1. How to collect IN data

There are two ways of data collection, one is buried point and the other is no buried point

1. Buried point

1.1 What is buried point?

A very traditional and very common way is to define this event by writing code. Load a piece of code where the website needs to monitor user behavior data, such as a registration button, an order button, etc. After the monitoring code is loaded, we can know whether the user has clicked the registration button and what order the user has placed.

All these ways of describing events and attributes in detail by writing code are collectively referred to as "buried points" in China. This is a very labor-intensive project, and the process is very tedious and repetitive, but most Internet companies still employ a large number of burying teams.

1.2 7 steps of buried point collection

So, what is the process of collecting data from the buried point? Generally can be divided into the following seven steps.

(1) Determine the scene or goal

Determine a scene, or a goal. For example, we found that many users visited the registration page, but few of them completed the registration. Then our goal is to increase the registration conversion rate, to understand why the user did not complete the registration, which step blocked the user.

(2) Data collection plan

Think about what data we need to understand to help us achieve this goal. For example, for the previous goal, we need to disassemble the data of each step from entering the registration page to completing the registration, each input data, and at the same time, the characteristic data of the person who has completed or has not become these steps.

(3) Collecting data from buried points

We need to determine who is responsible for collecting data? This is generally an engineer, and some companies have dedicated data engineers who are responsible for collecting data at the buried point.

(4) Data evaluation and data analysis

What is the quality of the data collected, and how should it be analyzed?

(5) Give an optimization plan

After discovering the problem, how to come up with a solution. For example, whether it is an improvement in the design, or whether it is an engineering bug.

(6) Implementation of optimization plan

Who is responsible for implementing the solution. Determine the person responsible for the implementation of the plan.

(7) How to evaluate the effect of the solution?

In the next round of data collection and analysis, return to the first step to continue iterating.

Knowing is easy and doing is hard. In this whole process, steps 2 to 4 are the key. At present, the method adopted by traditional service providers such as Google Analytics, Mixpanel, and Youmeng is called the Capture mode. By burying certain points on the client, relevant data is collected to the cloud, and finally presented in the cloud.

2. No buried point

2.1 Collection principle without buried point

Different from the Capture mode, the Record mode uses machines to replace human experience; in the data analysis product GrowingIO, there is no need to manually bury the points one by one; you only need to load a section of SDK (Software Development Kit) when you use it for the first time. ) Code, you can collect full, real-time user behavior data.

Because of automation, we control the format of the data from the beginning of the analysis process. From a business perspective, all data is divided into 5 dimensions: Who, what attributes the person behind the behavior has; When, when the behavior was triggered; Where, the browser or even GPS in urban areas, etc.; What is the content; How, how is it done.

Based on the deconstruction of information, it is ensured that the data is clean from the source. On this basis, we can fully automate the ETL, and what data is needed can be traced back at any time.

2.2 Technical advantages of no buried point

Looking back at the 7 steps of collecting data with buried points above, no buried points can well solve the needs of the second, third, and fourth steps, reducing the original multi-party participation to basically one party. Whether it is a product manager, an analyst or an operator, you can use visualization tools to query and analyze data, and truly achieve what you see is what you get. Not only the PC, but also iOS, Android, and Hybrid are supported, allowing cross-screen user analysis.

Author: GrowingIO
Link: https://www.zhihu.com/question/24939461/answer/1238521713
Source: Zhihu
The copyright belongs to the author. For commercial reprints, please contact the author for authorization. For non-commercial reprints, please indicate the source.

The purpose of user behavior data collection is to predict what will happen in the future by understanding the behavior of users in the past. There is no need to bury points, and you can go back to the data at any time, so that one person can handle the entire process of user behavior analysis. Such a simple, rapid and large-scale data analysis product can greatly simplify the analysis process, submit efficiency, and direct business.

3. Buried point + no buried point

Regardless of whether it is a non-buried point or a buried point method, it is necessary to be able to clearly describe each user's online access process with data; this is the basic goal of data collection and the original intention of GrowingIO.

3.1 "Buried point + no buried point" data collection principle

Let's take an e-commerce app loaded with GrowingIO no-buried point SDK as an example: a customer opens the app, searches for keywords on the homepage, and then selects favorite products to add to the shopping cart on the result page; then places an order for the products in the shopping cart and completes Paid. So in this process, what data needs to be collected, and how to collect it?

The user goes from "opening the app"-"watching first-screen ads"-"search keywords"-"entering the result page"-"adding to the shopping cart" to "payment completed", there are user behavior data (process data) in the whole process ), there are transaction data (result data). In the figure above, GrowingIO's No Buried Point SDK will automatically collect all user behavior data on this App, including visits, page views, and behavior events. At the same time, GrowingIO's data docking point program can collect more transaction data, including product SKU, price, discount, payment and other information.

In this way, we can collect a complete online shopping behavior with a combination of no buried points and buried points, and use data to completely describe and analyze the user's shopping history. In fact, no matter what online business scenario, we all hope to collect complete user behavior data and business data. And it is necessary to connect user behavior data and business data.

3.2 "Buried point + no buried point" data collection advantage

So why do we need a combination of no buried point and buried point to collect data?

First, because the method with no buried points is more efficient. Through practice, we found that the data indicators generated by no buried points are 100 times or more than those generated by buried points.

Second, the cost of data collection without buried points is low, and the release of App/website does not affect automatic data collection.

Third, the advantage of buried point collection is that it can describe the attributes of each event in more detail, especially for the result data.

User behavior data collected with non-buried points is the "antecedent" data for the user to produce the final result, and business data collected with buried points is the result data is the "consequence". The solution that combines no buried point and buried point improves work efficiency, and at the same time records "cause" and "consequence" data to help market, product and operation analyze customer acquisition, conversion and retention, and achieve rapid user growth.