In 5 months, DAU increased 9 times and exceeded 20 million daily active users. What did Huoshan Video do?

Today, taking Volcano as an example, let’s take a look at what Volcano has done in product and operational promotion in the past five months, as well as the corresponding data changes.

Peter Thiel, the author of "Zero to One", was once an entrepreneur and is now an investor . His most important point in the book "Zero to One" is that only failed companies compete, and successful companies only monopolize. This also indirectly illustrates that achieving user growth and improving competitiveness are what all companies are striving to do, and making users addicted is the best way to achieve the above goals; the love of hundreds of millions of users for the product has become a guarantee of unicorn valuation. In the industrial age, the relationship between users and products is like an arranged marriage. What products to buy are determined by manufacturers and channels , and you have no say over whether you like them or not. In the Internet age, the relationship between users and products is completely free love. Users have great autonomy to love or not love someone. Can it become a billion-dollar unicorn? In fact, it was destined from the beginning that "habit testing, addiction model, and core behavior quantification" are all necessary. You have no choice but to pass the level!

1. The core secret of building a $1 billion unicorn: Habit test of "core behavior"

Habit Testing works perfectly with the Hook Model. By using Habit Testing, you can better answer the following three questions:

1. Who are your deep users?
2. Which part of the product can form user habits?
3. Why do these parts form user habits?

The prerequisite for Habit Testing is to have a working product. Of course, before launching a product (even a minimum viable product), it is best to complete the business model hypothesis and explore how your product can stimulate user desire. Once the product is online, you must start looking at the data. Don’t monitor all the data, just focus on the user’s interaction nodes with the product and the user’s access path. Let’s take a look at the three steps of Habit Testing:

Step 1 – Identification: Who are our heavy users?

Define the characteristics of a heavy user: Ask yourself how often users would use your product if the bugs were fixed and the product was polished. Be honest with yourself. If your product is a social app like Instagram, you can expect loyal users to use the app several times a day. But if you are building a movie recommendation site, you'd better not expect users to visit more than three times a week.

Data analysis : Once you know how often your product should be used, you can look at the data to understand how it compares to expectations and create a cohort analysis (a method of analyzing the behavior of people with the same characteristics over a certain period of time) as a benchmark for measuring the effectiveness of subsequent product iterations. When defining frequency of use, base it on reasonable guesses and don’t be blindly optimistic by only considering die-hard users. The more frequently users use your product, the more likely they are to form a habit. A good way to measure whether users will actually use your product is to observe whether your team uses it. Take Twitter, for example. The social media giant was born out of Odeo, a company founded by Biz Stone and Jack Dorsey, who knew Twitter had potential because their engineers couldn't stop using it.

In any case, you will hopefully find users who interact with your product frequently, which we call "deep users."

Step 2 - Organize: Deepen the steps that users take to use your product

Sort out the steps that deep users take to use your product so that you can understand what attracts them; First, how do you know if you have enough deep users? A safe baseline is to have about 5% active users; remember this number, you need to have a higher active user ratio if you want your product to survive. If less than 5% of users find your product useful, you're in trouble. If you already have this 5%, you can start looking for habit paths, which are a series of similar behaviors shared by most loyal users; not every user will interact with the product in the same way. Everyone has a unique data "fingerprint" that can show the user's usage patterns, which in turn can help you discover habit paths.

Back to Twitter, the social app found that once new users follow enough other users, the chances of more users using the product increase. Here’s how to determine which steps in the habit path are critical to cultivating deep users:

1. Create a hypothesis: Which links will make users start to become fans? While this may seem like assuming causality, it’s the best you can do when you’re in the early, chaotic days of a new product.
2. Talk to users: Understand how and why they use the product. For example, you could ask them to show you the steps that attracted them.

Step 3 - Correction: Guide new users to use the custom path we set

Now that we have our hypotheticals out of the way, it’s time to get back to reality. We have to think about learning and guide new users along the habitual paths we set. For example, Twitter suggests people that new users can follow during the onboarding process in order to reinforce habit paths. Once you determine that a certain path can promote user attention and conversion, you will also want to do something similar.

Habit testing is an ongoing process that applies to every new version release. Continuously comparing the operating behaviors of new users and deep users can guide how the product should evolve and further cultivate habits. If a product cannot establish enough user habits, it will fail. And if no habit is formed, it will be forgotten by users in the rapidly changing trend.

2. The core secret of creating a billion- dollar unicorn: the "TBRR" addiction model of "core behavior"

How to make our products more "sticky"? Taking the TBRR model's research on "addictive" behavior as a thinking framework, addictive behavior follows four stages: Trigger-Behavior-Reward-Repeat. In the Trigger stage, in order to motivate users to use the product, you can use "physiological attraction". When users are motivated to try the product, the user value experience is improved by addressing top-level needs, thereby achieving a closed loop with a high probability of realizing the secondary scenario loop. Take "smoking addiction" as an example. I heard that smoking can relax my nerves (Curious-Trigger), so I tried it out of curiosity (Behavior). During the smoking process, I felt really good (Reward). When I was under great pressure, my brain automatically built a loop, and smoking can relieve the current situation (Repeat).

Simply speaking of "smoking addiction", this model is just an explanatory tool and has little practical significance. However, if it is projected onto Internet products, it provides some thoughts on how we can open the door to product "stickiness".

1. Trigger-Triggering and awakening of the scene

According to Maslow's theory, needs are divided into five categories: physiological needs, safety needs, social needs (Love and belonging), esteem needs, and self-actualization needs, arranged from lower to higher levels, as shown in the figure:

(1) Underlying needs – the most direct trigger point of the scenario

It can be seen from the figure that the lower the level of demand, the larger the user base, that is, the lower the level of demand is often the high-frequency area of ​​the product trigger source; taking application promotion as an example, when promoting DateAi, a photo of a long-haired beauty was selected as the trigger source of demand to directly stimulate the physiological level of men (the main paying objects of this type of social software are men) and achieve the effect of scene triggering.

(2) Top-level needs generate scenario awakening through good value experience

From the above, we can see that the higher the level of demand, the greater the user stickiness. This is mainly because the high-level demand is actually the concrete embodiment of the user value experience. For example, although NetEase Cloud Music is not as large as Kugou and QQ Music, it is unmatched by any other music software in terms of user reputation because it has opened up the top-level demand level of comfort, sociality, and style, as well as listening to songs and reading comments, and feeling the emotions conveyed by the music. As a user, the most direct love for the product is to treat using the product as eating, which has evolved into a habit, just like when working, you must listen to NetEase Cloud Music's pure music to improve scene concentration.

The good value experience generated by the top-level needs will trigger a secondary scene awakening. In addition to the perceptual experience of using a product and the behavioral experience during use, the most important thing is the value experience at the final reflective level. That is, the product must be impressive, and users will immediately think of you in a certain situation, at a certain time and at a certain moment! Establish a scene loop.

Take an experience of taking Didi Taxi as an example. Some time ago, I went to another place. As soon as I got off the bus station, many taxi drivers came up to me and asked, "Where are you going, young man?" I was not very familiar with the surroundings in a foreign place. In addition, it was getting dark, so I had a stronger need to take a taxi. However, I was worried that the driver would deliberately rip me off because I was from another place. In comparison, Didi 's private car service seemed to be more standardized (the value experience at the brand level triggered the scene loop of my using Didi Taxi). So I placed an order with WeChat , but it was inexplicably deducted from my order (this inexplicable reason shook my trust in the Didi brand, and I even planned to actively cut off the scene loop next time). Because I was in a hurry, I called again, and this time it was successful. When I returned to my place, I contacted Didi customer service. It turned out that I had an unpaid online order before (there was no money in the bank card bound at the time, so I had no choice but to use cash). In order to make up for my loss, Didi refunded the deducted money to me (I found the inexplicable reason, and the customer service also made up for my loss. The shaken brand was consolidated again, and the good customer service experience made me trust Didi more). As a result, the probability of the second scene being evoked has been greatly improved, as shown in the figure below.

2. Behavior – Business points about behavioral needs, don’t let a behavior vacuum appear

(1) Don’t let bugs in usage ruin your early efforts

Users are motivated to use the product because of the initial trigger, so they should not let the initial trigger go to waste due to bugs in use; Echo's unique scene music type 3D echo once attracted me to experience it, but one night I couldn't sleep, I thought of Echo, and chose a scene music of waves. When I was about to set up "scheduled play", the following screen popped up: you need to become a member, see the picture below; even the most basic behavioral needs such as sleep have to be blocked by commercial means, so it was uninstalled and never downloaded again.

Let’s look at LOVOO, which is also a commercial product. During the IM interaction, LOVOO implemented commercial embedding of “message pinning” as shown in the figure below. Compared with Echo, users not only have no plan, but also have the urge to pay.

(2) Why does this psychological difference occur?

There has been a deviation in the commercial positioning of behavioral needs. Other platforms have cultivated the habit of free use of "sleep timing", that is, no matter what it was like in the past, at least it is a basic need now. What are basic needs? In layman's terms, they are "natural needs". Once such needs are not met, users will feel disappointed. Of course, if such needs increase, user satisfaction will not increase! The business card in echo is stuck on basic needs, so users will naturally be disappointed because there is a behavioral vacuum.

In LOVOO IM, the basic behavioral demand of users is to successfully send messages to the other party. This process is smooth. If you want to be displayed at the top of the other party's message list, you need to become a member (commercial embedding point). If you fail to be given priority, the satisfaction will not decrease, because you must pay a certain cost to obtain the privilege. This is our normal logic. Therefore, commercializing such needs will naturally not cause a decrease in user satisfaction. On the contrary, it can also generate certain payment motivations. Therefore, in the Behavior stage, you should pay attention to the type of behavioral needs and the matching relationship with commercial cards to ensure that there is no behavioral demand vacuum.

3. Reward – Behavior perception results and activity reward gameplay

The Reward of smoking corresponds to "coolness". The coolness of the product comes from the user's perception of the result of the behavior and comparing it with the expected goal, so as to make a judgment: Has the goal been achieved? If it is achieved, it is cool, and if it is not achieved, of course it is not cool; in the process of using Didi Taxi, the taxi failed, and this failure will affect my Trigger for using Didi Taxi next time; in this process, Reward is a result of the interaction between Trigger and Behavior, and it is transmitted in a closed loop with the TBRR model, as shown in the figure below. Therefore, compared with Reward, as designers, we should pay more attention to the former two.

In addition to the rewards that come from the perception of behavioral results, there is another type of reward that corresponds to the activity of virtual wealth. For example, if you log in today, you can get 10 yuan worth of virtual currency. When you collect 100 yuan of virtual currency, you can exchange it for some privileges, and this privilege is often accompanied by a growth system. No one is willing to be a green leaf for life. If the platform does not provide a path to flowers, users are likely to give up because of hopelessness. Therefore, through activities such as accumulating sign-ins or completing daily tasks to win virtual coins, these green leaves will one day become wild lilies. The road to the goal of wild lilies is filled with many thoughts, and these thoughts are the touchpoints that awaken users every time. As the reward grows, the user's "wealth" is also accumulated, and the user's dependence on the product will become stronger!

Of course, activity-based rewards are often accompanied by "competition", because only competition can highlight the value of rewards. For example, QQ often offers diamond privileges (yellow diamond users can customize more beautiful QQ space skins, and yellow diamond users can kick out non-host users in the room, while ordinary people, regardless of whether they are the host or not, have no right to kick out yellow diamond users. When purple diamond users enter a live broadcast room, the room automatically prompts "Distinguished user XX has entered the room, everyone applaud.")

4. Repeat – Increase the lifetime value of your customers

The purpose of getting users engaged is to make them unknowingly enter the cycle of the addiction model again. Investment occurs when users provide their personal data and social capital, their time, energy, and money to a product. The acquisition of user behavior data is conducive to improving the service quality of the product, thus having a positive impact on the first three stages of the addiction model. For example, Mobike is simply a means of transportation. The use-and-go feature allows users to get instant satisfaction in the action stage, while the investment stage is mainly related to users' expectations of future rewards. Here, Mobike adopts the open approach of the Mobike Achievement Center to deeply explore the user's lifetime value.

Later, Mobike's treasure box car launched a new way to play: collect three stickers, "Achievement", "Achievement" and "Hall", and you can get a mysterious opening gift; after the opening, you can get corresponding time-saving coins, environmental protection coins, and health coins for each ride, which can be used to redeem gifts or participate in charity activities in the form of points; in addition to Mobike cycling coupons and Mobike peripherals (already redeemed), the current gifts online also include coupons launched in cooperation with Meituan , iQiyi , JD.com , NetEase Cloud Music, etc., covering a variety of product lines of major domestic Internet companies, which can meet the different needs of Mobike users; Mobike Achievement Hall is a form of points mall. The charm of the internal points system of the product lies in that points can drive user investment and form usage preferences; because we often try our best to be consistent with our past behavior to avoid "losses" caused by migration and change. In this way, a non-essential behavior such as riding a bicycle may form a continuous loading of the next trigger under the incentive of external factors, and cultivate users' behavioral habits of unlocking the bicycle and parking in an orderly manner; and the more active users there are, the gathering of massive user data will also become the explosion point of the next round of shared bicycles .

Therefore, the TBRR addiction model is a thinking framework, which is equivalent to what is expressed in the fable "The Crow and the Pitcher": a crow wants to drink the water in the bottle, which is a Trigger, but is restricted by the behavior. The stone at this time is the Behavior at the moment. In the end, the crow drinks the water and thus gets a skill (Reward and Repeat).

3. The core secret of creating a billion- dollar unicorn: the user engagement level model of “core behavior”

Think about the big question: How can you do your best to build an enduring, $1 billion-plus unicorn company? This requires a combination of art and technology: the art is to understand how the product works, and the technology is to know how to measure it; this requires a combination of art and technology:

The focus of different stages is quite different. In the early stage of the company, you need to think more about the "art" side to evaluate the effectiveness of product operation; as the company enters the mature stage, quantitative assessment, that is, the "technical" side, begins to play a greater role. At its core, it is essentially about understanding how to maximize user engagement through "core behaviors"; the user engagement pyramid model has a total of 3 levels:

1. The first level at the bottom is “user growth”, which aims to acquire as many users as possible and complete the core user behaviors defined by the product;
2. The second level in the middle is to “retain users”, which allows users to have more “stakes” in the product and thus increases the cost of users leaving;
3. The third level is when users create a positive cycle in the product. Ultimately achieve our goal: to build a company worth more than 1 billion US dollars.

As you can see, this has a lot of similarities to the “TBRR” addiction model we introduced earlier, and it’s a natural match. Let’s break it down level by level:

Level 1: New user growth stage

Silicon Valley startups pay close attention to user growth. Just look at the Google index curve for the keyword “growth hacking ” in the figure below and you will know that growth is indeed important, but does it mean success?

There are many companies in Silicon Valley that regard "growth" as their key task. Many companies have good growth in the early stages, but a little later, their growth performance is very different. Snapchat 's growth was fluctuating but overall stable; Musical.ly's overall growth performance was also very good; Yik Yak later experienced a cliff-like decline...

The so-called "user growth" does not mean that we should focus on the total number of users of the product, but that we should care about the growth of effective users who complete the core behaviors defined by the product.

(1) What should be the core action of your product?

In fact, the total number of users of a product is growing no matter what (because no product will exclude users who have logged out or churned when compiling statistics), so many times looking at the total number of users really doesn’t make sense. The so-called "core behavior" refers to the behavior that can form the basis of the product. For example, the core behavior of Facebook is adding friends, the core behavior of Pinterest is Pinning a picture, the core behavior of Snapchat is taking photos and sharing them, and the core behavior of YouTube is subscribing to videos.

Generally speaking, the core behavior of a product is often directly related to user retention rate . Let me give you two examples: if a user can add 7 Facebook friends within 10 days, he may be inseparable from Facebook for the rest of his life; if a person pins some photos to a board on Pinterest, we can always expect that he may come back to take a look next week.

However, users do not only perform these core behaviors in the product. Other peripheral functions also support the core behaviors. But the ultimate goal is to have more users complete the core behaviors. For example, when users upload photos and update personal status on Facebook, they provide support for maintaining the user relationship graph. If users do not upload or watch videos on YouTube, there will be no content to subscribe to.

User growth that is not aimed at increasing core behaviors is rogue and meaningless growth. Although it makes you feel good about yourself, it has no actual value to the company. How can we confirm that we are currently focusing on the correct core behaviors? Ask yourself: What features should we build to optimize the user’s core behaviors? On the right roadmap?

(2) Can the core behavior be scaled to attract enough users?

User motivation and user ability jointly determine whether core behavior occurs. The essence of this question is whether the core behavior we define is a real demand of the mass market?

It's hard to say. The social video sharing software Viddy once acquired 17 million users from the Facebook open platform in one week, but there were not enough users to complete the core behaviors defined by Viddy.

Pinterest is the opposite. In 2011, Pinterest grew very fast because of several iconic things:

• First, the product has a clear subjective value;
• Second, the product has compound growth;
• Third, more than 50% of weekly active users will Pin pictures (completing the core behavior).

Level 2: Active user retention stage

Assuming your product has acquired a critical mass of users, the question that must be answered now is: Will these users stay and remain active? Before answering this question, we need to understand how important retention rate is. Let’s take an example again. There are two companies A and B. Company A adds 5 million new users per month, with a monthly retention rate of 80%; Company B adds 2.5 million new users per month, with a monthly retention rate of 95%.

If the retention rates of the two products remain the same, after six months, their monthly active users (MAU) are shown in the figure above, with Company A leading;

Three years later, Company B has 42 million users, surpassing Company A's 25 million users. Moreover, the growth curve of Company B is much more beautiful than that of Company A. This is the huge impact of retention rate and the value of compound growth.

Since retention rate is so important, how do we create products with high retention rates? The core is that the product needs to allow users to increase their "benefits from continued use" and "losses from leaving the product" at the same time. In other words, these two dimensions are actually two sides of the same coin.

• The so-called "increasing returns from use" means that as users contribute more and more data to the product (whether intentionally or unintentionally), the product can continuously improve the user experience based on this data, and finally reach the situation of "the more I use the product, the better my product experience"; it can be understood this way: in the process of using the product, the user explicitly or implicitly increases the available data of the product, and the company further improves the user's product experience through this data.
• The so-called "increasing loss upon leaving" means that the longer you use a product, the more it becomes your dependence, your identity, or a place where you accumulate some kind of value. Finally, the situation is reached where "the more I use the product, the more I will lose if I leave it"; it can be understood this way: the more times/time a user uses a product, the more they will lose if they leave it. Such a product represents the user's personal identity to some extent and is an indispensable part of a certain aspect of the user's life.

(1) For example, in Evernote, the core behavior is note-taking. As you take more notes in Evernote, the value you get from it increases, and the harder it is to leave it; that is, Evernote's core behavior creates both increasing returns from use and increasing losses from leaving.

(2) On Pinterest, the more you Pin, the more you lose by leaving the product - your bookmarks, expressions, and identity system (increasing loss of leaving); initially, the product creates "increasing returns of use" by focusing on features, and then adds some related features to reflect the increasing loss of leaving, such as Pinterest's "Featured" feature, where the more you Pin, the more your homepage meets your personalized needs.

(3) Some anonymous community applications such as Yik Yak and Secret have done well in terms of user growth, but they do not have obvious "increasing returns from use" or "increasing losses from leaving", so their subsequent value is limited. Although Twitter also has an anonymity function, overall the user identity is fixed, so it has two major characteristics: "increasing returns from use" (new follows/followers) and "increasing losses from leaving" (personal identity recognition, such as relationship chains), and has long-term value; in short, if a product wants to retain users, it needs to provide continuous value over time and emphasize the losses faced by leaving the product.

Level 3: Generating the virtuous cycle of the product itself

Assuming your product has acquired a critical mass of users with good stickiness, the question you must answer now is: Are the engagement patterns of existing users sufficient to generate multiple virtuous loops within the product? A virtuous cycle is like a flywheel that converts user engagement into fuel that drives your company forward; the strongest virtuous cycles have network effects.

(1) Let’s use Pinterest as an example

The leftmost loop in the figure below (the "increase product value" loop) is the most important virtuous loop for Pinterest. It actually forms a "network effect loop": the more users pin pictures, the more content there is on the platform, and the better the user experience. The two loops in the middle and on the right are important product mechanisms in terms of user growth and increasing user activity.

The detailed generation mechanism is as follows: The more you Pin → the richer your interest graph → the better your content discovery experience → which in turn encourages you to generate more Pins…

At the same time, there will be other similar circulation systems in a product. For example, in Pinterest’s growth cycle, users discover more good content worth Pinning → recommend it to friends → help the product acquire more users → generate more good content. For example, in Pinterest’s user engagement loop, user A repins a photo of user B → notifies user B → recalls user B → B may stay and generate more Pins.

You should maximize each naturally occurring virtuous cycle. For example, in the cycle mentioned above, in order to allow notifications to reach more users and achieve the effect of recall, the growth team will work hard to increase the number of dormant users.

(2) Virtuous cycles are difficult to create — the vast majority of products on the market have failed to do so, despite many attempts;

In fact, these are issues that need to be considered comprehensively at the beginning of product design. How big a product can be is actually determined from the beginning. If the design is wrong, no matter how hard you try or how much money you raise, it will be of no avail.

For example, Evernote performs well in user growth and retention, but is a little weak in sustainable development. It is more of a tool for users to obtain value, but it does not generate a virtuous cycle for the product itself. This is a common problem of all tool-type products, and it is also the fundamental reason why many tool-type products want to add community and user interaction elements after they have developed to a certain stage and scale.

For example, Tinder, a dating matchmaking product, has built some virtuous cycle systems. For example, after a successful match, the other user will be notified and they will be called back for more participation. But there is a serious problem overall: every time a user successfully goes on a date, the couple will leave the product. Successful matching does not bring more active users (as shown in the figure below), and it does not form a virtuous cycle for the product itself. In China, stranger social products represented by Tantan are difficult to escape from the dilemma of users adding WeChat and QQ for further chatting after getting to know each other.

(3) Virtuous cycles drive company growth. We should look for opportunities to maximize virtuous cycles. The most important thing is to be able to quantify:

The following are the three core quantitative indicators:

1. The growth of engaged users enables new users to continue to participate;
2. Retaining users can improve user retention rate;
3. A virtuous cycle that helps products achieve organic growth and retention.

The clearest way to understand user engagement for a company’s products is to look at the retention performance of the same cohort of users, which is also the ultimate metric. The performance of the same cohort of users includes two dimensions:

• First, look at the number of users who complete core behaviors each week;
• The second is to see what percentage of these users are among weekly active users.

Specifically, cohort user analysis includes:

1. Growth: the number of people in each cohort;
2. User engagement: the percentage of users who complete core behaviors;
3. Retention rate: The retention performance of this cohort of users over time.

When looking at the retention rate of a product, in addition to the regular next-day and next-week retention of new users, it is also more important to look at the retention rate of the same group of users. For example, when I was at Douban , I would define a few data: for example, there were 1,000 users who registered Douban accounts on a certain day. Among them, there may be 100 users who have joined at least one group (this is the core behavior). After one day, seven days, and one month, how many of these 100 people (identified by user account or UDID) are still retained? The retention rate of the same group obtained in this way will be more instructive for the product than the retention rate in a broad sense. There are many ways to reflect the retention performance of the same group, such as line graphs, tables, or triangular heat maps.

Here, we must remind you that although the emergence of these cycles is difficult to grasp, we should try to develop product genes that meet such characteristics as early as possible in the early iteration process of the product. It is a virtue to plan carefully before taking action on major issues. After all, only a virtuous cycle of products is the core driving force for the company to move forward. Please remember: user engagement is the fuel that drives most of the enduring non-transactional consumer Internet companies with a value of more than 1 billion US dollars. It is not about brushing water, brushing rankings , falsifying, or buying water armies...

Easter egg: Core behavior "Find a girlfriend". When my friend told me that he met his girlfriend by playing games , I immediately took action and started playing games! My friend came to my house yesterday and saw me playing games. He asked: Are you addicted? Me: What? I'm not looking for a girlfriend! My friend looked at my game and shook his head: You are playing a single-player game, you can’t find a girlfriend; I was completely devastated! I am not addicted to the game, I am addicted to the person who plays the game.