How to build a refined operation system?

As operations enter a refined stage, how should we identify users? How to formulate an indicator system? How to understand the needs of users at different levels and design different operation strategies based on their needs? The author of this article shares methods on how to build a refined operational indicator system from 0 to 1. Let’s take a look.

Before we start to do refined operations, we must first do some data analysis to guide our operation strategy. So when we analyze, which dimensions should we start from? First, we must think clearly about the following three questions:

1. What is the goal of our analysis?
2. What data do I need to analyze?
3. How can this data help me?

Before preparing to conduct data analysis, it is recommended that you first link the analysis with your own business, and this connector is the analysis target. The marketing team hopes to get more sales leads, the operations team hopes to get more users and higher transaction amounts, and user operations hope that users can visit frequently.

At different stages, we will have different analysis goals, which also guide the framework and process of our entire analysis. Before conducting formal data analysis, we need to establish a series of data indicators, which is the data operation indicator system we will discuss in this section.

Just as there are no two identical leaves in the world, there are no two sets of indicator systems that are exactly the same. Different product types, different product development stages, and different ways of thinking of the data analysis teams will all lead to differences in the established indicator systems. However, the basic ideas and methods for developing an indicator system can be traced. The following methodological framework can quickly identify analysis goals and develop a data operation indicator system suitable for one's own products and services.

1. How to define the analysis objectives?

Analysis goals are closely related to KPI. Data analysis and insights actually have many application scenarios. The most common use is to improve our performance, or we use data analysis when we want to improve KPI performance. What determines your performance? It could be sales leads, order volume, or the number of newly registered users. This measurable performance is the goal of our data analysis. For example:

• The purpose of an e-commerce platform is to allow users to complete more transactions on the platform, and the KPI and data analysis goal of the platform manager is to increase the order volume.
• ToB enterprise service websites hope to obtain more registration leads, and the KPI and data analysis goals of website operations are to increase the number of registrations.
• Banking apps hope to attract more users to purchase financial products, and the KPI of app operation is to increase the total purchase amount of financial products.

2. How to formulate an indicator system?

The conventional indicator system includes the first key indicator, first-level indicators, second-level indicators, etc.

The first key metric, also known as the North Star metric, is the metric most relevant to our data analysis goals.

When we start to analyze data for a product (website, APP, mini program, smart hardware, etc.), many data indicators will be involved, but only the most important core indicators can be called the first key indicators. The characteristic of the first key indicator is that it is directly related to the goal. Our work revolves around promoting data changes in the first key indicator, and these data changes will also help us achieve our goals. For example, the number of new registered users on a website every day is closely related to the goal of acquiring new users, so we can use the number of registered users as its first key indicator.

It should be noted that although the first key indicator is the most important, it is not the only one. For example, for e-commerce websites, we pay attention to the order volume while also paying attention to the number of new user registrations. Moreover, the first key indicator is not constant and will change with business changes. For example, in the early stages of our product, we paid most attention to attracting new users. After accumulating a large number of users, we will increase our attention to user retention. At this time, the first key indicator may be weekly active users (WAU) or monthly active users (MAU).

How to judge whether an indicator is the first key indicator? One criterion is: if we improve this indicator, will the long-term performance of the product be improved? Examples of common first key indicators on websites or apps: corporate service websites: number of registration leads; e-commerce retail and other general transactions: number of successful payments; categories that waste user time and attention: DAU or average stay time.

1. Primary indicators

Primary indicators refer to a series of indicators that directly contribute to the first key indicator or can help the product develop in a better direction.

For example, the first key indicator of a website is the number of registrations. The first-level indicator may be the number of visitors to the form page. Increasing the number of visitors to the form page can directly increase the final number of registrations. The first-level indicator may also be the conversion rate from form page visitors to successful registrations. Increasing the conversion rate from form page visitors to successful registrations can directly increase the number of registrations.

2. Secondary indicators

Secondary indicators can directly contribute to primary indicators, or they can be a series of indicators that can help products move in a better direction.

For example, the primary indicator of a website is the conversion rate from form page visitors to successful registrations, and the secondary indicator may be the number of times the first field completes verification. Under the same number of visitors, the more the first field is completed, the higher the conversion rate from form page visitors to successful registrations.

We can develop a multi-level indicator system, but one thing that must be made clear is that each level of the indicator system developed will have a direct contribution to the indicators at the previous level. We recommend that you simplify your indicator system as much as possible and think clearly about what data you want to analyze. A simple and controllable indicator system is of great help in data analysis and allows us to focus more on core analysis.

It is important to figure out which analytical indicators are the most important and spend your analytical energy on those that can make your product better and provide a better user experience. If the entire team jointly develops a data operation indicator system, it is necessary to communicate the requirements in advance to ensure that each team doing data analysis can understand the meaning of the current indicators.

3. Management indicator system

After the indicator system is completed, it is necessary to ensure that each level of the indicator system is directly related to the previous level. For example, the change of the secondary indicator can affect the primary indicator, and the change of the primary indicator has a direct contribution to the first key indicator, as shown in the following figure:

4. Classification of key indicators

When a visitor becomes a user, they will go through a user life cycle (Custom Journey) from active to lost. Of course, not all users will go through the complete user life cycle, and users may leave us at any stage. One thing we have in common when doing product and user operations is that we hope users can participate in the product as much as possible, and we also hope that users will return to visit as many times as possible.

No matter what type of product you have, there is a typical set of user life cycles, and you can build an operational indicator system around the life cycle. The user life cycle mainly includes five stages: contact, conversion, activity, engagement and retention.

1. Contact refers to the total number of users who reach the website or APP from external traffic channels. It is mostly used in the stage of acquiring new users. It indicates the maximum value of your users. If it is a website, it is related to UV, and if it is an APP, it is related to launch. During the user contact period, the data indicators we focus on should contribute to the key indicators of this stage. Usually we need to know the composition of visitors who arrive at our products within a fixed period of time. At this time, traffic channels will be involved. Understanding the user composition status of different channels will help us optimize channels and improve visitor quality.
2. Conversion has different meanings in different applications. If it is a website or app that retains information, conversion refers to registration; in the e-commerce industry, conversion requires two steps: registration and payment of the order, so registration is not considered a true conversion.
3. Active users refer to people who have taken actions and gained value from the product within a period of time. Whether it is daily active or weekly active, it reflects the user's level of engagement with the product. How to define activity? Just like conversion, different products define activity in different ways. For a website, opening the website again within a certain period of time is the key behavior of activity; while for the e-commerce industry, multiple purchases are a typical manifestation of user activity.
4. Participation Engagement refers to the proportion of users who have completed certain key behaviors among all active users, which is used to evaluate the user's participation in the product. The degree of participation means the product's stickiness to the user.
5. Retention rate reflects the stickiness of the product. When defining the retention target, you will get the corresponding retention list. The retention target can be product opening or a certain function (for example, whether users who use function A will come back to use function A within the next 7 days). Such indicators reflect the function retention well. It is generally recommended to use 7-day retention rate, 30-day retention rate, and 90-day retention rate for retention analysis.

5. Develop your metrics system

Developing an indicator system is an important step in starting to analyze user behavior/data-based operations. First, clarify the first key indicator, then clarify the first-level indicator to contribute to the first key indicator, and clarify the second-level indicator to contribute to the first-level indicator. In the process of formulating the indicator system, try to ensure the principle of superior contribution of the indicator system. Of course, there may be some special business indicators outside the system. As long as the indicators are helpful and valuable to making the product better, we can also put them into the indicator system.

In order to have a clearer understanding of the construction of the indicator system, we take the most daily e-commerce industry scenario as an example. The actual shopping process is much more complicated than the following example. We will use the simplest process as an example:
Step 1: Browse the product details page; Step 2: Add to cart; Step 3: Submit order; Step 4: Pay for the order.

Then we can start to develop the indicator system based on this

The first step is to define the first key indicator. The first key business indicator of most e-commerce companies is defined as the number of orders with successful payments. The business goal in the form of selling goods is to sell more goods. The entire shopping process indicator system of e-commerce companies such as Taobao and JD.com is very complicated and difficult to understand. Here we only select some common basic indicators.

Next, clarify the first-level indicators in the indicator system. At this time, you need to think about what reasons make users willing to pay for more orders?

Looking from the back to the front in the entire shopping process, choosing a payment method is a necessary step in the payment order link. This data can be used to understand whether different payment methods affect payment. You must look at the number of payment failures and find out what caused the payment failure; you also need to look at the number of clicks on the Submit Order button and the number of users who successfully submitted orders. The more items added to the shopping cart, the more likely it is that a payment order will be placed. So you also need to look at the number of clicks on the Add to Cart button and the number of cart views. The shopping cart reflects the potential purchasing behavior and the user's shopping preferences. In addition, the number of clicks on the confirmation button is also important. In the shopping cart, confirmation means entering the order submission page.

From browsing product details to paying for an order, we also need to focus on the three conversion rates in these four links (the conversion rate from browsing products to adding to the shopping cart, the conversion rate from adding to the shopping cart to submitting an order, and the conversion rate from submitting an order to successful payment). Improving these three conversion rates will directly contribute to the final number of successful payment orders.

The secondary indicators are the blue indicators seen in the above picture, such as the number of specification clicks and quantity clicks, which can promote the achievement of the primary indicator of adding to the shopping cart. Although product collection, review checking, etc. will not directly contribute to the number of successfully paid orders, they will directly contribute to the act of adding to the shopping cart, which is also the indicator we need to focus on.

The above is a simple data analysis indicator system for e-commerce shopping scenarios. By determining the indicator system, we can use the existing user behavior analysis/intelligent user operation platform on the market, and build various data indicators that we need to know at any time through the dashboard. We can then get a data analysis dashboard that is online at any time, updated in real time, and collaboratively shared.

After building a data operation indicator system, you should consider how to obtain relevant data on user behavior such as the number of clicks. By adding code to your website or APP, whether it is visual or coded, you can complete a complete data-based operation closed loop from collecting, organizing, calculating, and using numbers through the tool platform related to intelligent user operations.

Author: Jiuxianqiao Ishihara Satomi

Source: Jiuxianqiao Ishihara Satomi