Umeng + recently released the white Paper on User Growth of Small Programs, which summarizes the methodology, solutions and success stories of Umeng + in statistical analysis and refined operation of small programs. The following is part of the white paper:
Both enterprises and governments hope to upgrade all data, whether their own enterprise data or external data, in the cloud to improve the overall digital operation capability of enterprises. The enterprise should make a data center, which can be the private domain user center within the enterprise, or the cloud on the center. This trend is called ** “digitalized operation” or “digitalized new infrastructure” **.
There are three stages of intelligent operation of enterprise numbers
Taking Umong + as an example, all public domain traffic data are gathered together and then placed in the whole fusable field. Then, enterprise business data is cross-analyzed in the private domain and the data after normalization is correlated one by one.
So that enterprises can see different platform, regional, or user data in different channels, finally forms the enterprise internal a user operating the whole report, help enterprises to check where is the user, the user what kind of behavior, should eventually through what channels to better manage and promote the user’s active and after purchase.
The digital operation map of the enterprise is divided into three stages:
Stage 1: Better statistical analysis and access of all data.
This access includes not only external contact data, but also small program data, H5 data, App data, CRM data and so on. After all data access is completed, the more data, the more operational opportunity points can be seen and generated. These data access into the data bank, the entire enterprise internal digital bank can do better digital management.
The second stage: digital private domain fusion and normalization.
Not only digital normalization under existing accounts, but also all accounts of anonymous potential customers with the help of external media, small programs, etc. Umeng + helps enterprises identify not only the account level, but also the device level. Whether it is mobile or App side, private domain cross-end analysis, with the help of these devices, and then to do the crowd selection and insight, including intelligent tags, tag sorting and so on, to better understand what their private domain users are like.
Stage 3: Combine all the contact methods, whether official or private, to achieve better operation through these contact methods.
Four stages of small program statistical analysis
According to the specific statistical analysis of small program, the user life cycle of small program can be divided into four parts.
Customer acquisition analysis, comprehensive understanding of customer sources.
The decentralized nature of applets makes it necessary for developers to have a 360-degree view of the distribution of user sources and observe the main channels of user sources in order to capture the characteristics of users from the source. At the same time, the depleted nature of miniapps leads to low retention rates, so developers focus more on acquiring new customers.
Activation analysis, multi-dimensional grasp of user usage.
Help developers in the fission process to find the user experience path and key business path, so as to do a good job of small program fine operation.
Sharing analysis, real-time view sharing backflow effect.
Help developers understand the information transmission path, the nature of the people covered, the number and distribution characteristics of Kols and KOCs, and target their portraits to help developers quickly find out where the potential audience is.
Transformation analysis, view user transformation results.
Identify core operational metrics that need to be optimized and improved by customizing events and conversion funnels.
Umeng + small program statistical analysis supports the mainstream small program platform in the market. Starting from the data analysis of each stage of user growth model, umeng + small program statistical analysis provides small program developers with a full-link analysis path of “customer acquisition – activation – share – transformation”. In addition to providing insight into basic business trends such as new additions, active user size and usage duration, it also helps developers have insight into customer base analysis.
In section 4/5/6/7 of this chapter, specific introduction and case interpretation will be carried out around small program acquisition analysis, activation analysis, sharing analysis and transformation analysis. In Section 8 of this chapter, detailed introduction and case interpretation will be carried out around refined user operation to help readers better understand the refined operation of small programs.
A guest analysis
As an emerging entrance under the mobile Internet, small programs have many advantages such as low start-up cost, saving users’ mobile phone memory, and gaining customers efficiently. They were once replaced by developers and users as low-frequency apps, and competing with apps for traffic and sharing traffic is an inevitable trend.
But on the other hand, small programs are not easy to acquire user base, user retention is more difficult. As a result, developers are putting more emphasis on attracting new customers, based on the applet run out user scenario. Wechat and Alipay, two large and small program platforms, play more ways to get customers by virtue of their respective advantages.
Different platform applets access differences
There is a very significant difference between applet raxin and App. Since small programs are mainly based on several “super apps”, new scenarios are closely related to several platforms. Each small program relies on its own platform advantage, in the small program battlefield running horse enclosure.
In order to improve the efficiency of customer acquisition, it is necessary to understand the differences between small programs on different platforms and find the entrance with the largest leverage effect.
(1) wechat small program
Wechat mini programs naturally trigger consumer demand with people and content as the core, and users are scattered in wechat groups, public accounts and moments of friends. Connect developers and users through small programs to form their own private domain traffic.
In wechat mini program promotion, the common ways to promote and gain customers are as follows: online paid promotion, self-owned other channels for drainage, offline scenes, sharing scenes, and various entrance traffic of wechat.
Alipay small program
In the promotion of Alipay, the general promotion method is similar, but at the same time, it has more self-operation entrance and linkage advantages of other Platforms of Ali, such as Dingding, UC and other platforms can jump through sharing links.
Alipay opens up all fields of social life through small programs. At present, the scenes mainly focus on the centralized distribution scenes such as search and conference. The decentralized scenes mainly focus on the exchange and offline scenes.
Baidu applet
It is mainly distributed through search engines and information flow.
Headline applet
For its various platforms of content services, the main focus on centralized distribution, entertainment, games, e-commerce.
QQ small programs
The users are mainly young groups, and both centralization and non-centralization are equal. The primary traffic distribution is through social fission, as well as recommendations for mini-programs and mini-game centers.
According to the above characteristics of various platforms, small programs to attract new customers mainly need to pay attention to several core scenarios:
Link and code: including all kinds of advertising channels, advertising platform and change quantity entrance, such as public account, small program, etc.;
Sharing: users share fission;
Scenario value: the centralized and non-centralized traffic source entrance provided by the platform;
Own matrix operation: own small program, public number, enterprise number and other matrix flow operation;
Offline scenario: scan codes for offline traffic scenarios.
Through the analysis of the corresponding five scenarios, developers can clearly understand the source and transformation of users, and have an all-round insight into users. Let’s take a look at how five scenario analysis tools work.
Scene 1: Links and codes:
In wechat and Alipay, online paid promotion and self-owned channels, as well as self-operating entrance, are generally promoted by adding links.
Because there are so many promotional activities and channels, it is important to generate custom links flexibly and analyze new users, active users, retention, and conversion through links.
Through the “promotion source” page of Umeng + small program statistics, developers can:
1. Quickly create multi-platform (Alipay, wechat) promotion links/codes, simple and convenient;
2. View real-time data of activities, new channels and active channels in real time, and help students to adjust activities and launch strategies in time;
3. View and monitor new, active, active quality and retention data based on activities and channels through complete data reports.
[Promotion Sources] includes two pages of “Promotion Activities” and “Promotion Channels”, each app supports 200 events and channels respectively. “Promotion activities” can be added by the way of activity links to monitor the data of promotion activities, support the generation of promotion TWO-DIMENSIONAL code and promotion path, can quickly view the Top5 activities of new users, understand the trend of all activity data changes, can also view the new, active, open and other detailed data of a single activity.
“Promotion channels” supports viewing the data of a single or multiple activities in the same channel. You can quickly view the Top5 channels of newly added users and understand the data trends of all channels.
Getting customers Scenario 2: Sharing
Sharing fission has the characteristics of free and fast transmission, relying on sharing fission to quickly acquire customers is the biggest feature of small program different from other channels.
There are two main forms of applets sharing:
Share page links or TWO-DIMENSIONAL codes of small programs directly on the platform.
Share content in the App to the platform, you can jump to the corresponding brand in the platform of the small program.
Customer acquisition Scenario 3: Scenario value
Wechat and Alipay have a lot of built-in scene values to help users open mini-programs. Wechat provides “recently used”, “search results page of the search box at the top of wechat home page”, “Advertisement in moments of friends”, “template message”, etc. Alipay also has a “Favorites,” “Add to desktop,” “Multiple operating channels,” and “Life.”
Further, by analyzing and comparing different scenarios, we can understand the active degree of users in different scenarios, so as to better understand the contact scenario and the correlation between the scene and activity, and effectively do the functional design and operation of the scene contact. Through the analysis of the scene value entry, we can intuitively understand the user from which touch points to contact the small program.
For example, in wechat applets:
If a large number of users come from “recently used”, it indicates that the small program is used frequently and the scene and user retention fit well.
If a large number of users come from the “search results page of the search box at the top of wechat home page”, it indicates that the search scene and natural traffic are well done.
If it comes from “small program message card in single chat session” or “small program message card in group chat session”, it indicates that sharing is the main touch point, and the product operation needs to produce more value and operation points of sharing.
In alipay mini program:
If a large number of users come from “Alipay Membership channel” and “Life Channel”, it indicates that the mini program is relatively related to the operation scenario of Alipay.
If a large number of users come from the “search results page at the top of the search box”, the search scene and natural traffic are well done.
If a large number of users come from the “mini program favorites application entrance”, it indicates that the mini program is used frequently and the scene and user retention fit well.
Further, by analyzing and comparing different scenarios, we can understand the active degree of users in different scenarios, so as to better understand the contact scenario and the correlation between the scene and activity, and effectively do the functional design and operation of the scene contact.
For example, here is the main scenario source for a shopping transaction applet. It can be seen that the average time spent by users in group chat is twice that of search and single chat, indicating that group chat is a scene with high user activity, and users have a stronger demand for “browsing”. And search with strong purpose, the user clear purpose.
“Real time” is also an important granularity in the operation of La Xin. Especially in the promotion of paid purchase volume, it can pay more attention to the flow and flow quality indicators at the hour or even minute level in a timely manner, and can also avoid buying traffic of lower quality and make immediate adjustment.
The “Scene Source” function page of Umong + small program statistics provides developers with all scene values of multiple platforms (Alipay, wechat), and supports the view of core traffic and active indicators such as the number of new and active users, open times, and stay time under multiple time granularity; Provide real-time data from all scenario sources to facilitate real-time observation of traffic, users, and active indicators after product launches or updates of major functions and activities. It also provides scenario value trend and comparison functions.
Customer acquisition scenario 4: Multi-matrix drainage
There are two aspects to the effect of self-owned matrix operation. First, pay attention to the flow effect of each small program exchange in its own matrix; Second, you need to focus on how active users are in the matrix.
Its own matrix each small program exchange flow effect, that is, the number of users brought by the matrix accounted for the small program daily flow ratio; And focus on the ratio of small programs flowing into other matrices.
Secondly, pay attention to the active status of users in the matrix: generally, pay attention to the average number of applications used by users, the average length of application used in the overall matrix, and the number of active days.
Customer acquisition scenario 5: Offline push
At present, offline scenes mainly include store code, merchant code, ground push, screen advertising, item code, etc. Similar to promotion, short-term focus on flow size and flow quality.
Channel evaluation method
The short-term effect of promoting new customers is mainly evaluated from two aspects: flow size and flow quality, in which flow quality is evaluated from shallow to deep by flow activity, flow transformation and flow retention.
(1) Flow size
Traffic volume mainly focuses on two core indicators: new users and active users;
New users: used to evaluate the ability of various channels or traffic sources to pull new applications;
Active Users: Active users include new users that are used to evaluate the channel’s ability to bring total UV to the applet.
(2) Flow quality
Flow quality from shallow to deep can be evaluated from three perspectives: active, transformation and retention.
Active: Common indicators include open times, page access times, and average access duration. These three metrics can be used to assess whether the users brought to the channel are using the app, rather than just popping it open and popping it out.
Conversion, generally according to different business characteristics, conversion indicators are also different, may be “watch ads”, “browse goods”, “add to shopping cart”, “order payment”, “complete the game level”, etc. You can focus on the number of core conversion events completed, the number of users, and the amount paid.
Retention, the next-day retention rate of new users in a channel, is generally used to measure whether traffic has a demand for continuous attention to the product.
For sharing, in addition to understanding traffic size and traffic quality, we also need to pay attention to the fission size of sharing. Fission scale focuses on the whole, users and content, and the main core indicators are the overall share reflux ratio, the share reflux ratio of core sharing users and the share reflux ratio of main shared content. The larger the number and frequency of sharing, the higher the sharing reflux ratio, indicating the larger the scale of sharing fission.
Activation analysis
People come into our app from a variety of channels, and then how the product is used, how active it is, how much retention it has, what content they prefer to access, that’s the activation analytics we’re looking at. Many small programs have been doing new and promotional activities, but lack of attention to activation or retention; As a result, the user growth is like “drawing water with a bamboo basket”.
We suggest analyzing users’ product usage from the perspectives of overall user activity indicator trend monitoring, user engagement, page access preference, retention, etc.
User trends
Monitor user access statistics in the past 7 days or 30 days and observe the following indicators: new users, active users, page access count, average stay duration, average stay duration, and average open count. Through these indicators, the user data changes in each stage are analyzed in detail.
New users: the number of users who visit the page of the small program for the first time.
Active users: the number of users who visit the page of the small program.
Open times: The total number of times a small program has been opened. The interval between the user opening the small program and the exit of the small program to the background is counted as one opening. When the interval between two consecutive opening of the small program is less than 30 seconds, it is counted as one opening.
Page access times: the total number of visits to all pages in the small program. Multiple page jumps and repeated visits to the same page are counted as multiple visits.
Average duration of stay: the average total duration of staying on the page of the mini-program each time you open it, the total duration of staying on the page of the mini-program of all users in the instant segment/the total number of times you open it;
Per capita stay time: the total time per capita stay on the page of the mini program, the total stay time of all users within the instant segment/the total number of active users;
Opening times per capita: the total opening times per person in the selected period, the total opening times of all users in the instant period/the total number of active users.
Participation analysis
By monitoring the average stay time, average stay time, average open times and other indicators, the stickiness and dependence of users in small program products are analyzed.
By stratification of participation, we can view the distribution and change trend of the number of people and times in different sections of the index in the current selected period, so as to analyze whether the product is healthy.
Page analysis
By analyzing the page view data of interview pages and entry pages in the past 1, 7, and 30 days, we found out which applets and entry pages attracted the most attention from users. There are two concepts that need to be distinguished here, the interview page and the entry page.
Interview pages are all the pages that the user browses when entering the applet. For example, the user enters the small program from page A and jumps to page B, where A and B are the interviewed pages. The number of visits, number of users and average visit duration can be used to know the pages that users use most and how long they use them.
The entry page refers to the first page that the user accesses when entering the applet. For example, the user enters the applet from page A and jumps to page B. A is the entry page, but B is not. The number of entry pages and the number of users help you understand the main pages from which users enter the mini program, and the bounce rate data can help measure the main page experience of users when they enter the mini program.
Retained analysis
Users who start using the mini-program for a certain period of time and continue using it after a certain period of time are considered retained users. The proportion of these users to the current users is the retention rate, which is one of the most direct indicators to judge the effect of attracting new users and user stickiness.
Retention analysis can be done from two perspectives: retention of new users and retention of active users. Retention analysis comes in two of the most common forms of graphical visualization, retention curves and retention tables.
The first is the retention curve, as shown in the figure above, which shows the retention trend of a user over the retention cycle. In general, the retention curve is decreasing, and we look at next-day, 3-day, 7-day, 14-day, 30-day retention.
The second type is retention tables, as shown in the figure above, which contain more information than the retention curve. Retention tables can look at retention rates of users who enter on different dates over subsequent retention cycles, but the information is less intuitive than retention curves.
Share analysis
Sharing fission has the characteristics of free and fast transmission, relying on sharing fission to quickly acquire customers is the biggest feature of small program different from other channels. Taking wechat applets as an example, there are two main forms of sharing:
The first is to share the page link or QR code of the mini program directly on wechat.
Page links can only be shared to wechat friends/groups, not to moments. Other friends click on the small program link, you can directly enter the small program, so as to achieve the small program to pull new or promote the effect of living. The QR code can be shared with wechat friends/groups or moments. Other friends can also access the mini program by scanning the QR code of the mini program.
The brand & service of small program is recommended and spread in the user’s circle, and spreads fission through sharing relationship, bringing more private domain users to small program.
Second, after sharing the content in the App to wechat, you can jump to the mini program of the brand.
For example, if you share the product content with your wechat friends in an e-commerce App, the e-commerce company will show the product content to wechat users in the form of small programs. Users can directly enter the e-commerce mini program, or jump to the e-commerce App. This can not only improve the small program to promote the new activity, but also improve the activity of the App.
Compared with wechat, alipay mini program has more ways of sharing and more diverse sharing channels: friends, Alipay friends, Dingding, Sina Weibo, wechat, QQ and so on. As shown in the figure above, different channels are chosen to show different forms of sharing.
Umeng + small program statistics module has three secondary pages: sharing trend page, user sharing page and sharing page. Below we will explain one by one, to help you better analyze and master their own small program to share the effect of fission and develop strategies.
Share the trend
Developers can see the overview, trends, and details of the most important sharing metrics for viewing applets in the selected time period.
For example, developers can view the trend chart of the number of users newly added by the mini-program every day, so as to discover the abnormal situation and the effect after the operation of the mini-program.
You can view the trend of the number and times of users sharing and the effect of sharing in any 1, 7, or 30 days. Sharing function is an important way for small programs to attract new customers, and an important indicator to judge the health of small programs.
Users to share
This page allows developers to view all user sharing and the fission effect of sharing.
For example, which users share the most, which users bring the most new users to the small program, and so on. Once you have identified the key KOL users in the sharing scenario, you can tailor your business strategy to those users to achieve better results with less money.
In the diagram data center above, the first user shared a lot, but didn’t bring any backflow or new users, indicating that the user was probably a fleece. The following user return and increase are very low, indicating that the operation effect of this activity is not good.
Share the page
The more times a page is shared, the higher the value of the page. For example, in e-commerce applets, each page Path corresponds to one item. Developers can see which items have been shared the most to find the most popular items.
Transformation analysis
In addition to the basic data of small program channels, new, active and other data, some developers are very concerned about the conversion of small program products, such as the conversion of small program orders. These transformation behaviors need to be analyzed through custom events and transformation funnels.
Custom event analysis
Custom events help you perform buried statistical analysis of the interactions of key events within applets. Displays the number of clicks and times of all events. You can also view the number and number of triggers of a single event, as well as the trend of attribute values.
The figure above shows the number of times a transformation event is triggered within a applet product, such as the number of successful orders.
For a specific event, we can upload properties to split the transformation event. For example, for the event of successful order, we can analyze the event trigger times, percentage and number of people under different commodity categories.
Transformation funnel analysis
Funnel can help developers understand the conversion behavior of users in the product. By analyzing the conversion or loss of each link, they can find the “obstacles” that hinder users from taking the next step, and then improve the conversion rate through product optimization or operation activities, and finally achieve the product goal.
In electricity market view, for example, through the establishment of “browse commodity – add to cart – submit orders – payment success” this funnel monitoring, can clearly understand each link of the transformation situation, how many people joined the shopping cart, how much is the conversion from the add to cart to submit orders, etc., and on this basis to optimize product or operational strategy adjustment, Guide the user through more transformations.
Alliance + transform funnel
Or take the new user registration scenario as an example. The user experience process is to click register – fill in information – complete registration. Through funnel analysis function, it is found that the conversion rate from filling in information to completing registration is low (lower than the industry average). Through user feedback, it was found that some users would fail to deliver verification codes, so they could not register for the next step. Finally, multiple SMS channels were added to improve the SMS arrival rate.
Refine user operations
Most small program operations in the early development of extensive flow operations, after several years of development, users were all kinds of head application play after cultivation, extensive operation began to gradually refine the way of operation.
Refined user operation is mainly to match different services and content to users with different needs, so as to meet their personalized user needs, so as to better complete the objectives of attracting new, activating, activating and transforming in operation.
There are two core problems in fine operation: 1. Identify user needs; 2. Match different content or service paths based on user requirements.
Three ways to identify the user needs of applets
The main methods to identify user needs are:
(1) Analyze user portraits
User portrait generally includes two parts: natural population attribute and user’s personalized portrait:
Demographic attributes are usually user gender, age, region, occupation and other demographic attributes;
User personalized portrait is divided into two types:
One is the behavior data of users in the small program, including login, browsing, purchase order, evaluation and sharing, etc.
One is the user’s interest preference, consumption ability and so on.
Here is an example: if the mini program to be operated is “female, 25-29 years old, second and third tier cities, faculty”. Based on these basic user portraits, we can select hot topics, hot information, education information and other related content and activities for operation.
Interest preferences can help promote collaboration with other applications. For example, if the main interest of the users of the application is “e-books”, you can consider adding more e-book products into the operation.
Research on users according to their preferences and group attributes can better optimize user activities, match operation documents and improve operation effects.
(2) Analyze the behavioral characteristics of different groups of users
Users drawn from different channels usually have different crowd portraits, which can be compared so as to guide users in different scenarios and channels with different content and service paths.
Again, this is an example of user activity data from part of the scenario of a shopping transaction applet. It can be seen that the average time spent by users in group chat is twice that of search and single chat, indicating that group chat is a scene with high user activity, and users have a stronger demand for “shopping”, while searching users have a strong purpose and clear purpose.
For users who enter through the scene value of group chat, they should consider “browsing” and “watching” functions such as shopping venue and live broadcast room after entering the mini program. For users entering the search scene, more behaviors such as “add, favorites” should be guided, so as to pave the way for the next user return.
(3) User research
User research is a commonly used way to identify user needs. Operated in an ecological environment, such as Alipay, wechat, QQ and Toutiao, users can be guided to pay attention by the ecological facilities of public accounts, life accounts and enterprise accounts, so that users can be easily reached and user research can be carried out.
Applet user touch
Once you have some insight and understanding of user needs, the next step is to match product content or service paths based on user needs. There are two types of core ground contact: the first is the contact in the application; Second, external contacts.
(1) In-application contacts
The primary goal of in-app contacts is to facilitate conversion or increase revenue, or improve the user experience. For example, the ability to present different features, pages, or page content to different users.
Examples of some usage scenarios:
According to the user set system language, display the welcome language of different languages;
Small game updated new levels, first visible to a small number of users, verification and then full release online;
Activities and festivals promote props within the time limit;
New players of the game free trial commemorative skin;
The opening and closing of advertisements;
For specific user groups, present different recall language;
AB tests are carried out in different groups of people with different attributes, and data operation is utilized to improve the application scientifically.
(2) station outside contact
External contacts of applications are mainly channels for active users to reach, such as SMS and push.
At present, wechat and other social channels have also become important user touch points. Small programs, especially wechat small programs, can better combine wechat service number and subscription number to form a better operation system with user experience. But in the operation, also need to do a good job positioning and mutual guidance. For example, small programs generally undertake trading, transformation and other functional scenarios, subscription number, service number notice to undertake the ability of active marketing; And the service number, subscription number and other user messages, can be used as a dialogue channel with users.
In an offline restaurant, guide users to pay attention to the service number and public number when ordering food; When users leave the restaurant, they can also recommend and notify the food through the service number and public account, and undertake online business scenarios such as food and take-out through small programs. In the delivery of goods, goods on the new, but also according to the user’s preference, user notification, enhance the small program active and transformation.
In order to reduce the use of cost, improve work efficiency. You can also provide “online parameters” through the umeng + small program product function, support to provide online parameter configuration, as well as real-time content delivery and display according to the conditions, without multiple editions.
In addition, it also provides the ability to unify the service number and small program user ID, set the crowd tag, select the crowd circle, set the service number message push, and provide product and technical ability for the small program developers to operate in an all-round way.
The above excerpt is from the “White Paper on Small Program User Growth”, which summarizes umeng + ‘s methodology, solutions and success stories in small program statistical analysis and refined operations, as well as the interpretation of many enterprises’ real practices.