Practical experience: Talk about the practical experience of "user growth"!

The concept of User Growth (UG) originated from the United States. Judging from the trend, it will definitely become a standard feature of major Internet companies in the future.

However, there are relatively few people and companies that have actually done UG, so most books and articles still present cases of foreign products. People in the industry know that the Internet industries abroad and at home are two different things. Reading successful cases from abroad is of no use other than making you feel good.

Everyone can pay more attention to and study UG in the future. It will be useful and this type of talent will be in great demand. This article is my understanding of UG. It may be one-sided, but it is ultimately extracted from actual combat.

Many people around me say that this is just a cool concept and has no practical significance. Everything we do is for growth.

That’s right. In the past 20 years, the domestic Internet has been focusing on growth, from the initial UV and PV to the subsequent DAU and GMV. Every product iteration and every operation activity is to improve these numbers. Isn’t this growth?

Of course they are different, I will explain in detail below.

Point 1: User growth is a set of methodologies

Some bosses think that UG is the key, and once they get it, they can open the door to product growth and quickly see changes in data. This is definitely wrong.

User growth is a set of methodologies, just like product development and operations. It is not a panacea and will not hit the mark. If it is just unsystematic individual operations, there is no way to talk about user growth.

Simply put, it is the AARRR model in the figure below. No matter what product it is, attempts are made to segment based on this main line. Only after repeated optimization and AB testing can a better solution be found.

As can be seen from the above picture, UG is not just about attracting new customers , but also includes retention and even revenue growth. For UG, closed-loop thinking is a very important quality, which requires a more rational and digital approach, focusing on the entire user life cycle until commercialization.

Suppose you make an app and need to spend money on advertising and promotion activities. It costs 20 yuan to attract a new customer. But on average each user can bring you 30 yuan in revenue, so the ROI is positive.

This way you can spend as much money as you want, there is no upper limit on the promotion cost, and the only thing that limits you is the number of users you can acquire. Because the more money you spend, the more money you earn back, this is closed-loop thinking.

To give an extreme example, there is an app for dating. By changing the name and icon of the app, dozens of identical apps were copied, all with names like "XXX in the same city". There is no upper limit on their advertising costs, because all the money spent can be earned back, and this closed-loop model is successful.

The above is my understanding of UG, which can be summarized as follows:

User life cycle is the main line, data is the guide, user or revenue growth is the goal, integrating products, operations, and technology as the means of execution

To put it another way: no matter what method is used, as long as it can drive growth, it is right.

Corresponding to UG is the practice of most Internet companies at present. Although the KPIs of product, operation, technology, marketing and other departments are all user volume, if you think about it carefully, the work content of each department will not directly drive growth and are separated from each other, which brings us to the second point.

Point 2: User growth is a teamwork approach

Currently, the division of labor in most Internet companies is based on products or business lines, corresponding to roles such as product, operation, R&D, and design, and usually, this is also the way the team is divided, that is, PM is a team, and the same goes for operations and R&D.

This is a very strange phenomenon. There is very little need for collaboration within the PM team, which is far less than the collaboration between PM and R&D. Most of the time, PMs work with R&D, design, and interaction. So why let a bunch of PMs form a team instead of letting those who often collaborate form a team?

Maybe there are historical reasons, or maybe a team consisting of "a bunch of PMs" can find a more senior PM to manage them. In any case, this model has become unreasonable now.

What’s more serious is that this division of labor model has led to the emergence of a “silo phenomenon” within the company. Like frogs staying at the bottom of their own wells, they only look up and wait for the boss to speak, while information sharing and collaboration are poor, and the entire company is divided.

In order to solve these problems, many companies have disrupted the traditional division of labor and established multiple small-scale project teams with all the necessary facilities. This can solve the efficiency problem by optimizing collaboration, but it is still not enough.

A UG team is needed to connect all departments of the company, break the "silo phenomenon", and truly do something with the goal of growth.

This is still a stretch. Let me give you an example. If an e-commerce product hopes to increase purchase conversions from new users, should it optimize here or there, or run a powerful promotion? In fact, neither is correct. This is because you have not found the source of the problem, or you are still thinking in a fragmented way.

According to UG's thinking, in order to achieve good purchase conversion of new users, we must first analyze the source of new users. List the major channels for acquiring users, such as:

1. GuangDianTong
2. App Store
3. OV Store
4. Baidu SEM

These four channels cover 80% of new users, and each channel has different delivery characteristics and target user groups, so we need to analyze these four channels one by one and break them down into the following questions:

Refined delivery strategy → Obtain user portraits → Make targeted products or operational measures → Form a mature plan

1. Analyze the channel data and delivery strategies to see if they can be more accurate. Only by ensuring that the channel acquires high-quality and accurate users can subsequent retention and conversion be meaningful.
2. Obtaining the user portrait and behavioral characteristics of each channel is likely to be a no-man's land. Although channel colleagues need it, they don't have much motivation to push forward.
3. Based on the portraits obtained, different channels should conduct corresponding conversion experiments. This is the most core execution item. The specific strategies are not limited to the product or operation end. When it comes to converting new customers, we certainly cannot generalize the issue to all new customers.
4. Form a mature plan and implement it, and maintain normal operation afterwards. UG will not continue to follow up on this project. Once it is completed, it will be handed over to the corresponding colleague.

The above is UG’s general idea for doing this. You can imagine that if this matter is handed over to PM, he will definitely constantly sort out the product conversion funnel, and the focus will definitely be on product-side optimization; if it is handed over to operations, it is estimated that he will continuously carry out refined promotional activities and use operational tools to promote, and he will think less about the product side.

Which method is more reasonable and thorough? Now it is very clear. This is a specific case, but this is roughly the difference between UG and traditional business division of labor.

The specific approach should follow the above figure, that is, the user life cycle AARRR model, and send it again.

Among these 5 steps, the focus is on the first 3 steps: acquisition, activation and retention.

Step 4 "monetization" is the focus of commercial products. I rarely touch upon this part, so I will avoid it here. The fifth step, “recommendation”, is an extension. It mainly depends on the product being well made and some incentives and stimulations can be added to achieve it.

The specific methods are discussed below one by one. Due to space limitations, I can only touch upon them briefly.

Step 1: Acquisition

It’s all about attracting new users. It’s simple to say, as long as it’s a way that can attract new users, it’s fine. But you need to pay attention to the following indicators: new users, next-day retention, cost per user, and user acquisition ROI.

These numbers actually tell you that you need to look at the number of new users, the quality of users, and whether the money spent is worth it.

The specific methods can be divided into three categories:

1. Channel placement

For companies with a promotion budget, this is the most important way to acquire customers. There are many tricks here. I am not an expert and will only talk about it from the perspective of growth.

Let’s not talk about channel expansion and business relationship maintenance for now. There is a lot of room for exploration in the delivery strategy alone.

①Optimize the delivery materials and increase the CTR of information flow , manufacturers, and store promotion resources . This is a process of continuous experimentation, communication, and results observation, which involves relatively many trivial tasks, especially the communication link.

The ideal situation is to replace manual labor with algorithmic strategies. The algorithm can generate delivery materials for different people and can quickly replace the materials, thereby increasing ad exposure and CTR. This is another relatively large branch, which I will cover in two sentences.

②Refine the delivery strategy to improve ROI. In other words, we want to make the investment more reasonable and "spend money where it counts."

For example, you have to water 10 pots of flowers every day, but you only have a fixed pot of water every day. You can evenly distribute them to these 10 pots of flowers and water each pot a little;

You can also take a look at these 10 pots of flowers first to see which ones need watering and which ones don't. You only need to give this one pot of water to the flowers that need watering. Perhaps you can save water, and then you can raise 15 pots of flowers in the future, and the amount of water needed will remain the same.

This is the principle behind refined delivery strategies. It is not about blanket delivery, but about analyzing users first, developing delivery strategies, and spending money only on reasonable things.

For example, by placing information flow ads , you can choose to acquire new users or awaken old users. For new users, you can select basic information such as gender, age, region, and interests, and try to identify the target group that is suitable for your product to achieve refined delivery. For old users, you can choose to deliver to non-high-activity users, such as users who have not visited in the past 30 days, to avoid delivering to users who would have visited every day.

For example, in more extreme cases, you can identify users that the product cannot reach, or users who have already left, and use campaigns to reach them and recall them. This is more purposeful, and although the scale may not be large, the strategy is reasonable.

③ Maximize resource utilization, including pre-installation, manufacturer and third-party store promotion resources.

This is centered on channel business, and requires obtaining as much information and resources from resource parties as possible, connecting external resources with internal teams, and obtaining more methods for growth.

For example, if you have pre-installed products, there will definitely be unactivated inventory. Using the manufacturer's promotional resources to activate the inventory is a quick and efficient way to acquire new users.

For example, for Android manufacturers, there are many resources and ways to explore, some of which are official and some of which are explored privately. Even though there is very little room for big manufacturers like OV, Huami and Xiaomi, we can also study some relatively small brands , perhaps there are blue oceans there.

2. Task system

Pinduoduo's monthly GMV is said to have reached 40 billion, while JD.com took 6 years to reach 10 billion, and its daily order volume has already exceeded JD.com.

It took Qutoutiao (pictured below) one and a half years to reach a DAU of over 10 million.

The task system of Qutoutiao

These are two products that are currently very popular. Their specific forms and positioning are different, but they both use "profit" as bait to get users to split.

Through the task system, users can "recruit disciples" and acquire new users; they can also set tasks to improve retention, comments, sharing and other behaviors, and users can receive cash or gold coins as rewards after completing them. The product obtains advertising revenue through user behavior to cover the cost of rewarding users, forming a closed loop.

Gold coins can also be directly exchanged for cash, and the specific exchange rate is floating, which can ensure the stability of ROI.

This is a representative form of acquiring new users. It is not intended to encourage everyone to "recruit disciples", but rather to draw on this fission task system.

3. Activities

There are so many different interpretations of the definition of activity. For example, the big promotions by JD.com and Meituan are activities, the book throwing event by Xin Shixiang is an activity, the communication H5 by NetEase is an activity, and the Million Heroes by Toutiao are also an activity.

But the activities I’m talking about here are entirely for the purpose of attracting new customers or improving retention, and benefits such as brand exposure are not included. Therefore, among the activities mentioned above, the more typical UG-oriented activity is Toutiao’s Million Heroes.

If the video answering format had not been stopped at the time, it would probably have become the standard. This type of activity brings a good amount of new users to the product because of its low participation threshold, attractive cash rewards and resurrection card dissemination mechanism.

What kind of activities can bring about user growth? What is the next "Millionaire"? I don’t know now, but there are ways to explore.

Establish an event project team that focuses on growth, whose members include event planning , product, interaction, vision, front-end and back-end research and development, etc. Set common goals, let everyone sit together, and keep trying quickly. You can even copy verified forms from abroad, like answering questions via video.

4. Seize the dividend

This is a bit vague, but it is indeed a direction. UG needs to keep an eye on the latest developments at home and abroad, quickly seize the benefits of new technologies or new ways of playing, and use strong execution capabilities to implement them to acquire new users for the products.

To give a random example, sending text messages via Alipay to receive red envelopes (see the picture below) is a good way to increase DAU. The cost is controllable, which is the SMS cost + red envelope cost; the operation is simple, just copy the SMS and then open the client. This operation has been taught before and many users know it.

There are endless new ways of playing, ranging from a small text message to a video quiz. UG students must maintain their curiosity about new things and the ability to execute quickly in order to seize the benefits.

Step 2: Activation

The word "activation" is ambiguous. In fact, what we need to do at this stage is to retain new users , which is the most important link to some extent.

Because when you get 10,000 new users in a day, only 1,000 of them will stay, and the secondary retention rate is 10%. Then the cost of losing 9,000 people is wasted, and the actual number of new users is only 1,000, not 10,000. If we can increase the retention rate of new users to 20%, we can bring in 1,000 incremental active users every day, which means 30,000 per month.

Retaining new users is the intermediate link between acquiring new users and activating old users. If this step is not done well, it will cause a break in traffic. New users will be lost and old users will not be accumulated, and the entire market will not grow.

For example, if you have just opened a small restaurant, the general idea is to serve the first batch of customers well so that they can come again and bring more friends. It’s the same when making an app. The first batch of new customers must be served well, and then more and more regular customers will come.

The specific methods are as follows:

1. Optimization of user access paths

Sort out the paths that users take to access your product from different sources, list them step by step, and see which links can be optimized. Ensure that new users have a smooth and enjoyable experience using the product and leave a good impression, which will lead to subsequent retention.

Taking Mobike ’s new user usage path as an example, the following figures show the situations before and after optimization respectively. It can be seen that not only steps are saved, but also the experience ( scanning speed) is optimized.

It is important to take all aspects of the user experience into consideration, such as page loading speed, negative experience caused by advertising, etc. Users may experience slow loading or errors on different models (especially low-end machines) and in different network environments (non-WIFI). These are difficult for PMs sitting in the office to imagine, so more testing and research are necessary to analyze the data funnel.

2. Interest incentives

Let’s take the small restaurant as an example. When you visit for the first time, you will be given a coupon for 20 yuan off for purchases over 100 yuan, but it can only be used next time and is valid for one month. This is how new user retention is achieved through interest incentives.

How to define "retention" is actually not just "next-day retention" in a narrow sense, but an "indicator" that can help retain users.

For example, through data analysis , we can find that the retention rate of new users with behaviors is much higher than that of users without behaviors. Therefore, to retain new users well, we need to increase the proportion of active users among these new users. This is the path to achieving the goal.

It's a bit confusing, let's use an example:

• Social products: Users who leave their personal information have a higher retention rate than those who don’t;
• Mutual financial products: The retention rate of users who have purchased financial products is higher than that of users who have not purchased them;
• E-commerce products: The retention rate of users who received coupons is higher than that of users who did not receive coupons;
• Information products: users who comment have a higher rate than those who do not;

How were the above cases achieved? Profit incentives are one of the methods.

During the subsidy war a few years ago, Meituan stimulated new users to place orders by issuing coupons. At this stage, it is impossible to obtain sufficient user portraits, so we don’t know the preferences of new users and can only distribute coupons in the categories that users browse.

There are also some products that give red envelopes to new users, which can be received every day for 7 consecutive days. This is the most direct inducement, which increases user retention in the first few days. However, the cost of this kind of activity must be controllable. It is unreasonable for the cost of a single retention or 7-day retention to be too high, but lowering the cost will cause users to be unaware of the red envelopes.

3. Resource Tilt

Just think about it, what are you thinking after you post a message to your friends ? Are you expecting friends to comment or like your post? If there are, or even a lot of them, you will get pleasure and satisfaction; if not, you will be disappointed.

Users will also have this experience when using your product.

Suppose you are a new user on a community product like Zhihu or Tieba . If after you post a piece of content, it is displayed in an important position, or quickly gets a few replies or likes, you will be very happy and will be very likely to stay.

This is why people who work in community operations will constantly check posts and reply immediately when they see new people posting something. This is to make the other person feel that "pleasure" and think that there are many users discussing here and it's very lively.

Rumor has it that only a few drivers were using Didi when it was cold-started , so Cheng Wei asked several employees to hail taxis and drive around the city, which would make the drivers who were using Didi feel that this was a reliable service and that they would be able to attract customers. The effect of this behavior is actually retaining new users.

Let’s go back to the example of the small restaurant. When the boss finds out that this table of guests is visiting for the first time, he might try to chat with them, try to get closer, and maybe even give them a fruit platter or something. The boss's behavior is also to retain new users.

In the above cases, community operations reply to new users’ posts, Didi spends its own money to hail taxis, and restaurant owners try to get close to them. These are all targeted investments of resources for new users. Therefore, allocating resources to new users is also a way to improve retention.

Step 3: Retention

After acquiring users and doing targeted retention work, the remaining users enter the pool of "old users" and are no longer affected by the first two steps. The next step is to retain these "old users."

In fact, if the product itself is valuable and can form a closed loop, it can retain old users. But the actual situation is not so ideal, so we still have to do something.

1. Incentive system

This is a huge topic, and it is not appropriate to elaborate on it here. I will just talk about the principles.

The incentive system cannot be equated with specific plans such as points, levels, sign-ins, medals, etc. We must combine user needs and product selling points before considering solutions. The specific solutions are not limited to the above ones.

It can be seen that many "check-in" products only use this function as a tool to improve retention, but they have not designed a complete closed loop. Users have no reason to sign in, and there is no timely and effective feedback after doing this.

Similarly, the points levels of most products are just for show. On the contrary, more successful products, such as WeChat official accounts , Zhihu, Douban , etc., do not have these.

For example, in the picture below, the Weibo level is of little value, the level privileges of keep are rather useless, and the medal is intended to be purely spiritual motivation, but it feels not enough.

Weibo and Keep user system

Looking at the picture below, Didi’s tasks and memberships provide price discounts and priority dispatch, which are valuable to users. I don’t know the data, but I guess it will be better than Weibo and Keep in the above picture.

Another screenshot is the task interface of the game " Clash Royale ". Through tasks, users may obtain cards that can only be purchased with money, or even cannot be purchased with money. This is also valuable to users. Moreover, the multi-task setting can ensure that users can access the application multiple times a day and initiate multiple operations.

Didi and Clash Royale Quests

Didi’s user privilege system

The tasks of Qutoutiao mentioned above are also a kind of incentive system for old users . Through inducement, stimulate old users to continue visiting.

When building a user system, we must focus on user needs and formulate solutions. We cannot be bound by existing models, otherwise we will end up at a level similar to Weibo. We need to find ways to create something truly valuable that benefits both users and products.

2. User Reach

Old users are aware of the product brand and are loyal to it, but they still need to be "reminded" from time to time to attract them back to the product. This is user reach.

There are probably four common forms: push , system message, SMS, and email. The first two are still usable, but the effects of the last two are very poor in most scenarios.

There are several key points for push:

• content. There are many optional published contents, with many categories and strong timeliness, and the effect can be guaranteed regardless of full or personalized push;
• aisle. Ensure that push can reach more devices and find ways to get more users to turn on the push switch on their phones.
• Strategy. The push strategy needs to include which users to send to, when to send, how many messages to send per day, whether to add pictures to the push copy, whether the sound can be adjusted, etc.

In addition, push must be well managed to ensure that the published content and categories are stable and reasonable. Create a report to monitor data performance every day, and review the effect of each item in a timely manner to adjust the strategy as soon as possible.

Among these four ways to reach users, push is the most effective, but also the most difficult. I will just briefly touch upon a few key points here and will not go into detail on the other three.

3. Activities

The main benefit of online activities is to activate old users and increase the frequency of activity. For example, before the event, they come once every three days, and after the event, they come every day. On the contrary, it is very difficult to attract new customers through activities. From a cost perspective, it is better to let channels do it.

Toutiao’s “Collect Zodiac Signs” campaign during the Spring Festival (see the picture below) has the effect of attracting new users and promoting activation, but here we will only talk about promoting activation.

Today's headlines collect zodiac activities

If you want to collect all 12 zodiac signs, you need to launch the app every day and use such activity incentives to increase the retention of old users. Of course, this is a multi-dimensional revenue activity that can attract new users, socialize, and bind cards, but the cost is controllable. Judging from third-party data, it also has a pulling effect on Toutiao's overall DAU. This is the effect of a good activity.

After talking about the significance and practice of UG, let’s finally talk about common problems. After all, there is a big gap between theory and reality, and this concept has not been widely understood, so it lacks the soil for rapid growth.

Problem 1: Weak data foundation

Most employees of domestic Internet companies feel that their product data foundation is weak, which is normal.

First of all, most of the complainers are people who "use data" rather than "make data". From the perspective of my own needs, I always feel that the data is not satisfactory.

Secondly, small companies do not have the time, cost or need to pay attention to data. Many large companies are not data or technology-oriented. This is the current situation in the industry.

But the problem is that data is very important to UG. Instead of relying on data to make decisions, data is needed to verify subjective speculations, so that a large number of AB tests can be done quickly to implement the final plan. Companies that do well in this area run hundreds or even thousands of AB tests online every day. They have mature technical platforms and are completely independent of client releases.

It doesn’t matter if the data is weak. As long as it is judged to be necessary, you can build it bit by bit, even if it takes more time. Although this kind of infrastructure construction does not produce results quickly, if we persist in pushing forward its construction, the benefits will be huge in the long run.

Problem 2: Internal collaboration efficiency

My feeling is that the most difficult part of being a UG is not the business skills. As long as the main line is clear, the team is strong, and you continue to push for optimization, you will definitely see results.

The difficulty lies in the fact that the entire company has different levels of understanding of UG, which makes it difficult to implement. The main problem is the low efficiency of internal collaboration and poor communication within the company.

Because products, operations, R&D, channels, marketing, etc. belong to different departments and are managed by different bosses, and even the KPIs are different. UG happens to need to integrate these roles together to do things, which is bound to encounter collaboration problems and conflicts with existing businesses.

For example, from the UG perspective, a certain product module needs to be optimized, but this demand conflicts with the demand for the basic product, and it is not the UG students who will ultimately implement it. In this case, there must be sufficient reasons to be recognized for this demand before it can be included.

Sometimes, it may not be a problem of "demand" but a problem of "cake". The UG team will affect the interests of the original team, and the cooperation will inevitably not be smooth.

Solving this problem is difficult, but also simple. It must be recognized and promoted at the company and boss level. Only this way can it be fundamentally solved.

Problem 3: Lack of talent

When a company wants to do UG, the first thing it thinks of is to find someone with experience. However, there are not many people in the industry who have actually done this and have accumulated some methodologies. Compared with the demand gap, the gap is even greater.

For talents, this is a great opportunity. Students who have worked in product development or data analysis before will have an advantage in switching to UG. At a minimum, they need to have logical thinking and data analysis skills, as well as strong communication and promotion skills.

For a company, it would be ideal to be able to directly find suitable and experienced people. If you can’t find it, you can also dig from within. At least the executive team is easier to build. The difficulty is that we need a leader who understands UG. This is a large project and we need someone to steer and coordinate its progress.

The above is my understanding of UG. The problem is too big, so I can only touch upon many branch issues briefly. Due to my limited vision, it may not be the complete picture of UG, but at least it is extracted from actual combat. I hope it will be helpful to everyone.

Source: Han Xu