Don't have an analytical idea in the face of a problem? Detailed Analytical Thinking for Data Analysts (Part II)

In the previous article, we introduced the 5W2H analysis method, the industry analysis method, the comparative analysis method, the correlation analysis method, and the hypothesis testing analysis method. In this paper, we will focus on multi-dimensional dismantling analysis methods, group analysis methods, RFM analysis methods, AARRR model analysis methods, and funnel analysis methods.

1. Multi-dimensional dismantling and analysis methods

Multi-Dimensional Analysis (MDA), also known as Drill-Down Analysis, is an efficient technical method in the field of data analysis. This approach enables analysts to drill down from a high-level overview of the data to the specific data points at the bottom, revealing the underlying trends and patterns behind the data.

The core strength of Multidimensional Teardown Analysis (MDA) is its flexibility, which gives analysts the ability to quickly switch between macro level data aggregation and micro level detailed data points to uncover valuable information, depending on the analysis needs. This approach not only enhances the understanding of the deeper meaning of data, but also facilitates data-based decision-making, which in turn provides solid data support for enterprises to optimize their business processes.

1. Applicable scenarios for multi-dimensional dismantling analysis

(1) Analysis of the composition of a single indicator

Applicable scenarios: When you need to deeply understand the composition or proportion of a single metric, such as analyzing the playback distribution of different columns on a video platform or the proportion of new and old users.

(2) Business process analysis

• Scenario 1: It is suitable for global conversion process analysis, such as the entire process from users browsing products through different channels, to adding products to the shopping cart, and finally completing the purchase.
• Scenario 2: For products sold across regions, analyze the differences in the effectiveness of marketing activities in different regions. For example, you can perform a detailed analysis of the effectiveness of promotions in different provinces.

(3) Restoration of user behavior scenes

• Scenario 1: It is suitable for products that require a deep understanding of the specific situations in which user behaviors occur, such as live streaming services. As a product manager or operations person, you may need to analyze tipping behavior from multiple dimensions, including the level of the tipping user, their gender, and the channel they watch.
• Scenario 2: When analyzing user behavior, it is also necessary to consider the user's usage environment, for example, to understand whether the user is operating in a Wi-Fi or 4G network environment, which can help to more accurately grasp the user's behavior pattern.

2. How to choose the dimension of disassembly?

(1) Indicator composition dimension

First of all, we can start with the specific composition of the indicator to disassemble. For example, if the single metric we focus on is "Users", we can further segment our user base into two subcategories, "New Users" and "Returning Users", to analyze user behavior and characteristics in more granular terms.

(2) Business process dimension

Second, it can be disassembled according to different stages of the business process. This approach is useful for analyzing key business metrics such as pay-per-user rates, especially across different marketing channels or user acquisition strategies. In this way, efficiency bottlenecks or advantageous channels in the user conversion process can be identified.

3. Application scenarios of multi-dimensional dismantling analysis methods

The Multi-Dimensional Dismantling Analysis (MDA) approach plays a key role in the field of data analysis, which is mainly reflected in the following key areas:

• Deeper data insights: MDA allows analysts to drill down into data through multiple perspectives to reveal complex connections and underlying patterns between variables. This approach helps identify key trends, outliers, and key contributing factors.
• Decision-making: Provides decision-makers with a comprehensive view of data analytics, enabling them to make more informed decisions based on more complete data insights. For example, by analyzing sales data in multiple dimensions (by time, region, product category, etc.), management can more accurately grasp sales trends and thus allocate resources more effectively.
• Rapid Problem Localization: When faced with business challenges, MDA can help enterprises quickly trace the problem to its source. For example, if there is an unexpected drop in sales of a product, a detailed, multi-dimensional analysis can quickly determine whether the problem originates from sales in a specific region or at a specific time period.
• Optimize business processes: Through comprehensive and multi-faceted data analysis, enterprises can identify and improve inefficiencies, and then optimize strategies and operational processes to improve the overall efficiency of the business.
• Customized marketing strategies: MDA enables companies to more accurately grasp the specific needs and behavior patterns of different customer groups, so as to design and implement more accurate and customized marketing strategies.

Through this multi-dimensional analysis method, enterprises can not only gain a deeper understanding of data, but also formulate a more scientific and reasonable business strategy to cope with the continuous changes and challenges of the market.

The data analysis template mentioned in the example is shared with you——

https://s.fanruan.com/x3k5k

Zero-based quick start, but also according to the needs of personalized modifications

2. Cluster analysis methods

Cohort analysis, sometimes referred to as cohort analysis, is an analytical technique that divides a dataset into different groups based on specific characteristics. With this approach, we can group data objects into groups with similar characteristics, and then compare and analyze the data from these different groups. Essentially, the purpose of cohort analysis is to reveal the underlying structure and patterns of data through group comparison.

1. Application of group analysis methods

Cohort analysis is often used to track and evaluate changes in user retention (or churn) over time. With this approach, it is possible to dig deeper and identify the underlying factors that lead users to choose to continue using the product or ultimately leave.

When dealing with retention issues, user grouping strategies are often based on the month in which users first came into contact with the product. This can be the month the user signs up or the point in time they make their first purchase. Once the grouping is complete, the next step is to monitor and analyze the retention rate of each user group over time. For example, you need to evaluate the retention of users for one month, two months, or more after signing up or purchasing (as shown in Figure 2-115).

For groups of users with high retention rates, it's important to analyze and understand why these users are keeping high retention rates. This may involve factors such as their high level of satisfaction with the product, the irreplaceability of the product, or the deep engagement of users with the product.

2. How to use the cohort analysis method

Apply cohort analysis to identify groups of users who exhibit low or high retention rates. Next, statistical tools such as hypothesis testing and correlation analysis are used to delve deeper into the reasons for the varying retention rates of these cohorts. Once the key factors have been identified, the product can be optimized in a targeted manner.

When faced with a cohort analysis table that contains a large amount of data, it can be difficult to analyze it directly. To simplify this process, data visualization techniques can be used to transform the data into a line chart and other forms. Such charts can visually reveal trends in data over time, allowing us to quickly capture key time nodes and pattern changes, allowing us to more effectively identify and understand factors that affect retention.

3. Key points of using the cohort analysis method

When using cohort analysis, the key is how to group the data effectively. While it is common to group groups in chronological order, the criteria for grouping should be flexible according to specific business scenarios in order to better meet actual business needs. This means that, in addition to the time dimension, different user groups can be divided into different user groups by considering various factors such as user behavior, geographical location, user preferences, product type, etc.

3. RFM analysis method

The RFM model is a widely used analysis tool for users to refine operations, and its name consists of the first letters of three English words:

• R (Recency) refers to the time interval between the customer's most recent purchase, which reflects the customer's activity and loyalty to the brand.
• F (Frequency) indicates how often a customer buys within a certain period of time, and it measures how frequently a customer trades.
• M (Monetary) refers to the total amount of money spent by a customer over a specific period and is used to assess the economic value of a customer.

Together, these three dimensions form the RFM analysis framework, which not only describes the value of customers, but is also one of the popular methods of user stratification, especially suitable for in-depth research on paying user groups.

1. Characteristics of indicators

When applying the RFM analysis model, you need to customize the definition of three core indicators according to the characteristics of specific services, and the characteristics of the three indicators are as follows:

• Recency (R): This metric measures the length of time that has elapsed since a customer's last purchase. In most cases, the smaller the R-value, i.e., the closer a customer has recently made a purchase, indicating greater loyalty and value to the brand.

• Frequency (F): This metric reflects the number of times a customer makes a purchase within a certain period of time. The higher the frequency, i.e., the higher the F-number, it usually means that customers interact with the brand more frequently and its value is correspondingly higher.

• Amount spent (M): This metric represents the total amount of money spent by customers in a specific period. The higher the M value, that is, the greater the amount of customer spending, indicating that its contribution and potential value to the enterprise are also greater.

By sorting these three indicators by user value from low to high, and using them as axes, user groups can be segmented in a multidimensional space. This approach allows the entire market space to be divided into eight distinct segments, each representing a user group with specific RFM characteristics.

2. RFM user classification process

Here are the detailed steps to apply the RFM model to classify users:

• Raw data calculation: First, based on the user's historical transaction data, the specific values of each user's Recency (R), Frequency (F), and Monetary (M) indicators are calculated.
• Indicator scoring: Then, each R, F, and M indicator is scored, using a scoring scale of 1 to 5, where 1 is the lowest value and 5 is the highest.
• Qualitative scoring: Then, the average score of each indicator is calculated. Based on this average score, the score for each metric is compared to the average score: if the score is below the average, the indicator is marked as "low"; If the score is higher than the average, it is marked as "high".
• User classification: Finally, the "high" and "low" markers of the R, F, and M indicators of each user are matched with the preset user classification rules to obtain the category to which the user belongs.

Through this process, companies are able to divide users into different value groups, and then implement differentiated marketing and service strategies to improve user experience and business revenue.

3. How to carry out refined operations according to the classification results

(1) Valuable users

When a user's RFM performance is good in all three indicators, it indicates that they are a high-value customer. For this type of user, providing a VIP-level service is key to show recognition of their loyalty and spending level.

(2) Important development users

Although these users do not spend frequently (low F-value), they have a high time interval between their most recent purchases (R-value) and amount of money (M-value). Enterprises should strive to increase the purchase frequency of these users through personalized marketing strategies to maximize their potential value.

(3) It is important to keep users

Although users' recent spending behavior is far away from the present (high R-value), they have shown higher spending frequency and amount (higher F and M values) in the past. This shows that they are loyal customers, but they may not have looked back in a while recently. Businesses should proactively communicate with these customers and take steps to promote their repeat purchases to maintain their loyalty.

(4) Important customer retention

These users have a longer time interval between their most recent purchases (high R-value) and less frequent (low F-score), although they still spend a higher amount (high M-value). This means that these customers may be at risk of churn. Businesses need to be proactive, communicate with customers, understand the reasons behind them, and take appropriate measures to regain these potential lost customers.

4. Key takeaways

(1) Business customization of R, F, and M indicators

In different business contexts, the definitions of R (last consumption interval), F (consumption frequency), and M (consumption amount) will be different. Therefore, these metrics should be flexibly defined and interpreted according to specific business scenarios when applying.

(2) Scoring rules for R, F, and M indicators

Scoring rules should be determined based on the value of the metrics and take into account the specific needs of the business. Scores are usually graded from 1 to 5, but the specific score ranges and criteria should be adjusted according to the characteristics of the business. In addition, RFM scoring rules can be developed by working with business teams to brainstorm ideas or using cluster analysis techniques to group R, F, and M values and assign a score to each group.

(3) Integration of R, F, and M indicators with other analytical methods

These three metrics can be used not only on their own, but also in combination with other analytics techniques such as predictive modeling, trend analysis, and more to provide more comprehensive customer insights. This flexible integration enhances the depth and breadth of analytics to better support decision-making and strategic planning.

4. AARRR model analysis method

The AARRR model is a core framework in product operations and covers five key phases, each of which corresponds to a different level of user engagement within the product lifecycle:

Acquisition: Explore the process by which users discover and engage with a product. The core question is "How do users hear about our products?" ”

Activation: Focus on the user's initial experience and evaluate the quality of their interaction when they first use the product. The key metric is "Is the onboarding experience smooth for users?" ”

Retention: Analyze whether users are willing to return after the first use, focusing on the user's continued engagement. The main consideration is "Do users continue to return to the product?" ”

Revenue: Focus on how to monetize your user base, including pricing strategies, paying user conversions, and more. The core question is "How can we increase revenue?" ”

Referral: A measure of a user's willingness to recommend a product, i.e., whether they will introduce the product to friends and family. The key metric is "Are users willing to become recommenders of the product?" ”

Through the continuous optimization of these five steps, the AARRR model helps product operators build a complete set of user growth and maintenance strategies to drive product success.

1. The role of AARRR model in user growth

The AARRR model covers the entire life cycle of user interaction with the product, so it can effectively assist in analyzing the behavior patterns of users and provide decision-making support for product operations, thereby promoting the growth of the user base. For example, when the low retention rate is found to be the core of the problem through other analysis methods, we can learn from the strategy of improving user retention in the AARRR model to take targeted measures to improve the continuous engagement of users.

With this model, teams can more systematically identify and address critical issues in user operations, enabling optimization at every stage. This involves not only the initial stages of acquiring new users (Acquisition) and activation, but also the subsequent steps of increasing user retention, increasing revenue and promoting user referral. The AARRR model provides a comprehensive set of analytical and action frameworks for product operations to achieve sustainable user growth and product success.

2. How to use the AARRR model analysis method

(1) User acquisition strategy: How do we attract users?

Here are a few key metrics that are critical to user acquisition:

• Channel Exposure: A measure of how many people see a product promotion message, reflecting the reach of the marketing campaign.
• Channel conversion rate: Tracking the percentage of users who actually become users after seeing your ad message is key to measuring the effectiveness of your ads.
• Number of new users per day: Records the number of new users added each day to evaluate the dynamics of user growth.
• Daily app downloads: Counts how many users download a product each day, which is a direct indicator of user interest and acquisition efficiency.
• Customer acquisition cost: Calculating the average cost per customer acquisition is critical to assessing marketing effectiveness and financial viability.

In the Internet industry, the challenges faced by many startups are not due to the inadequacy of the business model or the product itself, but due to the high cost of customer acquisition and the lack of effective cost reduction strategies. High customer acquisition costs can quickly erode profit margins, limit a company's growth potential, or even lead to a company's failure. Therefore, optimizing user acquisition strategies and reducing customer acquisition costs are crucial for startups to survive and grow.

(2) User activation: Optimize the user's initial experience

Although many products have managed to attract a large number of user sign-ups, the actual usage rate (open rate) of users is not ideal. At the user activation stage, our goal is to spark the interest of users and encourage them to actually start using the product. In order to achieve this, the key is to identify and reinforce the "aha moment" of the product.

What is an "aha moment"? It refers to the moment when the user suddenly experiences the unique value of the product in the process of using the product, and the resulting strong emotional resonance. This moment often allows users to deeply feel the charm of the product, so as to establish positive habits.

(3) User retention: How to promote the continuous return visit of users?

Once a user has completed their initial activation, the next step is to turn them into loyal repeat customers and ensure that they continue to use the product. The core of user retention is to cultivate users' habits and enhance their stickiness to the product.

In order to effectively improve user retention, we need to pay close attention to the key metric of retention rate. Retention rate reflects the percentage of users who continue to use a product after a certain period of time (typically a day, week, or month) and is an important measure of product appeal and user loyalty.

(4) Revenue growth: how to improve the profitability of products?

On top of user retention, the next goal is to increase revenue. There are two main types of revenue sources for products: service revenue and advertising revenue. Service revenue usually comes from users' purchases of in-product paid services, such as the membership service provided by NetEase Cloud Music, where users need to open a membership to enjoy the right to listen to specific songs. Advertising revenue comes from the advertisements displayed in the product, such as the revenue generated by the promotional articles published by the official account.

In order to effectively increase revenue, you need to focus on the following key metrics:

• Overall business metrics: such as total turnover and number of deals, which reflect the overall size and growth trend of the business.
• Per capita metric: For example, the average order value, which measures the average revenue per user brings to a product.
• Paying behavior metrics, such as the pay rate and repurchase rate, which reveal user willingness to pay and loyalty.

In addition, there is one concept that needs to be paid special attention to - "grips". The so-called pinch point refers to the potential revenue loss point that may occur in the process from the discovery of the product to the completion of the payment. For example, in the e-commerce shopping process, after selecting an item, a user may abandon the purchase at any stage of adding to the cart, selecting a payment method, or filling in payment information. Identifying and evaluating these pinch points requires an in-depth analysis of why users churn at these key points so that you can take action to reduce losses, optimize the user experience, and increase conversions.

(5) User recommendation: how to stimulate the user's communication potential?

After completing the first four steps, we move on to the final stage of the AARRR model – user recommendations, which is also known as viral marketing or product self-propagation.

In this session, our goal is to motivate existing users to recommend products to their social circles, resulting in word-of-mouth and organic growth in the user base. The success of the recommendation process depends on the satisfaction of users and the recommendation value of the product itself, that is, whether the product has enough features or advantages that attract users to actively share.

In order to promote user recommendation behavior, the product needs to make a strong impression in the user's mind, provide value that exceeds expectations, or create an interesting sharing mechanism. This could involve a well-designed product experience, a user incentive program, sharing rewards, or harnessing the power of social media to amplify the reach of your product.

5. Funnel analysis method

Funnel analysis is an analytics technique that breaks down a user's path to behavior by detailing it into key steps. This approach aims to identify and resolve barriers to user conversion, thereby improving conversion efficiency. By visualizing the conversion process as a funnel model, we can gain more intuitive insight into the user's behavior patterns and intrinsic needs, which provides a basis for product and service improvement.

1. What is the function of the funnel analysis method?

The main role of funnel analysis is to "locate the problem node", that is, to identify the specific part of the business process where the problem occurs. This approach is particularly useful for the analysis of user conversion and churn, so it's important to focus on two key metrics: conversion rate and churn rate.

(1) User conversion value

Users who go through various business links and ultimately convert tend to have higher value. Not only are they more loyal to the brand, but they are also more receptive to business processes, and this increase helps improve the overall quality and monetization potential of users.

(2) Retention and profitability

As the number of converted users grows, the base of retained users will also expand accordingly, which will drive the growth of product profitability.

(3) Churn analysis

The churn of users varies depending on the business. Through analysis, it is possible to identify where users are most likely to churn and why. Possible causes include the complexity of business processes, inadequate product features, or other unmet user needs.

(4) Optimization goals

The ultimate goal of funnel analysis is to identify and address these issues that lead to user churn, continuously reduce user churn by optimizing processes and improving user experience, thereby improving the efficiency and profitability of the entire business process.

Funnel analytics provides a powerful tool for businesses to fine-grained manage the user journey and drive continuous business improvement and growth by systematically evaluating user performance at every step of the business process.

2. How to conduct funnel analysis

There are a few key steps that need to be completed:

• Set conversion goals and key metrics: Clearly define the conversion goals for your analysis, such as purchase actions or user sign-ups, and identify key conversion rate metrics.
• Data collection: Use data collection tools to collect user behavior tracks, covering data on key conversion nodes such as application launch, user entrance, registration process, product viewing, and purchase behavior.
• Build a funnel model: Based on the collected data, build a funnel model to subdivide the user's entire behavior path into multiple key steps, and calculate the conversion rate and churn rate of each step separately.
• In-depth analysis: Conduct an in-depth analysis of steps that exhibit low or high conversion or churn rates to explore potential causes, such as poor page layouts, cumbersome conversion processes, or poor user experience.
• Develop improvement measures: Based on the results of the analysis, formulate targeted optimization measures, which may include page design optimization, simplification of user operation paths, introduction of user feedback mechanisms, etc.

Through this series of orderly steps, funnel analysis can help enterprises systematically identify and solve problems in the conversion process, and ultimately achieve business process optimization and conversion efficiency improvement.

When applying funnel analysis to explore user conversions, it's important to recognize that business processes vary from industry to industry, and as a result, the corresponding funnel analysis charts will vary. If the funnel analysis method is directly applied to a certain industry without combining it with the specific business characteristics of the industry, the resulting analysis results may not be effective in guiding the actual business.

6. Summary

After introducing in detail the methods of multi-dimensional teardown analysis, group analysis, RFM analysis, AARRR model and funnel analysis, we can draw a comprehensive conclusion: data analysis is not only to reveal the appearance, but also an important means to gain in-depth insight into user behavior, optimize product experience, improve business efficiency and enhance market competitiveness. Each analysis approach provides us with a unique perspective that helps us understand the story behind the data from different dimensions.

With the continuous advancement of big data and artificial intelligence technology, the methods and tools of data analysis are also evolving. Businesses need to constantly learn and adapt to new analytics techniques to maintain their competitive edge in a rapidly changing market. At the same time, the results of data analysis need to be closely integrated with business decisions, and data insights need to be transformed into actual business results through continuous testing, learning, and optimization.

With its superior performance and flexibility, FineBI data analysis tools provide enterprises with a comprehensive data insight platform. It supports the rapid processing and analysis of massive amounts of data, ensuring immediate business decision support. Users can easily create complex data models through FineBI's drag-and-drop interface, and its rich visualization components make the presentation of data vivid and intuitive. In addition, FineBI's real-time data processing capabilities enable enterprises to keep up with the pulse of the market and respond quickly to changes. Whether it's a data analyst, business manager, or decision-maker, FineBI provides the deep analytics and instant insights they need to stay ahead of the data-driven business competition.