Product Promotion Operation | How to improve user activity of low-frequency products?

Having 1 million users is not a problem. The real difficulty lies in: how to maximize user activity among these 1 million users? In this article, the author will analyze a low-frequency product project that he has personally experienced based on the modular data analysis method, and explain how to better improve the user activity of low-frequency products.

We often talk about the 0-1 of a product. Do you know how many users are needed to reach 1 for a C-end product?

——One million users is only 1, but according to the current popularity of mobile Internet and the education level of users, it is really not difficult to obtain one million users.

Maybe 1 million users is a lot in your eyes, but in the eyes of investors, 1 million users is just the starting line. In addition to the user scale, there is often a deadline. The shorter the time it takes to gain 1 million users, the more likely it is to be favored by the market. (I once led a product that gained 1 million users within one month of its launch. As far as I know, there are other products that can gain 1 million users within a week of their launch).

For today's mobile Internet, the biggest challenge is no longer the 1 million users, but how many of these 1 million users are active and how many can continue to stay.

This is a very acute issue and also a challenge we face.

If your product already has 1 million users, but there are less than 50,000 active users every day, what should you do?

Sticky perception of the product

In the past, we believed that the culprit for low activity was the low frequency of business needs, so when establishing a project, we would carefully consider the frequency of the demand - the so-called high-frequency or low-frequency demand.

This is the view of the older generation of Internet people, but it is no longer so applicable today. In fact, many excellent products have been born in low-frequency demands, the most representative of which is the disruption in the e-commerce field.

Strictly speaking, e-commerce products also belong to low-frequency demands, and most users only make online purchases once a month. In such a low-frequency scenario, the up-and-coming Pinduoduo has surpassed the established JD.com in some data.

We talk less and less about the frequency of demand. In the Everyone is a Product Manager community, it is almost difficult to find articles analyzing high or low frequency of demand. This is because current product means can already make up for the frequency problem of demand.

The frequency of user usage has shifted from a market issue to a product issue, or in other words, we have a better understanding of user activity: the frequency of demand is not equivalent to user activity.

In fact, we have other ways to increase user activity and improve product stickiness, and demand frequency is only one of the factors that determine product stickiness, and it is not even the main factor.

Wu Jun (the god of letters from Silicon Valley) mentioned the concept of "toothbrush function" in his book "Insights". For some low-frequency products, we can effectively improve user activity by implanting some toothbrush functions. Only when the stickiness of the product reaches a certain indicator and users are highly active can our business be used continuously, generate sufficient value and have the opportunity to become an excellent product.

The toothbrush function means: let users use it three times a day like brushing their teeth, although each time only for a few minutes. Over time, users will form a habit of using it.

If you still believe that products should be pure and single and cannot accept modules that are not related to the core business, then you might as well think about this question:

"Pinduoduo" and its "Duoduo Orchard", the latter has nothing to do with the former's e-commerce and group buying, but it provides great stickiness support for the former. Although there is no data support, I believe that users of Duoduo Orchard have a higher order rate. And now, Pinduoduo is undoubtedly on the side of success.

Duoduo Orchard: Users can cultivate fruits, and after a period of continuous cultivation, they can get 1 kilogram or several kilograms of fruits for free. The so-called training requires users to log in to the product every day and complete some tasks. Overall, Duoduo Orchard is a gamification module. Pinduoduo also provides many gamification modules, such as Golden Pig Makes Big Money, a gamification module that uses red envelopes to purchase user usage times.

Sticky modules and low-frequency services (example)

Whether the user is active depends on whether the product has some "toothbrush functions" that can increase the number of times the user uses it.

High-frequency demand itself can bring a great number of usages, but for some products with low-frequency demand as the core, implanting a "toothbrush function" can also achieve the same effect as high-frequency demand.

In fact, high-frequency demand itself is just a "toothbrush function" and can be replaced by other "toothbrushes".

We once believed that demand was more important than products, and products were merely appendages of demand. But for now, the reality is: products exceed demand. Products are the closest to a business model and are the things that directly carry commercial value.

Demand is only a part of the product system and is very important, but there is much more that can influence the value of the product than just demand. It's like the load-bearing walls in a house are particularly important, but the location, floor, and even property management can affect the commercial value of the house.

The following are projects that I have actually experienced. I have blurred the data, but the overall proportions and the information conveyed by the data are real.

Low frequency products (example)

We already know that e-commerce products are low-frequency products. Most users only make online purchases once a month, so rebates are an even lower-frequency product than e-commerce.

This is a project I personally experienced. The user scale is in the millions, and we will temporarily call it Product A.

Project introduction: Product A is a rebate product. The platform establishes a cooperative relationship between JD.com and Taobao. Users can purchase the same goods on JD.com and Taobao at a lower price through Product A.

This product is well received by users. After all, it can bring real benefits to users. However, the actual data results are rather bleak.

User activity is relatively low compared to other types of products, and the user life cycle is also relatively short. There are many rebate products, but only a few survive, and more rebate products die in unclear circumstances.

We have organized many user return visits. The root of the problem lies not in the core business of the product, but in the stickiness of the product. When users shop online, they forget the existence of the rebate product and even express doubts about having ever registered for the product. When we explained the function of the product to the other party, they highly recognized the rebate model, after all, it allows users to obtain real benefits.

Despite this, in our subsequent monitoring, the user still did not generate any activity.

The biggest difficulty with low-frequency products is not the quality of the business demand itself, but that the products are forgotten in the intervals between demands. During non-demand times, these products have no sense of existence.

In order to verify one of my conjectures, I conducted a data analysis with an observation period of 30 days. The object of analysis this time was not the entire product, but several business modules in the product were extracted for modular analysis.

Here I will share some illustrative data with you. Of course, the data itself has been blurred, but in terms of proportion, it is not much different from the real data.

Definition of noun:

• Shopping: Using product A to complete shopping on Taobao or JD.com is the core business of the product;
• Friend Help: A new customer acquisition function that allows you to share with friends and earn more commissions.
• Red envelope shaking: a sticky function, which allows users to shake red envelopes every day, with a 5-minute interval between two red envelope shakings. On average, users will participate in 10 red envelope shakings per day.
• Early wake-up check-in : This is also a sticky function, the same as the common early wake-up check-in;
• Initial users: refers to new users of this module, such as the first time to shake a red envelope or the first time to get up early and punch in;
• Retention: refers to the retention of this module. For example, if a user continues to participate in the red envelope shaking on the second day after the first shaking, it will be considered as the retention of this module. If the user uses other services in the product but does not use the red envelope shaking, it will not be counted as the retention of this module.
• Proportion of active users: The proportion of active users of this module in the entire product's active users. 60% means that users who shake red envelopes account for 60% of today's total active users. If there are 1,000 people shaking red envelopes today, then the total number of active users of Product A is 1,666, and users who shake red envelopes account for 60% of the total active users.

The various proportions of this data are very close to the real data. It is not difficult to find that the user activity of shaking red envelopes occupies a very important position. To put it another way, if we abandon this module, the daily activity of the product will drop by 60% or even more.

From the perspective of retention, the data of shaking red envelopes far exceeds that of other modules. We can even think that users who have participated in shaking red envelopes will have a higher probability of retention and even have a higher frequency of use.

In Product A, shaking red envelopes completely serves as a "toothbrush function". On average, each user will participate in shaking red envelopes 12 times a day. Although they only stay for 2 minutes each time, they will use the product 12 times a day.

This is a modular analysis method. During the data analysis process, I abandoned the interpretation of the overall data and instead conducted an in-depth analysis of the data of a certain module.

Because the analysis is more granular, we have indeed discovered many problems that we had not considered before.

For example: We found that user activity does not depend on our core business, but rather a branch business drives user active behavior.

I have done some research through some third-party data platforms. Product A ranks among the top 5 rebate products. Rebate products with better development all have some functions that can significantly increase user activity, such as Huasheng Diary's MLM ability.

As for some dead rebate products, almost none of them contain modules that can effectively play a sticky role.

Now, among so many excellent products, it is not difficult for us to find some modules that seem to be unrelated to the core business but can greatly improve user activity and even fission capabilities.

With the help of this modular data analysis, we made a very wise decision to direct new users to the red envelope shaking module. This decision increased the retention of the entire product by 30%. I remember that the daily active users also increased by 15%.

Module-based data analysis

In large companies, modular data analysis is very common because they have so many modules and also because in large companies, a project team may only be responsible for one module.

But I think that more factors lie in the fact that large companies have many seniors, product people with rich practical experience, who will apply modular data analysis as a requirement and a specification in the team. It is actually difficult for self-taught product people and some potential teams that have not yet become large companies to be exposed to the concept and methods of "modular data analysis".

It's a shame, but at least it's not too late to start now.

I can tell you with certainty that the product is made up of several modules, and the core business needs are only a larger part of the many modules, just like the load-bearing walls in a house, which are very important but not the only one.

In the product, each module plays a different role. Some realize paid conversions, some realize commercial revenue, some are responsible for fission and attracting new users, and some are responsible for increasing user activity. They are interrelated and yet completely independent.

Especially now that the product market is mature enough, and product design methods are emerging one after another, it is completely impossible to compete with others in the current environment by simply building a core business.

Qutoutiao has over 100 million users. Unlike traditional information products, the quality of articles in Qutoutiao is not high, and there is a huge gap in reading level compared with Tencent News and Toutiao. However, Qutoutiao's fission module is indeed something that other information products do not have.

What Qutoutiao did was once considered as a deviating from its main business in the Internet world. It's not that no one had thought of deviating from its main business and finding a new way, but it was not accepted by the team, or it just couldn't get past its own test.

The same is true for Yunji. E-commerce is not the embodiment of Yunji's competitiveness, but the store manager system is. 3 million store managers have brought 30 million users to Yunji.

Products have begun to enter the era of modularization. The stage of market competition has gradually shifted from core business competition to module competition with clear data orientation. What we are competing for is whose fission module can bring more new users, and whose stickiness module can bring higher user retention and user activity.

Given this market situation, it becomes particularly important for us to master modular data analysis methods and have corresponding capabilities.

Many problems are difficult to find solutions for the entire product, but it is easier to find solutions if the problems are assigned to corresponding modules.

In modular data analysis, we advocate dismembering the product, which will be divided into several stages.

Let's take the sticky module as an example:

1. First, label each module.
2. Store each user's usage behavior in separate data;
3. Compare the user's first usage date and calculate the user's daily retention;
4. Normally distribute users across the platform;
5. Get a module-based retention data report.

The whole process is equivalent to making a mini version of the BI system, which is relatively complex and costly.

Such a modular data retention report can help us more clearly judge the quality and value of the module, which will ultimately affect our three important decisions.

1. Which module should new users be directed to? Because the retention capacity of the module can directly affect new users, thereby greatly improving the retention capacity of new users.
2. Modules that seem important but actually have low retention rates mean that users will not continue to use the module, or the module is used less frequently and cannot serve as a "toothbrush". We can stop investing resources in this module or even abandon it.
3. Retention and stickiness are also an indication of user trust. Users in this module tend to have higher trust in the product. Compared with users in other modules, they are more likely to pay, even pay multiple times and repurchase. Users in this module are more likely to become super users of the product.

In some countries, just relying on one city can drive the number of tourists for the entire country. For example: Cambodia and Angkor Wat. For the former, more than 80% of foreign tourists will concentrate in Angkor Wat.

Products, like countries, do not need every module to be so attractive. Many modules are only for supporting certain business, and these modules are valuable. But we must understand that value and attractiveness are not exactly the same. In addition to valuable modules, we also need some modules that do not seem to have much business value but are attractive.

How to perform modular data analysis?

For most teams, it is very difficult to conduct modular data analysis because we cannot afford the R&D costs of "data modularization", which is equivalent to a local version of a BI system, which is very costly and difficult.

When I was trying modular analysis, it took me two weeks of development time just for data tracking and mining, not including the time I spent on calculations and template design.

Excluding the option of self-development, there is one thing that we as practitioners should be thankful for. Some third-party data statistics systems also have the ability of modular statistics. With simple access and settings, modular data analysis can be carried out. For example, Umeng+’s custom retention can meet our needs for retention monitoring of each module. We can easily see which modules have high retention and which modules have low retention.

This feature is indeed very applicable. We can connect our own sticky modules to the custom retention system, set the event name, and then track and monitor the new users (initial users) of each module, as well as the user retention status under the module.

This gives every team the ability to perform modular sticky data analysis, although we are still exploring more modular statistics.

Summarize

We discussed the issue of user stickiness together. When we try to improve user stickiness, we can try to implant some "toothbrush" functions, and the user's stickiness to the module will be transferred to the product itself.

Through actual cases, we also learned the method of modular data analysis. We can divide the data into modules and conduct independent data analysis on each module to obtain more targeted information.

For sticky modules, you can try to analyze the module’s own retention capacity and the ratio of active users to product active users to determine the quality and potential value of the module.

Finally, I shared with you the custom retention function provided by Umeng+, which allows us to easily implement modular data statistics and analysis on stickiness.

If there is an opportunity in the future, we can continue to explore the data analysis methods of the fission module and the paid conversion module.

Related reading:

1. Product operation and promotion: How to compete for traffic?

2. Community Operation丨The essence and gameplay of "private domain traffic" in 2019!

3. Product operation: 2 major ways to get started to accurately capture private domain traffic!

4. Promotion and marketing: A brief discussion on vlog’s brand marketing and traffic monetization!

5. User operation: user growth in the post-traffic era!

6. Online marketing promotion: How to spend big money? Fine-tuned operation of large traffic!

7. Product operation and promotion | 5 underlying ideas for traffic growth!

8. Product operation: application of data system under the growth model!

Author: Dead Leaves

Source: Product Manager Charging Station