred) [Flash chat peek] You can directly unlock the other person's avatar when pairing with flash chat

[TanTanCoin] TanTan's virtual currency can be used to purchase super exposure and flash chat peek

3. Jimu

【VIP】It is mainly used for slipping mistakes, changing positioning at will, unlimited "likes" times, priority recommendations (increasing exposure), and advanced filtering (filter options include the number of user photos, tags and constellations).

(※Jimu-VIP price list)

[See who likes me] After activation, you can check which users "like" you.

(※Jimu-"See Who Likes Me" Privileged Function Price List)

After Jimu was acquired by Yingke at the end of last year, it began to try monetization plans in December. The VIP value-added service was first activated. The "See Who Likes Me" service is still free. In January this year, the scope of testing has been gradually increased. "See Who Likes Me" requires payment to view it. Judging from the feedback from Tieba, it will be a big blow to some ordinary users who have no payment habits and enjoy "free travel" from the beginning. For them, they have undoubtedly lost their original important advantages, the user value has dropped sharply, and the price of paid functions is relatively high.

4. Summary

Soul tends to package function payment mode, and members have additional discounts when purchasing other virtual props, while Tantan and Jimu adopt multi-point payment mode to facilitate the subsequent expansion of new monetization functions.

All three models have the function of "See Who Likes Me". For Soul users, visiting the homepage once and twice is not worth mentioning. If you check it repeatedly, or even "care" to a certain level every day, it is enough to stimulate users to pay to reveal the answer. Of course, it depends more on the user's curiosity and whether they care.

(※The picture above is the soul-"Who has seen me" list page)

For Tantan and Jimu users, "See Who Likes Me" is a "plug-in" function, which is equivalent to mastering the "activity" (you can select the one you like from the list of likes you to complete the match). Naturally, the price is also relatively expensive. Tantan's "See Who Likes Me" is 60 yuan/month, while Jimu is 58 yuan/month.

June 2019 data:

【Tantan】

41.7% of users are between 25 and 34 years old, 41.2% of users are corporate white-collar workers, 40% of users have incomes between 5-10k, and 9.2% of users have incomes between 10-20k.

【Abstract】

65.8% of users are between 18 and 24 years old, 43.9% are students, 43.3% have less than 3k income, and 10.5% have between 3-5k income.

I did not consider the reference targets of the price carefully, and I did not understand my own user attributes clearly. More than 50% of Tantan's users are white-collar workers, office workers and high-income people, and have sufficient payment ability.

Jimu's main user group is students or newcomers in the workplace who have just entered the society to work. For them, Bao Ji's Tencent Video membership and a McDonald's giant may be "more popular".

5. Expand

In the past two years, the live broadcast industry has entered a mature period and has shown a downward trend. The trend is sluggish and has a tired period. In order to stimulate the business ecosystem, Momo and Yingke have begun to plan for the interest dating field with relatively high matching efficiency for strangers, and also have the intention to bring live broadcast into the stranger dating field.

In February 2018, Momo completed the acquisition of Tantan for US$760 million. In July 2019, Yingke completed the acquisition of Jimu for US$85 million.

If you only watch the live broadcast business, Tantan will have more advantages with Momo endorsement. There is also news that Tantan's current live broadcast team is Momo's original live broadcast team. Momo has been very steadily in recent years and has always been at the top of the live broadcast industry. Let's take a look at a set of data first:

Momo's operating profit in Q4 2019 was 1.4585 billion yuan, compared with 1.0926 billion yuan in the same period in 2018. Tantan's operating loss in Q4 2019 was 2027 million yuan, and its operating loss in the same period in 2018 was 219.9 million yuan.

Tantan did not bring considerable benefits to Momo after being acquired. Moreover, in April 2019, Tantan was involved in the "pornography-related" storm, and the product was removed from the shelves and rectified on the entire platform. It was only after two months that it was re-launched. It has not yet recovered to its scale before it was removed.

In 2018, Momo's monthly active users increased by only about 10% from 103 million in Q1 to 113 million in Q4.

In 2019, Momo's monthly active users in the first three quarters were 114.4 million, 113.5 million and 114 million, respectively, and 114.5 million in December, an increase of only 2 million from 113.3 million in the same period in 2018, and even declined slightly in the second quarter.

At this time, Momo's business model reached its peak and user growth stagnated. It is urgent to find the next large-scale transformation scenario. The hope may be based on Tantan, which was acquired for US$760 million in 2018.

There are signs that Tantan's users have a considerable scale of paid users. According to the official 2019 financial report, Tantan paid users in Q4 2019 had 4.5 million, and the overlap between Tantan and Momo users was low (of course, it was also one of the reasons why they could be acquired). At this time, the goal of launching the live broadcast business was to complete the exploration of Tantan's existing users, thereby driving Momo's revenue growth. Tantan also needs new elemental stimulation at the product and business level to accelerate the commercialization process.

By introducing live broadcasts and "nearby people" to enrich the form and business model of Tantan's products, in the future live broadcast operation strategy, Tantan should be more inclined to the tail live broadcast users, focusing on nearby anchor pages, emphasizing the attributes of live broadcasting and making friends with strangers, which will be more in line with users' demands and expectations for the original product.

(※Picture ① is a live broadcast-recommended page;Picture ② is a live broadcast-near page;Picture ③ is a discovery-near page, publish-dynamic/live broadcast function;Picture ④ is a discovery-near page-recent live broadcast recommendation)

Will the two old elements of live broadcast and "sliding card dating" collide to create new sparks, and whether Tantan's new model can lead the next trend, it is still unknown, and what can be expected is:

• Recent iterative functions will extend the average usage time of Tantan users and increase activity
• Expand user payment channels (live broadcast rewards)
• Activate the existing users of live broadcast (viewer/anchored)
• Drive high-quality central users to become anchors

06 Comparisons about other aspects

1. Comparison of App PUSH notification policies

On the evening of March 27 this year, I launched a vote on Soul on Soul's Soul app notification. 115 people participated in the vote. The results showed that the proportion of application notifications closed was 76.5%, and nearly 80% of users would choose to block application notifications.

This is just a choice for sampling and statistics of soul users. You should know that the soul is relatively open to the permissions provided to users by application notifications. The message reception mode is as follows:

• Notification displays message details
• Receive all message reminders (operation notifications include regular official topics and popular recommended topics, etc.)
• Only receive reminders of square and chat messages (equivalent to choosing to close App operation notifications)
• Only receive chat message reminders
• *And when the app is not opened, messages of unopened types will no longer be received

(※The above picture is the soul-notification setting interface)

Tantan

• Show message preview
• Friends circle get "likes" and notify with private messages

Accumulation

• Message Details
• Added attention notice
• Comment Notification
• Who likes me to notify

(※The above picture is the "operation activation category" notification pushed by Tantan and Jimu)

summary

Applying PUSH notifications is a very important and most common operation strategy in apps to activate and retain users. If used well, it will be a pure profit and zero cost transaction. If used poorly, it will remind users to uninstall the app.

Soul is relatively restrained in this regard and chooses to hand over the initiative to users. It provides a customized combination of notification messages, which is set by users according to their actual needs. It not only ensures a good user experience, but also forms a significant experience difference with other apps, thereby enhancing favorability and loyalty.

Tantan is layered according to the user's frequency of use, and then adopts a push plan of different frequencies: if there are more open frequency on that day, no operation notification will be pushed; if the App is not opened for one to two days or more, an activation wake-up notification will be pushed every two hours between 8 am and 10 pm (there will be up to 8 pm in a day) . The main type of notification is recommended for users nearby, and the recommended users will be ranked first in the card list after opening.

The push strategy adopted by Jitu will be relatively random, and the content mainly belongs to the induced click category . After the click copy is opened, there will be no corresponding landing page.

Each company has adopted different notification operation strategies, based on its understanding of users under the platform. Soul is more like a friend, taking the emotional route, while Tantan is putting it to death and then reborn. The worst is to uninstall it. The best situation is to wake up successfully. No matter how bad it is to close the application notification and continue to remain silent. You will definitely turn it on when needed.

2. Regarding the issue of transferring users to WeChat

You often encounter this situation in the stranger social app. Many people will ask for WeChat if they are less than a few words to talk. We can briefly analyze the possible reasons for this approach:

Excluding other special circumstances such as fraud and traffic diversion, the seeker is not a loyal user of the platform, but only uses the platform as a channel or tool to expand friends. This type of person attaches more importance to social efficiency. The purpose is to complete the closed loop of "chasing up" by themselves, that is, asking for success of WeChat as a result, which is an inertial action. This practice will be rejected in most cases. In a few cases, even after exchanging WeChat, it will only become a silent "friend" in the WeChat address book.

In addition, it is because the relationship between friends has progressed to the point where the two parties are willing to reach a consensus that reality or close to reality relationships, so we will add each other. WeChat is a social tool for acquaintances in the minds of some users, and there are certain thresholds. This situation is generally the fact that strangers are most worried about and have to accept.

All the platform can do is to work hard to increase the cost of being replaced, improve the user experience of the product itself, improve user value and differentiated experiences different from WeChat. The following are the responses and strategies made by the following three apps to address such situations:

Soul mainly attracts users to retain through a unique community atmosphere and a combination of friends’ chat levels.

Chat Level: As the number of chat rounds on both sides increases, the love in the upper right corner of the chat interface will accumulate little by little. 1 letter can be lit for every 3 hearts, and so on. lit soulmate will reach the full level, and the more letters you will need to have a more chat round. (Blue will be displayed between the same sex, pink will be displayed between the opposite sex)

soulmate space: After lighting up soulmate, you can choose to establish a soulmate relationship, and you can also have a soulmate space exclusive to two people and publish news belonging to both parties.

(※Pictures ① and ② are the soul chat interface; Picture ③ is the personal homepage, providing a soulmate space entrance; Picture ④ is the soulmate space; Picture ⑤ is the soulmate space dynamics recommended by "Plaza")

Tantan and Jimu are different. Most users want to find a partner through the app or broaden their social circle to get to know more people. If the relationship between the two parties seeks further development, communicating within the Tantan or Jimu App is definitely not the best choice. Instead, they will choose WeChat that is more like an app that can carry acquaintances. If both parties develop into lover relationships, they are likely to become lost users.

In addition, the main reason why Tantan and Jimu can retain users is that they provide close to infinite possibilities for relationship matching, and users will have matching relationships. How to maintain users' enthusiasm and reduce frustration is a special issue that needs to be considered.

3. Regarding the issue of users being harassed

Tantan and Jimu are born with high thresholds. Only when both parties match can they have a conversation. Soul's situation is also introduced above. Female users are extremely susceptible to harassment in apps where private messages are basically no threshold. Soul adopts a "prevention of excessive disturbance" mechanism. When the number of private chat users reaches the upper limit on the day (estimated to be 99), other new private chat users need to give gifts to open the chat channel.

Not only does it "concert" raise the threshold for anti-harassment for users, it also naturally promotes the transaction of virtual goods. Users will receive notifications for gifts and can be displayed publicly on the "gift wall" of their personal homepage.

However, I am basically unaware of this "concerned" mechanism (unless others inform you or know it yourself), and this is somewhat opaque. If the "anti-over-interference" mechanism is triggered, the parties have a reminder sign, and they can even choose to turn off "anti-over-interference" independently, which will be more balanced in terms of user experience and business value.

Of course, this depends entirely on decision-making tendencies. In a certain situation, some mechanisms operate silently without disturbing users as the starting point, which is also a good user experience.

(※The picture above is the soul-chat interface-"Anti-overdistance" pop-up window)

In terms of anti-harassment, "blocking" can only be a temporary solution, and it is just a sub-plan in the combination of the problem. If we look at it from a long-term sustainability perspective, establish a user credit system and strictly implement reward and punishment measures, it can greatly restrict user behavior and better manage users in a layered manner, and it also responds to the previous Soul mentioned in the promotion copy:

Your soul cannot be too boring, otherwise we recommend you not to come. If you have a bad view, we do not welcome you. If your purpose is not simple, we forbid you.

Although the reporting function is the most basic and conventional function of the app, on Tantan, the execution level will be more convincing. Once the user is discovered and confirmed to have harassment, "hooking" and other behaviors, they will face the result of direct ban.

This also indirectly confirms that female users are protected and sought after in the stranger social app, and are at the top of the community food chain. A good-looking girl can gain nearly 10,000 likes in a day, and hundreds or thousands of ordinary people are possible. A few days ago, I learned from a user that I had registered for two years, but I played very little, and 260,000 people liked it . Male users were proactive and active, female users were relatively passive, and a mediocre male user may not have received a few likes for several months.

Tantan founder Wang Yu once shared a data:

In Tantan, male users will like 60% of female users, while female users will only like 6% of male users . There is also a very high matching success rate for female users, with a 60% success rate. In our data, the least popular girls are also more popular than the most popular boys.

Of course, it is not ruled out that some male users do not have filtering conditions, while female users will first carefully check the photos and other information in the information, such as careers, hobbies, and views on certain things (values), and then make a decision after thinking about it. A large number of "no thoughtful" "likes" will gradually reduce the expectations of matching relationships for female users, resulting in female users only contributing a very small amount of "likes".

Even in the most special situation of the peak of the domestic epidemic in February this year, Tantan's per capita usage time increased by more than 30%, the number of pairings and chat messages in the app increased by 20%, and the proportion of female users who "like" men who swipe right has only increased from 6% to 9.8%. (Breaking the highest record in the six years since Tantan's establishment)

(※Picture ① is a Jimu user who said, "I didn't care much about this software, and it suddenly exceeded 10,000 yuan"; picture ② is a female user who has registered for two years but is not active, and has gained 26W+ "likes"; picture ③ is an ordinary female user of Tantan, who has gained nearly 500 "likes" a day)

4. Conclusion

Therefore, when it comes to the issue of the differences in social initiative between men and women, it is difficult to reverse the balance even if the whole society is used to mobilize the strength of the whole society. I think this is why the social needs of strangers have always existed.

Final Thoughts

Thanks for reading, I hope it will be helpful or inspiring for you. I am still not satisfied, and I hope I can talk about a certain app in the next article. If you have any ideas, suggestions, or opinions, please leave a message or contact us. Goodbye.

Author: Half a Year PM

Source: Half a Year old PM