A Beginner's Guide to User Growth!

The scope of growth is very large. This article focuses on the Internet and related industries, focusing on " user growth ". The following will explain the definition of user growth and the core concepts of user growth, specifically what user growth is, what user growth does, and how you need to pay attention to user growth based on your role.

1. From growth to user growth

Growth is everywhere

Growth is the general law of things. The growth of cities, the expansion of population, the spread of plants, the spread of viruses, the growth of companies, etc. are all part of growth. However, there is no unconditional continuous growth. At a certain stage, it is necessary to rely on some new conditions, or to break through some shackles, or to accept some changes.

(1) Urban Growth

The expansion of a city's territory first depends on the development of urban transportation. The network composed of highways, light rail and subways will gradually widen the city's horizontal boundaries; in addition, breakthroughs in the construction industry make the city's vertical expansion possible. Ultimately, the city forms an ever-growing "structure" that accommodates more people and allows for connections between any two entities. Such a city is like an organism that breathes, excretes, and suffers from illness. When its size reaches the limit of the negative pressure it can withstand, the city stops growing.

(2) Population expansion

In the early 20th century, the world's population began to grow rapidly after the medical community basically solved global deadly diseases (such as smallpox) and improved infant survival rates. At the beginning of the 21st century, many developed countries and cities have generally experienced negative population growth, such as Western Europe, Northern Europe, Japan, Shanghai, etc. The fast pace of life in big cities and the strong pressure of high-paying jobs have made childbirth an "option" for some young people, and population growth has stagnated. For example, in China, even after lifting previous policy restrictions, there is still no obvious population recovery. If the fertility rate remains low for a long time, population aging will become increasingly serious, and the entire social welfare system will be under tremendous pressure.

(3) Plant spread

Vines are the masters of growth in the plant kingdom. They rely on efficient photosynthesis and use new fulcrums to gain new space for spreading. Even in difficult environments, they can still climb wall after wall. When the nutrients they produce are no longer enough to support one more step of climbing, the leaves at the edges will begin to turn yellow, and the boundary of the spread will be determined. At this time, more sunlight and more nutrients are needed to break the current balance and continue to expand the territory.

(4) Virus transmission

The spread of viruses has terrified humans. They first invade a host, then spread from host to host in a horrific chain, causing the number of viruses to grow exponentially. The two most terrifying large-scale viral outbreaks in history were the Spanish flu in the early 20th century, which caused 30-50 million deaths, exceeding the total death toll of World War I; and the Black Death that broke out in medieval Europe, which claimed 15 million lives. Only when the virus transmission is cut off, that is, when the transmission coefficient R0 is less than 1, the virus cannot spread to more new hosts, and its impact can be controlled and randomness can be quickly disintegrated.

(5) Enterprise growth

The growth of a company may be due to a bright idea or by grasping a technological trend, leveraging the market with one or two leading products, gradually occupying market share, and eventually forming an industry oligopoly.

Over the past fifty years, the business world has been repeating this story. In the information technology and Internet segments alone, as described in Professor Wu Jun's "The Wave at the Top", dozens of companies have experienced such a rush to the top of the wave, but most of them have fallen with the receding tide and even disappeared on the beach.

When a company's revenue reaches its limit and lacks new growth points, it becomes very dangerous. At this time, even just one wrong decision can destroy a business empire built over decades, such as the Nokia mobile phones and Lehman Brothers that we are all familiar with.

Growth is everywhere, and the various types of growth introduced above are essentially connected: growth first requires some starting conditions or opportunities to be seized; when the scale gradually increases, sustained growth will face huge challenges and limitations; if the accumulated problems cannot be solved and the current limitations cannot be broken through, growth will stop or even quickly turn to decline. Next, let’s focus on growth in the business world: it’s brutal, because if you stop growing you’re likely to die quickly.

Growth in the business world

The total market value of the stock market reflects, to a certain extent, the overall size of commercial companies in a certain period of time. In the long run, the market value of the stock market has been increasing, which means that the overall scale of business is constantly increasing.

The Dow Jones Index, whose full name is the Dow Jones Composite Average, calculates the average stock price of a number of representative companies in each period (i.e., the component stocks, initially 12 and later stabilized at 30), and is calibrated to ensure comparability with the initial value.

The Dow Jones Index increased 580 times from 40.94 points when it was first published on May 26, 1896 to around 23,800 points at the beginning of May 2020. The calculation method of my country's A-share index is slightly different. From the 100 points when it was listed on December 19, 1990, it has increased nearly 30 times in less than 30 years. The macro scale is growing, and if a company grows too slowly, it will gradually lose its survival advantage and eventually face elimination.

The original 12 components of the Dow Jones Industrial Average have been completely replaced with the removal of General Electric in June 2018. Many century-old companies have experienced ups and downs, and it is not easy for them to survive to this day.

The business world is very cruel. If you don’t grow, you will die. The market generally regards the company’s market value as the most important growth indicator, which comprehensively reflects the company’s current value and future expectations.

Market capitalization is affected by many factors. For companies that mainly provide mass products and services, user scale and its growth expectations are particularly important. The vast majority of Internet companies belong to this category.

Since it falls within the realm of economies of scale, the more users or customers there are, the more marginal benefits can be generated and the greater the overall rate of return (this book refers to the audience of products or services as users, and users will be used to represent users or customers in the following sections); in the same segment, user scale and market value are highly positively correlated.

For companies that rely heavily on advertising revenue, such as search, social, and information flow, user scale is even the lifeline of the company's survival. Because this type of product or service is free to users, a large number of active users consume huge bandwidth costs, and can only generate revenue through advertising.

Advertising revenue = Number of active users * Average revenue per user

Over a certain period of time, average revenue per user (ARPU) is a relatively stable value. It usually depends on the bidding level of advertisers and is affected by the overall market conditions.

Therefore, overall advertising revenue is mainly determined by the number of active users, which also explains why Internet companies have attached great importance to user growth in recent years. The user dividends brought by the popularization of mobile Internet have been gradually divided up, and natural growth has almost stagnated. It is no longer sufficient to support the continued growth of Internet companies. A dedicated team is needed to focus on user growth.

So, what exactly is user growth?

What is user growth?

The term "user" in this article refers to the audience of Internet products and services (including B-end and C-end); user growth refers to the growth of user-related indicators, including the user scale and the various impacts it produces, mainly including the following three aspects:

• User scale, such as Monthly Active User (MAU) and Daily Active User (DAU)
• User duration, such as total duration, average duration per person
• Commercial income, such as user payment, advertising income

What is the difference between user growth and growth hacking? Essentially, their goals are the same, and their work scope is basically the same; the biggest difference is the perspective of thinking.

Growth hackers, as shown in Figure 1, believe that the premise of growth is the fit between product and market (Product Market Fit, PMF). They advocate the AARRR or RARRA model, and are more from the perspective of the enterprise. They are guided by improving indicators related to customer acquisition, activation, retention, recommendation, and recommendation, and ultimately achieve user growth, including user scale and commercial revenue.

Figure 1 AARRR model in growth hacking (picture from the Internet)

User growth is mainly considered from the user's perspective, emphasizing growth based on the improvement of user value. In addition to user scale and commercial revenue, the indicators of concern also include user time.

User growth also emphasizes experimental methods as the core to support scientific and efficient product and operational decisions. Compared with the technical style of growth hackers, user growth is more like a goal-oriented integration of existing product planning and product operation work. It is a series of thinking methods and working methods.

Figure 2 Demand and salary levels for user growth-related positions (Image from Lagou.com, screenshot from May 2020)

For Internet companies, growth is almost equal to user growth, so departments and positions for user growth have almost become standard.

Excellent growth managers are scarce, so the positions they can offer are high and the salaries they can earn are considerable (as can be seen in Figure 1.2).

In recent years, more and more user growth positions have emerged. At the end of 2016, Didi Chuxing already had positions such as "user growth expert" and "user growth operation". Since the second half of 2019, Tencent's product promotion channel has also added a growth strategy direction.

The user growth position is an iterative version of product planning and product operations. It has many good opportunities and new and broad room for growth. We will continue to discuss the application and recruitment of user growth positions in the future.

Next, we will first discuss how to view growth from the user's perspective, which emphasizes a global perspective based on user value.

2. The driving force of user growth

Continued use of the product comes from gaining value

There may be many reasons why users continue to use a product, but in essence, the product provides some value to users, and this value is likely that other products cannot provide.

Users will continue to use the product and even call their friends to use it together only when they have gained value from using the product, including but not limited to gaining friends, pleasure, knowledge, benefits, and income. The following apps that everyone uses on a daily basis provide clear value to users, which means that users will continue to use them and new users are constantly joining.

(1) WeChat

WeChat has brought people closer together, provided free calls and the lowest-threshold social network - Moments. The circle of friends provides everyone with the most convenient stage for display and expression; various groups of "loving family" and "old classmates" provide a good place for middle-aged and elderly people who are new to the Internet to express their care and maintain friendship; and now everything can be done by scanning a code, which provides great convenience in life for everyone. The unique social attributes of WeChat have formed a huge network of interpersonal relationships, in which users will join spontaneously and find it difficult to leave.

All of the above are the main values ​​that WeChat provides to users, and are also the reasons why users choose and continue to use WeChat.

According to data from QuestMobile at the end of March 2020, the 180-day retention rate of WeChat’s active users was 95.5%, ranking first among all apps and far higher than other products. For general apps, a 60% retention rate on the next day is already very good. Such a high long-term retention rate shows the great value that WeChat brings to users.

(2) Tik Tok

Today, no matter whether you are extremely obsessed with short videos or sneer at them, you can no longer escape the influence of Douyin and Kuaishou. Thanks to the low-cost, high-speed mobile network that has been fully popularized in our country, short videos and live broadcasts can be played smoothly in buses, subway stations, and even high-speed trains. As the main representative among them, Douyin has an extremely low threshold for content consumption. The main operations can be completed by just swiping or double-clicking with one hand, allowing users to easily consume a large amount of high-quality content while lying down.

Tik Tok has become the main form of entertainment for killing time, and statistics show that its overall usage time is second only to WeChat. On the other hand, Douyin has an extremely low threshold for content production. As long as you have creativity, you can use a mobile phone to produce works that get hundreds of millions of views and millions of likes. As a platform, Douyin is also constantly optimizing the rules on both ends of the content, allowing high-quality and popular content to get more exposure and playback, allowing excellent content producers to be rewarded, and ensuring healthy growth on both ends of the ecosystem.

Data at the end of March 2020 showed that Douyin’s average weekly active users exceeded 500 million, and its 180-day retention rate was 79.5%, second only to WeChat among all apps. Short videos have become an absolute national-level application, and the main value they bring to users is pleasure.

(3) Zhihu and Bilibili

If watching short videos is considered a pure waste of time by some users, then seeking knowledge and learning on the Internet can be considered a more narrow form of value acquisition. In this regard, Zhihu, as a long-established question-and-answer community, has successfully survived the desktop Internet era to this day. As it increases the diversity of its content, including Zhihu Live and professional courses, while embracing hot topics and launching hot lists, its users are constantly increasing.

Another similar platform is Bilibili (B Station), which has gradually faded the two-dimensional content and provided more young users with a place to consume high-quality content. The friendly and interesting community atmosphere, as well as the influx of a large number of knowledge bloggers (called UP masters on B Station), have made learning on B Station a new trend. Zhihu and Bilibili are no longer niche, and the changes or even destruction of the community atmosphere will inevitably be despised by a small number of old users.

But from a holistic perspective, serving a larger number of users and providing them with a place to learn and entertain is providing value to more users. For products and companies, this is their most ideal destination.

(4) Pinduoduo

Pinduoduo's journey has been full of controversy. In the beginning, it was nicknamed "Pingxixi" because of its counterfeit goods. At present, it has been renamed "Pin Diedi" through continuous subsidies of hundreds of billions and the long-term launch of the lowest-priced branded digital products on the entire network.

Pinduoduo illustrates the user growth brought about by providing value to users, on the one hand through low prices and subsidies, and on the other hand through extremely low-cost shopping experience. With the help of WeChat's ready-made accounts and payment system, Pinduoduo has swept a large number of users in third- to fifth-tier cities, as well as various towns and villages through social fission in the WeChat environment. Many middle-aged and elderly users started online shopping through Pinduoduo, and many rural merchants also started online sales through Pinduoduo. These are real user values.

At the end of April 2020, Pinduoduo released its 2019 annual report showing that its transaction volume in 2019 reached 1,006.6 billion yuan, and the number of annual active buyers reached 585 million.

With such a large user base, the 180-day retention rate of active users is 47.3%, second only to Taobao's 51.6% in the e-commerce category. These have posed a powerful impact on Alibaba's e-commerce empire. Pinduoduo continues to provide discounts to buyers and has always provided merchants with 0 commission and 0 platform service fees, providing more revenue. These user values ​​have made Pinduoduo deeply rooted in the minds of users.

Good products can continuously provide value to users, so users will choose to stay and even actively spread word of mouth, bringing more valuable users to the product. This is the most ideal and healthy user growth.

The driving force for growth lies in improving user value

Since the commercial use of the Internet, the underlying logic of user growth has not changed: the increase in user value has brought about lasting and healthy growth in user scale and revenue. So, how do we define user value enhancement? Many years ago, Mr. Yu Jun proposed:

User value = new experience – old experience – switching cost

This formula can be used to qualitatively describe the ways to increase value: 1) Improve the difference between the old and new experiences; 2) Reduce replacement costs. The difference between the old and new experiences ultimately requires users to feel it personally. In addition to the new experience being really good, marketing methods are also needed to convince users to be willing to try the new experience, such as the fast charging experience of "charging for five minutes and talking for two hours". Once you have tried it, it will be difficult to go back to slow charging.

In addition to the obvious price cuts, various subsidies are also available to reduce migration costs. For example, the popular 10 billion yuan subsidy mentioned in the previous section has made Pinduoduo the preferred platform for students to buy electronic products.

In addition to reducing payment costs in a narrow sense, it also reduces the cost of obtaining information and completing consumption. For example, just swiping the screen allows you to immerse yourself in watching the non-stop Tik Tok and Kuaishou videos; and there is no need to compare prices among different stores where you can just “buy it, buy it, buy it” in live streaming e-commerce. The absence of the need to search and compare greatly reduces the cost of obtaining information; the threshold for consuming short videos and live streaming is almost zero, which greatly reduces the consumption cost compared to text and image information.

The above-mentioned products are undoubtedly outstanding representatives of user growth in the past one or two years. Although there is no shortage of jokes about the user sinking strategy ("outside the fifth ring road" and "surrounding the city with the countryside"), it has to be said that user value is a subjective cognition. There are high-brow and low-brow ones, and there is no distinction between high and low.

The growth of users is inseparable from the improvement of user value, and this needs to be understood more as the improvement of the value of most users; ignoring the growth of user value, or failing to ensure continuous value improvement, is more of a short-term self-indulgent carnival.

Having made it clear that enhancing user value is the basis for user growth, we also need to consider how to deliver value to users who need it. Especially in some crowded market segments, there are already many competitors competing for users.

3. A global perspective on user growth

User Value

Figure 3: Global user growth perspective “iceberg chart”

The previous section introduced that user growth needs to be based on user value. However, when it comes to user growth, what is more perceived in the public eye is marketing strategy. For example, in recent years, there have been Spring Festival Gala red envelopes during the Spring Festival. WeChat red envelopes, Alipay, Baidu, Kuaishou, etc. have successively attracted a large number of new users from them; the annual 6.18 and Double 11 have become shopping carnivals; and, recently, celebrity live streaming with goods has just started, and Mr. Luo Yonghao bought a car and a mobile phone on Douyin. These are the top of the iceberg, the most perceptible marketing strategies.

Starting from the foundation of user value, it supports insights into macro opportunities and business models, and further drives the implementation of marketing strategies in a more scientific and effective manner through data. This is a global vision for user growth.

Macro Opportunities

According to QuestMobile data, the scale of domestic Internet users officially peaked in April 2019, and the year-on-year growth in November 2019 was almost zero. At the same time, we can see that user time still maintains a relatively high growth rate, and as the monthly active users of national short video apps represented by Douyin and Kuaishou reach 400 to 500 million, the year-on-year growth rate of user time in the second half of the year continues to increase. The "traffic pool" completed the transformation into the "time pool" this year, which contains macro opportunities.

Figure 4 User scale and duration trends, data from QuestMobile "2019 Traffic Growth Inventory"

In November 2019, the average daily online time spent by the entire population reached 6.2 hours. What does this mean? Assuming an average sleep of 6 hours, users still spend 1/3 of their day staring at their phone screens. It is no exaggeration to say that mobile phones have become "organs", and each of us is inevitably drawn into this vast "time pool".

If the growth in user scale is due to the popularization of smartphones, then the increase in usage time is undoubtedly due to the popularization of low-cost, high-speed mobile networks (in this regard, based on my comparison with about 10 developed countries in several continents in recent years, China is the only one in the world). But are technological innovations and popularizations alone sufficient to support the surge in duration? Let’s look at two examples.

The steam engine was invented and applied to industry in the mid-19th century. By the beginning of the 20th century, the United States was basically able to use steam engines to replace horses as the main power of transportation, but it did not. The reason was cruel: the steam engine itself was too heavy for ordinary roads to support. In addition to this, another important reason was the resistance from the public. The steam engines were too noisy and accompanied by thick black smoke, so people would rather endure the countless horse feces and urine on the roadside.

Also in the United States, in the early 20th century, when large household appliances had been produced and the prices had been low enough for civilian use, even though Americans were lazy, these still could not bring about a buying frenzy. The reason was cruel: these electrical appliances consumed too much power, and the lines at the time could not bear it. There were not even standardized current, voltage, plugs and sockets.

These two examples tell us that macro growth requires many conditions. In addition to the maturity of the technology itself, two important factors need to be considered: user willingness and the maturity of the ecosystem. Users' willingness determines whether the technology application is widely accepted, while a mature ecosystem and its norms can ensure that the technology is applied on a large scale.

Let’s go back to the reasons for the increase in usage time: If the growth is attributed to the low or even free mobile networks, then why is user usage time still increasing significantly in the WiFi environment of first- and second-tier cities where data traffic has long been free? The reason comes from another dimension - the cost of content consumption has dropped sharply, and people are more willing to consume. We can define the content consumption cost as the total time to complete the target content consumption, which can be further broken down into:

Content consumption cost = content acquisition cost + completion consumption cost

Figure 5 Content consumption costs of different content formats

Short videos have an extremely low content acquisition cost - no need to search, just scroll and you can't stop; they also have an extremely low completion consumption cost - immersive and not easy to jump out, and only 15 seconds to 1 minute, you can watch this one several times without realizing it. Therefore, the content consumption cost of Douyin and Kuaishou is very low. Let’s take a look at live streaming. If it’s a live streamer you follow, the cost of acquiring content is also very low - usually when you open the app, the entrance to the live streaming will be right in front of you; the cost of completing the consumption is even lower than that of short videos - you don’t even need to swipe at all, just put the phone holder on and watch while eating takeout, and you can grab a red envelope and buy something.

Almost at the same time, the norms of the content ecosystem are constantly being improved, and are contributing to the continued reduction of content consumption costs: apps are easier to use and recommendations are more accurate; tools make content easier to create and high-quality content easier to spread; high-quality producers receive a lot of material incentives and even become an industry with high incomes, and creation becomes sustainable.

Therefore, with the popularization of low-cost, high-speed mobile networks, the extremely low cost of content consumption has caused the time spent on content consumption to skyrocket. By the beginning of 2020, it had surpassed long videos, mobile games, etc., and content consumption is being deeply integrated with e-commerce, providing "time owners" with greater profit margins beyond advertising revenue. By grasping the macro opportunities, user growth can be carried out on an upward "surface", and the effect can be increased by leveraging potential energy (or the "trend" that was popular in the past two years). Otherwise, the result may be half the effort with twice the results, and the medium- and long-term benefits cannot be guaranteed.

Business Model

Innovation in business models is extremely difficult, and for most practitioners, it is something that only happens by chance. More often than not, it is about optimizing the shortcomings of the existing business model, or more specifically, repairing the breakpoints in certain links and finding new opportunities.

As shown in Figure 6, the content consumption field currently presents a complete ecosystem, and many of the roles in it did not exist a few years ago. For example, the Multi-Channel Network (MCN) service connects grassroots content producers and content platforms, providing ordinary content creators with professional packaging capabilities, better opening up their content consumption channels and the influence of producers. MCN makes the content consumption ecosystem smoother, while also seeking high revenue shares for itself. Its prototype is a brokerage company, but when it is introduced into a new field, it fills the gap in the business model very well.

Figure 6 Global ecology of content consumption

For example, live streaming sales, which has been extremely popular recently, also has e-commerce channels on the original content platforms (the shopping links below the short videos). However, this is not direct enough, and the quality of the combination of goods and content is still uneven. Television shopping has existed for decades. Now it has been improved and has been brought to the Internet live streaming platform. The anchors use more professional and provocative descriptions to promote their products, merchants replace products more frequently, and the platforms provide more convenient purchasing methods. These measures have optimized the business model of content-driven sales and injected new impetus into it.

Data-driven

Data-driven in a broad sense includes analyzing to discover opportunities, experimentally verifying cognition, data-driven decision-making, etc. In user growth, the main purpose of data-driven is to find specific entry points. On the one hand, problems or opportunities can be discovered by scanning existing data, performing correlation analysis, etc.; on the other hand, the existing strategies are evaluated through experimental methods, and the next iteration is carried out based on the strategy results.

Data-driven development must first clarify growth goals, preferably a core indicator that is recognized by the entire company from top to bottom. All growth work is broken down around this indicator. The most common method is to determine the North Star indicator, and then break it down according to the DuPont analysis method to find the entry point. This part will be expanded in Chapter 2.

In user growth, experimental methods are a crucial part of data-driven development. Since we are doing growth, we need to focus on the incremental indicators brought by the strategy, and the experimental method is the golden rule recognized by the world's first-tier Internet companies. Although experiments face many challenges and there are many scenarios where ideal experiments cannot be conducted, it is still the first choice when conditions are met. The experimental design and analysis parts have been introduced in previous articles. You are welcome to read them.

Some traditional operational strategies mainly rely on manpower to complete configuration and iteration. However, in areas such as algorithm-driven content recommendation, dynamic pricing, and order splitting strategies, algorithm models have intelligently completed personalized and globally optimal strategy issuance, and continuously optimized the results through reinforcement learning. Representatives of these algorithmic strategies include

• Content distribution strategies of Tik Tok and Tencent News
• Didi and Meituan’s order dispatching strategies
• Taobao and Didi’s dynamic pricing strategies

Some of them will be introduced later.

Overall, data-driven helps evaluate the effectiveness of marketing strategies and provide feedback to serve the iteration of marketing strategies and ensure the ultimate achievement of growth goals. This part is very important, but it is also generally lacking, so I will spend a lot of space to give a detailed introduction to the experimental methods, experimental analysis, efficiency tool design ideas, etc. involved in data-driven development for your reference.

Marketing strategy

The products and operational strategies involved in user growth are all included in the broad marketing strategy. In general, a marketing strategy is an interface directly facing users. Its functions include but are not limited to delivering information to users, explaining some interest points, interacting with users, guiding users to complete key behaviors, and finally promoting the achievement of growth goals.

The above contents can be used to design targeted experiments to explore the optimal effect. For example, the common product promotion interface in e-commerce apps will clearly tell users how much they can save by purchasing a certain amount. If you are not sure how to design the two numerical variables involved here, you can conduct experiments to see what combination of values ​​can bring the largest GMV or the best ROI.

For example, in content consumption apps, historical data can show that once users interact with others (forward, comment, like, etc.), their subsequent retention rate will be significantly improved. At this time, the product can design some small functions to guide users to complete forwarding, commenting and liking. For example, after a user posts content, he or she will get some likes by default, which will greatly promote his or her enthusiasm for posting content.

For example, in the sign-in function, users can share a bonus by checking in seven times. After the user checks in for the first time, they will be given another chance to check in. Experiments have shown that the proportion of users who are given the chance to check in seven times is significantly higher than the group who are not given the chance.

Marketing strategies are visible to the public. Although the previous examples may seem simple, they all have underlying logic that enables them to be effective. Some references can be found in "Consumer Psychology" and "User Behavior". The exploration and implementation of marketing strategies rely on the support beneath the iceberg, that is, based on user value, through insights into macro opportunities and business models, and with the help of data-driven scientific and efficient implementation. This series is the main content of user growth work, which will be introduced in the next section.

4. Main tasks for user growth

Core workflow for user growth

From the introduction in the previous section, we can see that user growth is a systematic project. Ideally, we hope to have a complete data platform, labeling system, experimental platform, and even a good algorithm model to support automatic delivery of strategies. But at the same time, user growth is also a race against time. Even if we have nothing yet, we need to start as soon as possible, accumulate positive experience bit by bit, and build tools on demand to improve efficiency. Assuming we are in the early stages of user growth and do not yet have the above conditions, how should we start?

(1) Clarify growth targets

The growth target is the first thing that needs to be made clear to the boss. We have decided to do this work, what is the background, and what indicators do we ultimately want to improve? You can refer to the "North Star Indicator" to make your choice. There are many related articles, so I won’t go into details here. Let’s assume that after careful consideration, we decide that the metric we want to improve is DAU.

(2) Building a growth model

When planning the overall UG work, we will focus on the overall situation and look at the various links in the entire user chain and life cycle (the most common AARRR and RARRA that emphasizes retention); when we get into specific work, we will find that these "big models" often cannot guide us in what to do.

Let’s go back to our specific example: having only a DAU target cannot guide any work. DAU needs to be broken down to an executable level. How can DAU be split?

a. Inflow and outflow perspective

Think of DAU as a container. The “inflow” refers to the daily new users and returning users, while the “outflow” refers to the lost users. Take "day" as an example for the observation period:

• New users refer to new users gained on that day.
• Returning users refer to users who were inactive yesterday but are active today.
• Lost users refer to users who were active yesterday but inactive today.
• Retained users refer to users who were active yesterday and are active today.

It can be inferred that:

Today’s DAU = Today’s inflow + Yesterday’s stock – Today’s outflow = (New users + Returning users) + Yesterday’s DAU – Lost users

So, if you want to increase DAU, you can start from two sides:

• Improve new additions and increase return traffic (that is, open source)
• Improve retention or reduce churn (i.e., save money)

What we need to decide is, resources are limited, which side should we start from? It is necessary to analyze the composition and trends of the above categories in DAU in detail. In principle, priority should be given to "where the gap is" or "how to do the best to increase the volume". It is also necessary to make specific resource allocation based on the current status and stage of the product (for example, whether there is enough budget for paid growth).

b. Perspectives of new and old users

Similar to the inflow and outflow perspective, but relatively simplified. Simply divide DAU into two parts, one is "new users" (that is, those added on the same day), and the other is "old users" (that is, those not added on the same day), then:

Today’s DAU = yesterday’s new users * new next-day retention rate + yesterday’s old users * active next-day retention rate + others

The equation mainly has 4 variables. If you want to increase DAU, you can substitute the known quantities into the above formula. According to current experience, we can see which indicator can obtain greater benefits. For example, if old users account for 90%, increasing the secondary retention rate by one point will increase the DAU by 90%*1%=0.9%, while increasing the new secondary retention DAU by one point will only increase it by 10%*1%=0.1%. The “others” here include today’s new users and the return of yesterday’s inactive users. To simplify the discussion, they are all categorized as others.

c. Activity perspective

From the perspective of activity, we can look at it from the perspective of weekly active users WAU, DAU=WAU*weekly active days/7, or from the perspective of monthly active users MAU, DAU=MAU*monthly active days/number of days in the month.

From this perspective, we will focus on how to improve "active days". First, we will look at the level of active days, how much room for improvement there is (what is the average, what is the proportion of low-active users), and which group of people to focus on first (those who are active 1-2 days a week, or 3-4 days). This involves user activity modeling, which will be introduced in detail later when introducing "user segmentation".

The above models help us know the general direction, how to exert force, or where to exert force first. At the same time, DAU is also broken down into several smaller indicators that are more proactive, more sensitive, and can be associated with strategies. Help us realize: "If you want to increase DAU, you must first increase XXX."

(3) Finding a strategic entry point

Now that you know how to exert your efforts, you still need to implement specific strategies in the end. When deciding where the strategies need to be implemented, you can consider causality and correlation to help.

a. Priority causality

If you know why a certain indicator is not high, it is recommended to find the cause directly from product logic and user feedback and make targeted repairs. The qualitative method is to find typical users to ask and verify, and the quantitative method can locate the "breakpoints" of the user path through the most intuitive funnel analysis. Understanding user needs and enhancing user value are essential skills for product managers and product operators. They are also the essence and driving force of user growth, so I will not go into details here.

b. Secondary relevance

Correlation analysis can help us find user behaviors that are highly correlated with key indicators, so that we can position specific strategies to improve the key indicators and verify whether our growth goals can be improved.

Continue to take improving DAU as an example. Suppose we clearly understand that the most important thing in improving DAU is to increase the number of active users. We can further look at which behaviors of active users are correlated with the next day retention (simple linear fit, or more complex "magic number" analysis).

Generally speaking, the user's active retention is highly correlated with the user's usage depth on the day of activeness, such as the duration of the information flow app, the PV of the graphic and text reading, and the VV of the video playback. Through the analysis of the retention rate of user groups of different stratified and correlation with the above indicators, we can draw on which indicators have better correlation with secondary retention, and we can give priority to improving these indicators.

It should be noted that: 1) There are some specific pitfalls in the correlation analysis. You can first make up for the statistical basis just in case; 2) The correlation is high, but it only provides us with a hypothesis: improving this indicator will most likely increase the active retention and further improve DAU. Whether it can be really improved needs to be verified through design experiments.

c. Use experiments to quantify effects

Experiments help us verify whether the hypothesis is valid. The effect of quantitative strategies: how much has been improved by applying strategies or not applying strategies, publishing or not publishing functions, what is the improvement, what is the short term, what is the long term, what is the accuracy, and what will happen after increasing volume.

What is the actual work of user growth: analyzing data - forming hypotheses - experimental verification, the three cycles are the core workflow of user growth.

This is also the key to differentiating traditional product operations and product managers: whether it is oriented towards growth goals, whether it depends on data decisions, and whether it is used for experimental methods.

Do you need to pay attention to user growth

The previous article introduces the vision and capabilities required for user growth and what specific things are being done. Finally, let’s ask yourself whether you need to pay attention to user growth or which part you need to pay attention to in your current position or role.

If you are a middle- and high-level decision-maker in the enterprise, it is recommended to pay attention to the macro opportunities in the field of the product, whether the current product goals can be leveraged by the potential energy of the trend, whether there are optimization points in the current business model, whether the current product operations and reality are being carried out around key goals, whether data-driven methods are used, and whether marketing strategies have data as guidance.

If you are a product manager who is responsible for planning a core function of a certain APP, it is recommended to pay attention to the source of user value and think about how to continuously improve user value. It is also necessary to pay attention to whether there is sufficient data collection in the product, which will be an indispensable part of comprehensive data analysis, especially user behavior analysis. The user growth of the entire product depends on the scale, duration and retention of its core functions. It is recommended to explore what the functions you are responsible for have to do with the activity and retention of the overall app? Is this function driving the activity of the market or being supported by the market users? Where is your value?

If you are a product operator who is responsible for improving the next-day retention of active users, it is recommended to pay attention to the retention data of similar apps in recent times, study products with higher retention, and what marketing strategies are used to improve or maintain high retention. After finding a reference method, it is recommended to conduct some low-cost experiments in the App to constantly find out which strategy can effectively reach and retain users. In addition, you can also find out which user behaviors are highly correlated with user retention rates from existing data, because these all have opportunities.

If you are an active self-media practitioner on WeChat, there is currently a lot of data in the backend of WeChat’s official account that can help evaluate whether the content meets the audience, whether the communication is wide enough, and whether the profit is large enough. It is recommended to treat the official account you manage as a complete product to grow users. In which problems and opportunities can be discovered based on existing data, and conditions can be created to design some methods that are close to experiments to continuously optimize copywriting, content organization, and layout styles.

If you are not involved in specific Internet products for the time being, you are not responsible for any growth-related indicators. It doesn’t matter, the experimental methods and their core thinking will also help you better evaluate the gains and losses of some decisions in life, so that the cause and effect are clear and clear.

User growth is not limited to a specific task. More importantly, it is a systematic way of thinking and working methods. The more you are careful, the more you can appreciate the benefits it brings.

summary

The article is long, so everyone randomly chooses the readings they are interested in. The following are the key points:

1. User growth mainly refers to the growth of user scale and its related effects
2. User growth is based on improving user value
3. User growth requires a global perspective: based on user value, through insight into macro opportunities and business models, and with the help of data, we will implement scientifically and efficiently
4. After clarifying the growth target, the core workflow of user growth: analyze data - form hypotheses - experimental verification, and gradually improve this cycle

Author: User Growth Practical Notes

Source: User Growth Practical Notes