One of the KPIs of the Internet product manager is whether the release of each version brings new data to increase the sequential rate, new users to attract new users, and the increase of payment rate.
The general direction, though, is to prioritize business needs and leadership tasks.
Every Internet product has its own data background. With the increase of product lines and business requirements, the data background will eventually become a big data platform.
But what are the key module designs for building Internet product APP, small program and Web big data platform from 0 to 1?
Today disassemble the case module of data platform and divine policy.
Big data platform building framework
To facilitate the understanding of big data platforms, first of all, each product has a common module. Module design is as follows
▲ Data platform case module
Like other backend products, each SaaS product has a system setup. If you read my previous b-side product design, you should know the system Settings section.
Meanwhile, the above analysis tools were developed in iterations after 1.0. Typically, companies combine user analytics, analytics tools, user profiles, and user tags.
The application list module is also not suitable for enterprise research and development. Because of the product line, each company has its own product line and no more than one application. Usually an Internet company only has one app or small program product.
Overview data reports
Overview data reports are for the boss, the head of operations, product owners.
▲ Overview data report
1. New users today
The only registered user under the account system, the increase of the month, month, week
2. Accumulative active regular users
The active definition varies for each enterprise, with some companies defining two logins per day. Some definitions are once a day.
3. Cumulative daily active number today
Users who meet the definition of active participate today.
4. Number of visits today
The cumulative number of UV visits to the product today
5. Increase the number of users by time sharing today
Indicates the number of users added in the current hour/minute
The mean 6.
The average growth value can be used as the north star indicator of product iteration
The time trend length is recorded in today’s day, week, and month. At the same time, the current data and average data are distinguished by color and summarized as the overall trend table.
The overall trend
▲ Overall trend report
1. Startup times
Open times of product APP, Web and small program.
2. Trend of new users
Percentage of new users to total users
3.7 days of retention
Users who visited the product twice in 7 days
4. Next-week retention
After 7 days, the percentage of users who have logged in or paid
5. Active user trends
The user activity is defined in 30 days and 365 days in a year.
6. Average single time trend
The average time between each user access to the product and the kill process
7. Average daily usage
The average amount of time a user spends using the product per day. Let’s say 10 users, 300 minutes. 30 minutes for a user.
User life cycle
In building this module, we need to clearly define the four objects of returning users, silent users, continuous active users and lost users in advance
▲ User life cycle report
By looking at the bar chart above, you can quickly get a sense of the percentage of a user’s life cycle.
A life cycle phase is represented when active users + incoming users = lost users + silent users.
New User Analysis
On the main kanban, we can see the overall functional data, but no new details. Therefore, it is possible to build additional detailed kanban as follows.
▲ New User trend
1. New users on a daily/monthly/weekly basis
New users in different time dimensions
2. 7-day retention of new users
Users who have logged in to the product in the last 7 days
3. New users are retained for 8 weeks
Users who visited the product again within 8 weeks
4. Six-month retention of new users
Users who visited the product again within 6 months
Active User rankings
Active user from page, page view to active user.
Since the definition of activity metrics may vary from enterprise to enterprise, here we use PMTalk’s definition of activity: users who have logged into the community.
▲ Active User rankings
1. Monthly/weekly active number
Number of active users in different time dimensions
2. Page view ranking
PV ranking per page
3. Cumulative page exit rate
The exit rate of users leaving a page
4. Page button click ranking
Click the leaderboard button on the page
Terminal equipment
The usage of terminal equipment can be disassembled by using the chart below
▲ Terminal device report
Number of active operating systems
Active number of Apple systems
Active Android
Active use of equipment brand
Ratio of device models
Android resolution active users
Apple resolution active users
Network Type Indicates the number of active users
Active user number of the application version
Active number of network operators
By device size, you can reduce development costs. For example, the android version of the package, can focus on the device model adaptation of a high proportion of users, reduce resources and time investment in time matching.
At the same time, by understanding the user’s network operation, timely update the application in the file size, cache technical solutions.
Channel data statistics
Retention, number of registrations, number of activations across different channels.
Users who download through channels can establish the above logic by detecting user login and account registration number.
Data statistics platform, starting from requirements
To sum up, the overall data platform construction. You can start from the above module, the source logic of data statistics from the front-end buried point, back-end server logic, index definition three points to deal with.
As a product manager of data platform, I gave priority to business demand and operation demand at the beginning. So almost everyone has the feeling of running errands.
After all, data is a little bit here today and a little bit tomorrow. There is no end and perfect day.
So if you’re just starting out with a data platform, it’s recommended to start with a summary report. In turn, the channel, equipment, user portrait, user life cycle to do subdivision statistics
Today’s share is here.
To learn more about product managers, you can follow the “Kevin Changes the World” public account and add vX: KevingbsjDDd0123 for more information