When you are confused about operations, you can use it to find a breakthrough!

A friend who does e-commerce asked on WeChat: "User participation is good during the event, but the user retention rate is low after the event. What should I do?

Me: "Have the coupons been issued? Are there any reminders for expiration? Are there any recalls for the instant discount benefits?..."

Friend: “I did it all.”

Me: “That may be a product problem. The essence of e-commerce is still about product quality and logistics.”

Friend: "You are right. Because of the low traffic, high-quality merchants are unwilling to move in."

Me: “What we can do now is to run the data to see which categories of merchants have good quality (relatively high click-through rate , conversion rate , and repurchase rate). We can start to make a breakthrough in user retention from these categories that can still be saved.”

….

I don’t know whether my suggestions can ultimately help him solve the problem, but at least they point out a way to start operational optimization immediately.

In my hometown of Shaxian, there are many people running snack shops. In order to make more money, most of the snack shop owners will analyze how much they have sold today, the cost of purchasing ingredients, which one has a higher consumption, steamed dumplings or wontons, the customer flow, the average customer spending, whether the customers are old customers or new customers, whether they came through word-of-mouth effect or just strolled in...

In the snack food industry where information technology is almost non-existent, there is no choice but to rely on the bosses' experience to make analyses and judgments. It is difficult for them to come up with specific data to analyze business conditions, and usually they can only decide on the location of the store based on the size of the flow of people.

However, in the Internet industry where data is the biggest selling point, all work and user behavior can be quantified. Big companies like Alibaba call themselves data companies and do big data analysis of user behavior, while small ones are just statistics of banner clicks on the website. Of course, today we will only talk about the small side: how to use data to better operate.

What is your sharing/collection rate for pictures and texts?

What is your product’s ROI?

Which of your channels has higher quality?

How many readers does your content have?

How many users do you have for content production?

…….

As a measurement method, data can truly reflect the status of product operations , help us further understand products, users, channels and optimize operational strategies. Use the results of data analysis to drive the operation mode. Different operation directions (content operation, event operation, channel operation, user operation ...) require operation practitioners to have different data sensitivity and logical capabilities.

1 / Data analysis of event operations

Planning activities is a common means of operation. In order to avoid problems such as "not knowing how to estimate the effect of the activity", it is usually necessary to determine a core data to be improved before the activity starts. Such core data includes new user registrations, user activity, user payment conversions, product transaction volume, brand awareness (Baidu Index, Sina Index, Media Index)...

In addition, the activities planned by operations can usually be divided into two categories. The first category is interactive activities that are not commodity transactions. The main data it focuses on is the amount of user interaction generated by the activities and the effect of improving the core data of the product. Digging deeper into these interactions is to analyze the interaction ratio between old users and new users. Such further analysis can help operations understand the differences in the impact of various types of interactive activities (interaction form + prizes) on new and old users.

Figure 1: Click-through conversion comparison between new and old users on a campaign page

The second type of activity is the promotion of commodity transactions, which mainly focuses on the clicks on key nodes of the promotion page and whether the various conversion rates are much higher than those in the daily non-active state: the number of clicks on the activity page products/the proportion of traffic entering the products, the conversion rate of browsing/adding to the shopping cart/ordering, the shopping cart/completion payment rate, the number of coupons verified/coupons issued", and of course the total transaction amount brought by the activity to the platform.

The granularity of data analysis for promotional activities should be made smaller. It is recommended to conduct segmentation analysis based on new and old users/payment ratio of users in channel A and channel B, analysis of the average paying customer price, and subsequent user retention, etc. Taking the sales of membership-based value-added services as an example, in addition to analyzing the total sales brought by the activity, you can also analyze the payment situation of users at each level: the purchase volume of old paying users and the distribution of the validity period of the memberships they purchased/the number of new paying users/the distribution of the number of months of membership purchased/the recall rate of lost users/the payment conversion ratio of experience members.

Figure 2: Hierarchical member data management

2 / Data analysis of content operations

What editors are more concerned about is the content display, including the content updates, exposure , clicks, dwell time, forwarding, collection, likes, comments and the conversion rates of the content browsing path "exposure-click-reading-sharing".

Image 3: Analysis of the automatic image-text conversion rate of the official account

For content operations that are specifically aimed at improving a certain product indicator, more attention needs to be paid to the role of content in promoting that indicator. For example, when recalling users through content, what needs to be paid attention to is the ratio of the number of people delivered the content to the number of silent users logging into the product . Now more and more e-commerce companies are trying to use content to guide sales. In addition to paying attention to the basic reading volume of pictures and texts, such content operations also need to look at the clicks on product links in the content, product browsing and order status.

3 / Data analysis of channel operations

Advertising channels are the main way for products to acquire users. Operations can truly understand the channel’s ability to acquire users through downloads and activations . Channel operations need to analyze the conversion rate of each link in the user acquisition chain of "channel click-download-activation-registration-deeper behavior", and then optimize the conversion effect of each link to reduce the acquisition cost of each user.

When a new version is released and there is a significant data fluctuation, simply looking at the data of the channel user acquisition path cannot help us make operational decisions. It is recommended to compare data in different dimensions, such as comparing with historical data, or making multi-dimensional comparisons of similar apps at the same time point. This will help operations find the root cause of changes in channel conversion rate data.

Let’s take an example to illustrate how to find problems in operations by analyzing channel retention rate : After a social app released a new version, its daily active user ratio dropped sharply compared to before the release. How should this problem be analyzed?

DAU is divided into new users and old users. Operations can let the technology run a set of data (as shown in Figure 2) to study the retention of old users who have recently used the new version to see if the loss of old users is caused by product problems; then observe the sedimentation (retention) of new users in each channel's retention rate to see if it is a product problem or a channel promotion problem.

Figure 4: Retention analysis of a social product

Through comparative analysis of this set of data, we can find that old users have a good acceptance of the new version and there is no major change. The main reason for the very low overall retention rate after the version is that the new version has a problem of low user quality in the promotion of the main channel A.

The above is a simple data analysis idea listed based on common operation jobs. You can reply to the keyword "data" in the official account "Product Rookies" to see the 6 commonly used data tools recommended by Xiaoxian.

4/ Data analysis of project operations

For the operation of a project or product, if you need to follow up on data analysis from the beginning of the project, you should consider the following data-related matters.

1) Data items and tracking points

Operations need to list as many data items as possible based on the understanding of business processes. To do this kind of work, you can make a data funnel model based on the user behavior path. The finer the path split, the more effective the model will be.

Figure 5: User behavior funnel of financial products

As basic data, data items need to be obtained through technology by tracking points on the corresponding paths. Operations need to analyze the key path conversion ratio based on these basic data and analyze the same data before and after time to clarify the project's operating status. The data items that need to be further improved are to extend forward and backward based on the core data surrounding the project. The backward direction is the statistics of the user's subsequent behavior, and the forward direction is the quality analysis of the user's source.

In addition, data analysis can be divided into report data and non-report data. Report data refers to the data that can be directly viewed in the data background every day. Usually, user key path data needs to be made into report data. For example, the comparison of path behavior differences between new and old users is a long-term and non-core data, which is not recommended to be made into a report in the early stages of the project. It is only necessary to let the technology run a period of time for data comparison when analyzing the staged product operation status.

2) Data analysis cycle

As the name suggests, the data analysis cycle refers to the time interval between operations to extract the data generated by the data report for further analysis. In the early stages of a project, the analysis cycle can be one day, and once every seven days or even longer as the project matures. In addition to analyzing the conversion rate data for each key link, operations can plot the basic data on a timeline in an Excel spreadsheet. A broken line can be made with time as the horizontal axis to see the data trend of each path, and a bar chart can be made with time as the vertical axis to see the conversion rate of each path.

Figure 6: Trend analysis of the number of functional users of a project

3. Data Analysis Conclusion

Data analysis is a means, and the most valuable things are the conclusions after the analysis. Low reading volume, few forwarding volumes, and decreased transaction amounts... Use data analysis to find out the causes of the problems; increased transaction amounts, high reading volume, and high forwarding volumes. Through data analysis, we can extract those operational means that are effective in improving data, and these means can be fixed into product forms to further expand the data effect.

4) Project operation adjustment

In fact, data itself cannot directly help improve products. Data is ultimately a measurement. If you want data to really play a role, it must be based on the operator's deep understanding of the business. As the person who knows the data in the product best, operations personnel use data as the basis for product improvement. They need to use data to prove the value of continuing the project and use data to drive product, technology, and channel colleagues to complete the optimization and iteration of their respective work.

5) Overall project report

Large projects can basically be divided into several different periods, which can be generally divided into "inspection period - entry period - growth period - high growth period - maturity period - decline period". Each period will have different focuses, and the overall operation status of the project needs to be reported to the higher authorities.

If the project is reliable, there are many data points, platform data analysis is done frequently, and the project summary document writing routine of "project background, project goals, project data, operational analysis, experience reuse, and follow-up plans" is used, there will be basically no problems when encountering the leader's surprise project overall report.

Here I would like to focus on the project data part of the report. As the focus of the report, project data can be divided into core data and peripheral data (including content data, user data and comparative data of various dimensions). The core data is placed at the front of the report with an overview of the project status. A clever way to do this is to benchmark a product that everyone is familiar with. For example, if you are working on an information sharing project for elementary school education products, you can use such data to open the report: "The sharing of educational product information and user consumption are both quite large. The product has 20,000+ daily active users, and users share an average of 200 high-quality product content per day. The overall per capita browsing PV of the project is 10, and the per capita sharing volume is 7 PV. It is very likely to become what is worth buying in the K12 field."

After explaining the core data, we start analyzing the peripheral data. At the content operation level, we can analyze what types of content the target users of the project like, how much of our content is PGC and how much is UGC, and how are their sharing and browsing situations? If you are really not afraid of offending people, you can compare your content data with the content data of other projects to prove that your project content is of high quality and is very valuable in promoting product activity.

Data analysis at the user operation level mainly includes the conversion status of each key path of the core business, sorting out the key nodes of user loss and functional optimization suggestions, and then briefly mentioning the user level of the project, how many core users participate in content contribution, how many share the content and participate in content dissemination, how many people who read more than * times are loyal users, and how many are lurking users.

If your project is just to improve a core data in the product, after analyzing the improvement of the product's daily activity, you should conduct data analysis on several key aspects of the project that affect this core data to see where there is room for optimization and improvement.

Finally, let’s get back to the core point of our topic: When you encounter planning confusion during the operation process, you can first analyze the data of the execution of the existing operation strategy. Don’t prematurely define the existing strategy as invalid or immediately look for a new strategy (there are only a few strategies, and blindly changing will only push yourself into a desperate situation faster). Just like the conversation between my friend and I at the beginning, there is nothing wrong with the activity strategy itself. The core problem of not improving retention is that the products in the mall are not good enough and even the logistics have been criticized.

Your APP