How does Xiaohongshu build a user rating system?

1. What are the user upgrade standards (parameters)?

When we design the user level mechanism, we first consider which parameters need to be met when upgrading. Is it the number of posts? Order amount? Or the number of monsters killed?

First of all, we must be clear about the purpose of user levels: through the psychological sense of honor and material rewards created by level rights, we can motivate users to perform key behaviors expected by the platform.

Therefore, when formulating user upgrade parameters, we need to break down the user's key behaviors, assign weights according to their importance, and use this as the upgrade standard.

Taking Xiaohongshu and Taobao as examples, let’s look at the differences in user upgrade standards on the two platforms due to differences in user key behaviors.

1. Xiaohongshu

Xiaohongshu’s slogan is “Mark your life”, and the word “mark” reflects the key user behavior advocated by Xiaohongshu - posting notes. Therefore, the only criterion for Xiaohongshu users to upgrade is “posting a certain amount of notes of certain quality”.

In addition, it can be seen that since the content development trend is more inclined towards video format, you can upgrade by publishing video notes which are far less than the amount of text notes when upgrading.

This is Xiaohongshu using user levels to guide user behavior, that is, "first split, then assign weights" to key user behaviors. After completing these two steps, the user upgrade standards of Xiaohongshu will be clarified.

2. Taobao

Taobao's user level is represented by the numerical value of "Taobao Value", and the standards for obtaining Taobao Value are divided into "Purchase Products", "Comment on Products", "Complete Information", and "Family Rewards".

It is obvious to see the difference between Taobao as an e-commerce platform and Xiaohongshu as a content community: Taobao uses the purchase and evaluation behavior generated around the products as the upgrade standard.

From the comparison between Xiaohongshu and Taobao, we can see that one of the main functions of user levels is to guide and strengthen key user behaviors advocated by the platform through different user upgrade standards.

2. Should we downgrade?

One of the reference factors for whether to add a downgrade mechanism to the user rating system is whether the rating parameters "continue to create value".

Taking Xiaohongshu and Baidu Tieba as examples, these content platforms are rarely downgraded because users “have not performed key actions for a period of time”, because the content historically produced by users (which is also the level parameter of the content platform) always exists on the platform for browsing, and the level parameters continue to create value.

Only when the user deletes the content will the level be downgraded accordingly, because the level parameters disappear and value can no longer be created.

It is also worth mentioning that although Xiaohongshu does not have a formal downgrade mechanism, the system will automatically downgrade high-level users who have not been active for a period of time, but the level can be restored by appealing.

The purpose of "downgrading" is definitely to control the proportion of high-level users, make high-level users rarer, stimulate users to upgrade, and avoid waste of rights and interests; while "appeal to restore level" is to screen out the real lost users, because the subsequent active users can use it normally after appeal, while the real lost users will not appeal.

Xiaohongshu uses this non-explicit downgrade mechanism to obtain a more flexible user level control method.

Let’s take a look at JD.com’s user level mechanism: high-level membership is valid for 1 year, and after 1 year, a certain growth value will be deducted and the level will be recalculated.

In other words, if users want to maintain or upgrade their level, they must constantly complete key user behaviors (buy, buy, buy) to obtain growth value.

JD’s reason for doing this is obvious: what you bought before is what you bought before, and if you don’t buy now, there will be no continued value creation, so if you don’t want to downgrade, you have to continue buying.

This is the difference between JD.com as an e-commerce platform and Xiaohongshu as a content platform in terms of "whether upgraded parameters continue to create value."

3. Two algorithms for upgrading and downgrading

Regardless of the product type, users' upgrades and upgrades need to see whether the value of the level parameter triggers the level boundary value. The setting and calculation methods of the boundary value are divided into the following two methods:

1. Cumulative system

That is, "the value of the accumulated level parameter triggers a certain level boundary value, which can lead to upgrading/downgrading."

Taking Xiaohongshu as an example, when the cumulative number of published notes reaches the fixed requirements of each level, it will be upgraded; when the cumulative number of notes after deleting notes reaches the value of the previous level, it will be downgraded.

The advantage of the cumulative system is that the system rules are relatively simple, but the disadvantage is that it is impossible to downgrade users based on the completion of key behaviors within a certain period of time:

Xiaohongshu uses a typical cumulative system. To reach level 10, a total of 800 notes must be published. However, it is not possible to require users to continue publishing content after the upgrade to maintain the level. For example, 50 notes must be published within a year or they will be downgraded. This requirement is impossible to achieve under the "cumulative system" because the system is always judging whether a total of 800 notes have been published. Assuming that the level is forcibly downgraded to level 9, the system will immediately restore it to level 10 because the user has already published 800 notes.

If you need a downgrade and relegation mechanism that allows users to continuously complete key behaviors, I recommend the "stage system" below.

2. Phased system

To put it simply, the stage system converts the accumulation of key behaviors in the entire life cycle in the "cumulative system" into the accumulation of key behaviors in each time stage. The accumulations in each time stage do not affect each other, and the rules for demotion, retention, and restoration of levels can be flexibly added.

For example, I modified the requirements on the level mechanism of Xiaohongshu:

• The highest user level is 8, and lower levels can only be upgraded but not downgraded;
• High-level users need to ensure a certain amount of notes published each year to maintain their current level, otherwise they will be downgraded;
• After being downgraded, you need to complete simpler tasks to restore your level.

Applying the "stage system", my solution is:

• All levels are upgraded instantly according to the fixed target difference in the number of notes (for example, level 1 to level 2 requires 3 notes to be published, level 2 to level 3 requires 10 notes to be published... instead of the "cumulative system" that requires 13 notes to be published for level 3).
• There is no downgrade from level 1 to level 5.
• For level 6 and above, the promotion and downgrade date will be recorded and updated after promotion and downgrade.
• If the note reaches the target difference for the next level within one year, it will be upgraded immediately and the upgrade and promotion dates will be updated.
• If the number of notes in a year is greater than 50 and is less than the target difference for the next level, the current level will be retained.
• If there are fewer than 50 notes in a year, you will enter the level recovery channel to determine whether your level should be downgraded to level 5 or level 6 or above.
• If you are downgraded to level 5, if you complete 10 notes within one year, the level will be immediately restored to level 6, the promotion and demotion date will be updated, and you will be promoted or demoted according to the original promotion and demotion rules; if you do not complete 10 notes within one year, you will not be downgraded.
• If you are downgraded to level 6 or above, if you complete 10 notes within one year, the level will be restored immediately, the promotion and demotion date will be updated, and the promotion and demotion will be based on the original promotion and demotion rules; if you fail to complete 10 notes within one year and are downgraded again, it will be determined whether the level after the downgrade is level 5 or level 6 or above, and the rules 7 and 8 of these rules will be followed respectively, until you are downgraded to level 5 and cannot be downgraded and remain unchanged, or the level is restored to normal promotion and demotion.

The flow chart is as follows:

It can be seen that the "stage system" combines the time stage with the target difference of each level, and adds the mechanism of downgrade, level retention and level recovery compared to the "cumulative system", which allows users to continuously complete key behaviors within the cycle.

4. User Level Incentives

In addition to establishing user promotion and demotion rules in the above-mentioned grading system, it is also necessary to build different levels of rights and rewards to incentivize users, so as to form a user grading system. User level incentives are divided into two parts:

1. Status symbol display

This is relatively easy to understand. Different designs are given to user level names and level images, providing users with tools to show off, allowing users to compete with each other, and continuously generating key user behaviors encouraged by the platform.

Taking Xiaohongshu as an example, the level names are based on the image of "potato" and are named in combination with the growth process. At the same time, they are accompanied by cute level images, giving users the desire to pursue higher levels.

2. Strengthen closed-loop product rights

As the level of Xiaohongshu increases, users can obtain more personalized watermarks, more background music libraries, longer video notes, higher note exposure permissions, and more.

These rights and interests all apply to the key user behavior of the Xiaohongshu product - posting notes.

At the same time, a closed-loop growth flywheel is formed: the more notes I publish, the higher my level and the more rights I get, which helps me publish high-quality and highly exposed notes.

The same is true for Taobao. As your Taobao value increases, you can get more discounts, more coupons, more red envelopes, etc.

These rights and interests all affect the key user behavior of Taobao products - shopping.

It also forms a closed-loop growth flywheel: the more things I buy, the higher my naughty points become, the more benefits I get, the more discounts I get, and the more I want to buy.

Therefore, when designing level benefits, we are not simply rewarding users, but giving users rewards that are in line with product positioning and can form a growth closed loop, so that the upgraded benefits can further stimulate key user behaviors.

V. Conclusion

The user rating system is one of the important means for products to motivate users' key behaviors from the bottom up and increase stickiness and activity. It is necessary to lay a solid foundation at the beginning of product design.

Author: Qiu Teng

Source: Qiu Teng