Establishing four strategic systems for user operations

What is the user operation system like?

I believe that every company has a relatively complete user operation system. I have read some articles introducing user systems before, which basically equated the user operation system with the user segmentation strategy and AARRR operation model. In fact, this is only a part of the entire operation system.

Combining the review experience during the operation process, I have explored a more practical user operation system in the process of community o2o user operation, which includes 4 major strategic systems:

Growth framework: user growth team + core growth channels + growth tools.

User modeling: User model construction, including label portrait, user value model, user preference identification model, user churn warning model, user activity model, etc.

Scenario-based stratification: 12 major operational scenarios. Each scenario is stratified and grouped based on user tags and modeling tools, and corresponding precision marketing methods are formulated for operation.

Data operation: core operation indicator system + data analysis model.

1. Building a User Growth Framework

The importance of user growth is self-evident. Without user growth, there is no way to talk about user operation. Before every company increases its user base, the first thing it thinks of is channel operations, and the channels are responsible for allocating operations personnel.

The general practice is: the marketing department recruits channel operations, who are responsible for the delivery and optimization of various app stores and online advertising platforms; new media recruits new media operations, who are responsible for content output of social channels; and the user team is responsible for the activation, promotion, and retention of users in the user pool.

At first glance, this team system seems to be completely fine. Everyone is responsible for their own part and is responsible for achieving their own KPI.

However, in the actual operation process, the following unavoidable problems may occur:

The boundaries of departments and the setting of KPIs will lead to poor coordination of the entire operation. Basically, each department is focused on achieving its own KPI. At the same time, channel operators often only assess the number of registered users. In order to achieve this KPI, the quality of users is artificially lowered, and fake users often occur, making subsequent activation, promotion, and retention work difficult to sustain.

In order to achieve KPI, the channel department has basically developed all paid channels. On the one hand, the core growth channels cannot be cultivated attentively, and on the other hand, CAC remains high.

Lack of effective growth tools, such as channel analysis system tools and offline community customer acquisition models.

Based on the above problems, the first task for enterprises to increase user growth is to build a user growth team. The growth team must first eliminate department boundaries and exist in the form of a project group or a growth department, including channel operations, event operations, products, and user operations.

Secondly, based on each operation node of AARRR, growth indicators are defined for each function to guide the entire growth work:

The main assessments of channel operations at the Acquisition node are: new users, acquisition cost (CAC), and new user retention rate.

The main assessments of products at the Activation and Retention nodes are: registration conversion rate and function retention rate;

The main assessments of activity operations at the Activation and Retention nodes are: DAU, MAU, and DAU/MAU;

User operations at the Revenue and Referral nodes are mainly assessed on: user conversion rate and K factor.

All nodes are coordinated in the form of project teams or growth departments, and ultimately only one departmental indicator is assessed at the KPI level. Each function is linked to this indicator, solving the problem of acting independently and blaming each other.

The second task is to establish core user growth channels. User acquisition channels generally include paid channels and free channels. The first goal of an enterprise in establishing core channels is to find channels with a low enough CAC. If the proportion of paid channels for user acquisition is high and the user acquisition cost remains high, then growth will be constrained by the promotion budget, and growth will stagnate when the budget is insufficient.

Secondly, the users brought by the core channels must be high-quality users. Many companies have obtained a large number of non-target users by means of brushing rankings. Although the growth data looks good, the conversion effect is very poor. This channel cannot be regarded as a core growth channel.

We will find that some products that perform well must have their own core growth channels. Mobike’s body QR code obtains enough riding users through offline placement, Didi’s red envelopes obtain enough taxi users through fission in sharing channels, and Pinduoduo obtains enough e-commerce users through social channels by selling good products.

As a community O2O platform, we are also building core growth channels by relying on stores and offline delivery personnel to promote daily necessities and life services in order to acquire enough family and elderly users in various communities.

Finally, there are growth tools. What are growth tools?

Growth tools are means that can help companies acquire users efficiently. They can be physical objects, analytical models, or coupons.

Mobike's growth tool is bicycles. By combining bicycles with users' short-distance travel needs, it has gained a large number of short-distance commuting users explosively.

Didi’s growth tool is the subsidy coupons, which convert a large number of marginal taxi users into actual users;

Pinduoduo's growth tool is to offer lower and lower prices, and convert a large number of Taobao and JD shopping users into group buying users by offering low-priced and good products.

Our growth tool is the development of an offline community user model. Through community portraits, user portraits, and big data modeling, we predict the needs and preferences of users in each community, label them accordingly, and guide stores to conduct targeted marketing in the community, converting convenience store and supermarket users into community O2O users.

In fact, as a customer acquisition tool, growth tools are closely integrated with the core business of the enterprise on the one hand, and can meet user needs on the other hand. Both are indispensable. If growth tools cannot be found, the company will not be able to achieve sustained user growth by relying on face recognition.

2. User Model Construction

If a company is unable to build a basic tag portrait model, user operations can only be done on paper. The establishment of a user model is the basis for achieving user stratification and grouping, and is also a necessary tool for precise user operations.

User models include label portrait models, user value models, user preference identification models, user churn warning models, activity models, etc.

The value of tags lies in helping operations personnel implement scenario-based stratification of users based on business and design targeted marketing activities. The value of portraits lies in helping operators understand the characteristics of each group; the user value model can identify high-value user groups; the preference identification model helps operators push products in a targeted manner; the churn warning model retains users before they churn, and the activity model can carry out targeted awakening and activation.

The construction of the model requires a dedicated data product team to complete. When operators conduct marketing based on user models, they need to focus on marketing effect analysis and iterative optimization of marketing plans.

Through multiple marketing attempts and the data product team, we found a more suitable way to build the model and gradually established a stable operation plan.

Every day at work, operations staff can combine the tags generated the previous day into marketing information for user groups and send it out (push or SMS), monitor its conversion status, continuously iterate, and gradually establish standard operation plans and programs based on user models.

3. Scenario-based Layered Strategy

Several operation scenarios can be derived based on the platform business. Different user groups need to be operated in each scenario. The user groups come from the label model and various user models.

In our specific operation process, operations are divided into two categories: one is growth hack, and the other is user-refined operation.

The two types of operations are divided into 12 major scenarios. Take one of the operation scenarios as an example:

Business scenario: The user repurchase rate of a channel on the platform is low. The channel operator suspects that there is a serious user churn. He hopes that the user department can help monitor user churn and predict which existing users may churn? Formulate corresponding retention strategies through churn warning.

In combination with this business scenario, we will filter out users tagged with the xx channel in the label system, and train the churned user samples through the user churn warning model. The training method is described in the previous article "Community O2O User Operation: Using the Laundry Channel Practical Case to Teach You How to Build a User Churn Warning System". The model can be used to find out the characteristics of churned users, calculate the churn scores of users with different characteristics, and group users according to the churn scores.

Specifically, they can be grouped into low-risk churn users, medium-risk churn users, and high-risk churn users. The low-risk user group can maintain the status quo and carry out daily push marketing. For the medium- and high-risk churn user groups, it is necessary to combine the user portrait system and user preference analysis model analysis to determine the reach strategy.

For example: analysis shows that the proportion of females in this user group is relatively large, the community attributes are mid-to-high-end attributes, and they prefer to buy imported fruits and high-end laundry. At this time, you can use this to formulate a retention strategy, and push corresponding activity information targeting the female group to these users to successfully awaken the users and achieve the purpose of retention.

4. User Data Operation Strategy

Data operations include a core indicator system and a data analysis system. The core indicator system can monitor the development trend of user operations and understand basic information such as user activity and health in real time. The user data analysis system can help operators locate problems and optimize products in a timely manner based on the problems.

The first step is to build a core indicator system, and the core indicators must be closely integrated with product goals. For example: The goal of bicycle products is to obtain rental income, and its core indicators should be built around paying users. The goal of information products is to generate traffic through user reading, and its core indicators should be built around DAU, browsing depth, and duration.

At the same time, people at different levels within the company have different focuses on core indicator data. The leadership level focuses on the overall user volume, cost, and revenue; the operations level focuses on user activity, retention, and conversion. In the construction of indicator system products, we build around the core indicators of consumer users from four dimensions: new customer acquisition ability, health, preference, and purchasing behavior.

1. New customer acquisition capabilities

User growth potential analysis: Cities, stores, and field sales personnel understand the overall situation and development potential of regional, business district, and community user development;

User source channel analysis: Each channel wants to know which channels are currently being promoted and from which channels do users mainly come? Which channels are high-quality, so as to optimize channel strategies;

New product acquisition analysis: Stores and sales personnel want to know which product in the area contributes the most to new customers. The product that customers order for the first time is defined as a new product.

Analysis of new customer preferences in each community: Store and sales staff want to know the preferences of new users in each community in the area. For example, community A prefers electronic products, community B prefers fresh food, so that they can carry out targeted promotions when attracting new customers in each community.

2. User health

User value analysis: If the channel wants to know who its loyal user base is, it can find these high-quality users and invite them to participate in activities. Similarly, offline sales staff can invite these users to participate in activities in stores.

User churn index: The channel wants to know which users in different groups will churn and how to prevent them from churn;

Community user contribution: Stores and sales personnel want to know the GMV contribution rate of each community in the area where the sales personnel are located, divided into weeks and months, and there must be a trend chart for the distribution within the area.

3. User Preference

Category preference: Stores, field sales personnel, and channels want to know which neighborhoods/regions are more inclined to consume what types of goods (the cross-relationship between buyer location and category);

Activity preference: stores, field sales personnel, and channels want to know which neighborhoods/regions prefer what type of activities (the cross-relationship between buyer location and activities);

Price preference: The channel wants to know what prices users of different categories prefer, so as to push products of various price ranges to the corresponding users (the cross relationship between category and price);

Touchpoint preference: Stores, sales personnel, and channels want to know which channels users of different categories prefer to purchase through (the cross-relationship between categories and touchpoints).

4. User purchasing behavior

Repurchase rate of different user groups: The channel wants to know the repurchase rate of new and old users and find out the high-repurchase products, adjust the operation strategies for new and old users in time, do a good job in product operation, and monitor it on a monthly basis.

User path analysis: The channel wants to know the user participation from the channel homepage to the activity page, and at what stage the users are lost, so as to improve the page operation.

The second is the data analysis system, which requires the construction of a series of analysis model tools to help operations personnel locate problems in the operation process. The model tools include funnel analysis models, attribution analysis models, micro-conversion analysis models, cohort analysis models, etc.

Common analysis scenarios include: How to attribute the decrease in DAU? How to attribute the low registration conversion rate? The retention rate of new users is low. How to attribute it?

Taking the low registration conversion rate as an example, let’s briefly describe the analysis method:

Step 1: Decomposition of impact dimensions;

Step 2: Disassembly of subdivision indicators under dimensions;

Step 3: Identify the problem.

The registration conversion rate can be divided into two major influencing dimensions: channel and product.

Segmented indicators are broken down under each dimension. Channel segmentation indicators include delivery media, advertising type, advertising content, and keywords; products include registration logic, product design, input method, product stability, etc.

To locate the problem, it is necessary to check the detailed indicators one by one and find data anomalies. For example, check the conversion rate of each link through the funnel, and focus on the links with low conversion rates. If it is a channel problem, optimize the media, conduct AB testing on the advertising content, and accurately locate the keywords. If it is a product problem, optimize the registration logic and interface, and improve the stability of the APP.

By summarizing the four major strategic systems, we can find that user operation is no longer a simple matter of finding a few operators to do segmented operations, nor is it a job that can rapidly increase the user value of an enterprise by using a few user models. Instead, it is an operation system that an enterprise needs to invest long-term manpower, energy, and material resources to build.

The significance of user operation to an enterprise is self-evident. The growth of an enterprise's overall performance is inseparable from the expansion of the scale of high-quality users, and more importantly, the improvement of the user's life cycle value.

Author: Zhao Wenbiao

Source: User Operation Observation (ID: yunyingguancha)