3 common models of user operation!

This article introduces three common models. I hope it will be helpful to new operators .

1. About User Operation

Operations are a combination of various methods implemented to achieve business line goals.

Then the goals will be different, such as growth, monetization, short-term user growth , and maximizing the long-term user life value.

Of course, there are good and bad methods. Some are time-consuming but stable and budget-saving, some are fast but costly, and some are occasional explosive phenomena. Good operations are always a combination.

As for how to combine them, it depends on how many people, budget, resources, and the priority of various business indicators you have. The methods will be different if you pursue rapid monetization first or array growth first.

Among so many methodologies, the efficiency levels are different. In order to improve efficiency, we will introduce various models and optimization methodologies.

User operation is a methodology that aims to improve efficiency, differentiate and group different users, and implement differentiated operations.

Therefore, different classification methods of user operations, such as RFM, pyramid model, normal distribution, etc., are all choices based on different business objectives.

• RFM is more of a classification method based on optimization of transaction amount to improve user LTV.
• The pyramid model is a classification method based on the stable growth of multi-level data to form a healthy AARRR system cycle.
• The normal distribution, on the other hand, is more inclined to extensive operations, which is a classification method that uses limited resources to maintain long-tail customers when user resources are limited.

What are the three types of user operations classification and the operation strategies behind them?

2. RFM Model

The RFM model can be said to be a classic. Starting from traditional CRM management to user operations on the Internet , it can be said that it will never go out of date. After all, sales are the core of all businesses.

RFM represents three indicators of user consumption: consumption amount, consumption frequency and the time of the most recent consumption, to build a consumption model.

Monetary amount of consumption can be said to be the core of business and the most important indicator. Frequency of consumption is the number of consumptions within a limited time and is a reflection of user loyalty . Recency, the time of the most recent consumption, is an indicator of user churn and also an indicator of user memory of your brand .

These three factors can form eight combinations depending on their levels, including high/low consumption amount, high/low frequency, and recent shopping time. Of course, the settings of the amount, frequency, and distance of shopping time need to be related to the consumption cycle of the product itself.

After all, the timeline settings for users who buy home appliances and users who buy facial masks are definitely different.

Then, based on the classification of these eight dimensions, different operation strategies are implemented. For example, if the consumption amount is high, but the recent shopping time is far and the frequency is low, then what happened behind this group of users? How to recall?

For example, if the spending amount is low but the frequency is high and the shopping time is recent, how should we maintain such loyal customers? To keep?

This is basic maintenance operation based on RFM classification. According to key operation indicators, we select key user groups that need to be optimized and plan operation activities.

This is the construction of the user system, which is a combination of operational events that needs to be done in the long term to optimize operational indicators.

This is just the most general principle. If you go deeper, you will have to model it separately based on the user group, product characteristics, and product cycle, and you will need to analyze the best indicators in detail to carry out differentiated operations.

Similarly, we can also reversely filter out user groups in RFM based on different operational activities to improve the effectiveness of the activities.

When doing e-commerce operations , Tmall ’s data has a special RFM model for you to choose from, but it is a bit low-end, so it is better to model it yourself. Generally, during large-scale events, such as Double Eleven , the discount intensity and even the copywriting are different for users in different categories.

This is called reverse filtering of users based on activity.

3. Pyramid Model

As for operations, there is another pirate model that is very familiar to us, which is the A AR RR inverted funnel conversion model.

Users go from new additions, retention , activity, dissemination, to profitability. Each level represents the loss of some users, and these users are constantly evolving layer by layer.

Then, the pyramid model is a model that classifies users based on this process.

According to the user's behavior trajectory, from registration, downloading, using products, recommending, evaluating, purchasing and paying, to logging out, uninstalling, and losing, this is a complete process, and there are a number of users in each process.

Our core indicator, of course, is to get users to pay.

The purpose of stratifying these users is to carry out differentiated operations for each layer of users, allowing them to evolve step by step, or directly enter the conversion and purchase stage.

This is the user operation classification method of the pyramid model. Generally, we use the users of an APP as an example.

Of course, user behavior is not limited to these five steps. There may be more detailed steps. However, we generally classify them according to these typical five steps because each step of the online process is very representative.

Therefore, for these five levels of user classification, our operations have different requirements for each category: we hope that new users will download, so we give new users benefits; we hope that downloaded users will use the product, so we provide fool-proof operation guidance; and we hope that using users will come more often to build trust, so we need continuous stimulation and optimization.

As for interested users, of course we hope they will pay, so naturally promotion is a common means. As for paying users , we must provide good service, hoping that they will make a second or third purchase, or even recommend it to others.

Different strategies are adopted for users at different levels, which ultimately lead to key operational indicators. This is the core of user stratification.

Of course, this is just one way of stratification. There are others. For example, when operating resources are very limited under the RFM model, we will stratify through a single dimension.

For example, based on the amount of consumption, the pyramid can be divided into three levels: those who spend more than 1 million, those who spend 500,000 to 1 million, and those who spend less than 1 million.

For example, there are those who have consumed more than 10 times, those who have consumed 5-10 times, those who have consumed once, and those who have not consumed but only looked but not purchased.

This is also another pyramid model.

The purpose of stratifying users is, on the one hand, to achieve operational business indicators, and on the other hand, because operational resources are limited, efficiency must be maximized.

4. Normal Distribution Model

When the user's operation and maintenance resources are more extensive and very limited, the normal distribution model can be used.

For example, the 80/20 rule is a form of normal distribution, where 80% is near the mean of the curve, and the remaining 20% ​​is the source of profit.

Later, I worked as an agent for new media operations. When targeting B-side customers, basically all I had to do was maintain the best and the worst customers. As for the large number of customers in the middle, I just needed to implement standardized operations.

Therefore, the normal distribution model and the RFM model are just different modeling methods, one is two-dimensional and the other is three-dimensional.

The normal distribution model is a model built on two dimensions, such as profit contribution and number of people. You will find that the customers who provide the most profits are a minority, and the customers who make no money but instead lose money due to operating costs and emergencies are also a minority. Most of the profits that remain are around a constant value.

Based on these three situations, we can allocate operation and maintenance resources, focusing on maintaining high-profit users. At the same time, we must also pay core attention to customers who do not bring profits but occupy most of the company's resources and abandon them.

For the majority of customers in the middle who are near the constant value, standardized services should be provided to save resources and reduce marginal costs.

Of course, if you want to continue to refine your operations, you can continue to segment by variance.

For example, in the above figure, users within 1 standard deviation are standardized in service, while users between 1 sigma and 2 sigma are mined to see if users can evolve to 2 of the 28 rules, that is, between 2 sigma and 3 sigma.

As for users between -2 Sigma and -1 Sigma, they are profitable but still have difficulty in operating costs. Can we raise prices slightly, or reduce product standards, optimize processing, and thus increase profits to within the normal 1 standard deviation?

As for those outside -2 sigma, they are discarded.

This is the user stratification method of the normal distribution model. This model is often used in B2B business, while B2C is more likely to use the RFM model. As for the pyramid model, it is a simplified version of the RFM model.

Therefore, the user system is established to facilitate the optimization of different operational strategies and achieve the ultimate business indicator - profit.

Author: Eggshell Peanut, authorized to be published by Qinggua Media .

Source: Eggshell Peanuts