How to build a product data indicator system?

Today we will focus on sharing with you the North Star Indicator and the Context Model.

1. North Star Indicator

When talking about growth hacking, we have to first understand the most important indicator in growth, the "North Star Indicator".

The North Star Metric, also known as the One Metric that Matters (OMTM), is the most critical metric for the product at this stage.

Simply put, it is the development goal set by the company. Of course, the company will have different goals at different stages, so the North Star indicators at different nodes are different.

Why is it called the “North Star” indicator? The general meaning is to guide the direction of the company like the North Star.

The role of the North Star indicator is mainly reflected in:

• The company's development direction, all work starts from this.
• Help company employees clarify task priorities to avoid being at a loss in company operations and being unable to concentrate on key points.
• Improve your ability to take action. Once you have chosen your goal, there is only one thing you can do: work hard to achieve it.

The North Star Indicator is different at different stages of the product, and it will take a while to monitor it after it is established.

Moreover, there may be more than one good North Star indicator. The company's products also need to be tested in business, and the growth model can be broken down in many ways based on different situations such as department structure.

So how do we determine our North Star Metric?

There are two common methods:

• Determine the North Star Indicator through the "Growth Closed Loop Diagram" of User Core Value and Business Goals
• Use the North Star Indicator Checklist to screen your North Star Indicator.

In comparison, the former is more core, and the latter is an additional self-checklist added based on the former’s inability to determine the North Star indicator.

Note that when screening, you must first confirm the product life cycle in which the product is located before defining the North Star Metric.

Pinduoduo’s growth closed loop diagram

North Star Indicator Self-Check Framework

Note: When you self-check, you can select three specific indicators and put them on the problem for analysis. If you meet the requirements, write Y; if you do not meet the requirements, write N; if you are not sure, write ? , and finally select an indicator that contains all Ys. That is your product’s North Star indicator.

In addition, we should note that the North Star indicator can only evaluate partial situations and cannot cover the entire business.

After all, everyone has a different understanding of the product development stage, which can easily lead to differences in the North Star indicator.

Also, for some companies or development stages, the North Star Indicator is not perfect. When the North Star Indicator does not meet all six questions in the self-checklist, you need to be vigilant in your business.

After determining the North Star Metric, we can then start building a “Growth Model”.

In essence, the process of building a growth model can also be understood as the process of breaking down the North Star indicator.

There are three main growth models: full-chain funnel, factorization, and fully quantitative.

Here we focus on introducing two more commonly used growth models: the full-chain funnel type and the factor decomposition funnel type.

So some people may ask, why do we need to build a growth model?

The reason is simple. Because there are often many relevant factors that affect a thing, we need to break down all the relevant factors that affect the North Star indicator, and then control the variables like doing an experiment, so that we can find the key factors that affect the North Star indicator.

To give a simple example, when our computer freezes, everyone may think it has crashed, but there may be many factors that may actually cause the computer to crash.

For example: Is there a problem with the system itself? Low on computer memory? Computer infected? Bad contact on the mainboard? etc. . .

These are all the reasons that may cause computer crashes, so if we want to solve the problem of computer crashes, we must break down and analyze each factor one by one, so that we can find the key factors and solve the problem in the first place.

Full-link funnel model

The first step of the full-link funnel model is to determine the North Star indicator, and then we need to draw the user conversion path. Break down your North Star Metric into the product of your individual conversion paths.

Full chain funnel disassembly

For example, for Pinduoduo, assuming that the North Star metric is GMV, the typical conversion path includes: the core paths of users downloading the app, visiting the app, placing the first order, and continuing to place orders.

Pinduoduo full chain funnel

The full-link funnel model is a multiplication model, while the factorization model is an additive model, which splits the North Star into the sum of several sub-indicators.

Factorization

Taking Graphite Documents as an example, assuming that the North Star Indicator of Graphite Documents is the number of paying users, it can be divided into the number of paying users of enterprises and individuals, and then broken down separately

Graphite Document North Star Index Factorization

When using factor models for decomposition, there are two points to note:

• In the process of disassembling the growth model, we must pay attention to ensuring that the logic of the decomposition conforms to the MECE principle.
• When disassembling the multi-factor decomposition model, we can first break it down from new and old users, and then gradually break it down into specific conversion channels.

Finally, based on the above disassembly, we can summarize it into the following process:

• Defining a North Star Metric
• Draw user conversion paths
• Assemble the growth model.

2. Context Model

After determining the North Star Indicator, how do we then establish an indicator system that can fully reflect the development of the company?

We can analyze and monitor the specific situation of the entire company by establishing a detailed analysis system, which I like to call the "context model."

The problem that the context model essentially solves is to establish an indicator system that can monitor the entire chain from the time a user notices your product to the time a user churns. It is actually a breakdown and refinement of the North Star indicator.

First, let me explain the relationship between the North Star Indicator and the context model.

The role of the North Star indicator is mainly reflected in:

• The company's development direction, all work starts from this.
• Help company employees clarify task priorities to avoid being at a loss in company operations and being unable to concentrate on key points.
• Improve your ability to take action. Once you have chosen your goal, there is only one thing you can do: work hard to achieve it.

The role of the context model is mainly reflected in:

• Get instant insights into your company's health for quick analysis.
• It allows management to quickly understand the parts they are responsible for and quickly locate where problems need to be solved.
• Help employees understand the operating status of the product and maintain it as well as possible.

The North Star Indicator is like when we drive from A to B, there are countless paths, but the path that is most suitable for you is the North Star Indicator. The "context model" is like the GPS in your car, helping you monitor whether you deviate from the planned path.

If the North Star is the indicator that can guide people forward in the dark, then the "context model" is the GPS that monitors your deviation from the direction.

So how does the “context model” work?

The context model is built based on the idea of ​​the growth funnel, so our first layer is to use the AARRR model to decompose the indicators of each link, so that no link can be missed in the analysis.

Then, by modeling the AARRR model and further splitting it into secondary categories, it can be divided into product, operation and marketing categories. Here we can find that the data for different service categories are different.

When we want to drill down to analyze one of the indicators, such as product category, we can also divide it into quantitative and qualitative indicators.

AARRR and context model breakdown

Breaking down qualitative data is more for analyzing the basic attributes of users’ portraits, while when analyzing quantitative indicators, it is more about analyzing the user’s quantifiable data from multiple angles to derive the relationship between the individual user and the entire product.

When multiple businesses are classified (only applicable to product analysis here), of course, the four major modules of the product are disassembled and analyzed in combination with quantitative data. Further subdivision can also be combined with specific technical data for analysis (such as interface response speed, success rate, etc.).

Because in fact, the main way of thinking in analysis is to be able to split the data into the smallest dimensions for analysis. Finally, by combining the time dimension to conduct further month-on-month and year-on-year analysis of the problem, we can truly meet the MECE principle.

Of course, some people may ask, can other links of AARRR be broken down like the above methodology? Then the answer must be yes.

We can also divide other aspects into three categories: product data, operation data and marketing data.

For example, the advertising and marketing data indicators we mentioned at the beginning can be simply broken down into channel indicators and then further classified.

Finally, we can divide it into qualitative indicators and quantitative indicators.

Note:

1. Qualitative indicators: refers to evaluation indicators that cannot be directly quantified and need to be quantified through other means.

2. Quantitative indicators: Assessment indicators that can be accurately defined, precisely measured and can set performance goals.

Qualitative indicators are more used to analyze the psychological attributes of individual users.

Quantitative indicators are mainly used to analyze the overall health of products or channels.

When we conduct a qualitative analysis of product users and then break them down, we can divide them into the following five categories:

• Basic attributes, such as gender, age, marital status, zodiac sign, education level, income/consumption level
• Social attributes, such as industry/occupation, position/rank, vehicle usage, mobile phone (price/brand), mobile operator, etc.
• Behavioral habits, such as the city where you live, your daily schedule, transportation methods, economic/financial management characteristics, dining habits, online shopping characteristics, etc.
• Interests, such as browsing preferences, music preferences, sports preferences, game preferences, travel preferences, etc.
• Psychological attributes, such as lifestyle, personality, values, etc.
• Psychographic attributes can also be analyzed in conjunction with the psychographic statistical segmentation tool: the VALS framework, which I will not go into detail here.

For example, users can be divided into different types based on their usage habits, which makes it easier for operations and data analysts to analyze the usage of different user groups.

For example, according to the terminal type, it can be divided into Android and IOS, and according to user activity, it can be divided into high-active, medium-active and low-active users, etc.

When quantitative indicators are split into business data, business data is essentially divided into three aspects:

• User perspective, such as DAU/MAU, new users, retention rate, channel sources, etc.
• Behavioral perspective, such as PV, UV, visit depth, conversion rate, stay time, bounce rate, etc.
• From a business perspective, such as GMV||visit duration, ARPPU||average visit duration, number of paying users||number of visitors, payment rate, payment frequency||retention rate, etc.

If we want to further split the business data, we can also analyze it according to business modules.

The common product modules on the market can be divided into four categories. The modules and indicators are as follows:

• Tools, such as usage, goal achievement rate, and frequency
• Transactional, such as detail page conversion rate, average order value, and repurchase rate
• Content categories, such as number of views, viewing time, and number of interactions
• Community categories, such as posting volume, number of interactions, and relationship density

Finally, someone may ask, is it okay if I don’t establish an indicator system for the company according to this logic?

In fact, the ideas for decomposing indicators are all based on the same principles. If you can understand this decomposition link, you can use three dimensions (users, behaviors, and paths) to analyze each major category, and you will still see a different world.

The existence of the "vein model" is like going to the hospital for a check-up. If it is just a regular physical examination, you will only be observed according to the North Star indicators, but there are some details that you cannot discover unless you take an X-ray, have a blood test, and other detailed examinations.

Therefore, the "context model" is more of an indicator system that the company should use when conducting a detailed inspection every two weeks or so to prevent the company from deviating significantly from the normal operating track.

3. Finally

Finally, let’s summarize. In order to build a product data indicator system, we first need to clarify the product’s North Star indicator. There are two ways to help determine your North Star Metric:

• Determine the North Star Indicator through the "Growth Closed Loop Diagram" of User Core Value and Business Goals
• Use the North Star Indicator Checklist to screen your North Star Indicator.

After clarifying the North Star indicator, we need to disassemble the North Star indicator and build a growth model. There are generally three models, and we focus on two of them:

• Full-link funnel model
• Factorization Model

Finally, we introduced how to use the context model to build a comprehensive indicator system for the company. The context model is built based on AARRR.

Starting from the AARRR link, we can further divide it into three categories of indicators: product, operation and marketing, and then further divide it into qualitative and quantitative indicators.

Author: Product Talk

Source: Product Talk