---
title: "Data Analyst: The Decoder of Data"
description: "Data analysts are the decoders of data's commercial value. They are scarce and crucial for digital transformation, requiring a blend of business and technical skills. The article outlines the purposes of data analysis, types of analysis, and the essential qualities of a data analyst."
author: "刘春雄"
publisher: "New Distribution"
email: "zhaobo258@gmail.com"
telephone: "+8615854817671"
published: "2021-02-23"
language: "en"
canonical: "https://xinjignxiao.com/en/articles/data-analyst-the-decoder-of-data-f4e9c38d/"
markdown: "https://xinjignxiao.com/en/articles/data-analyst-the-decoder-of-data-f4e9c38d.md"
original_source: "https://mp.weixin.qq.com/s/4XAc_nI3IsiMztlKmT2-OQ"
translation: "https://xinjignxiao.com/zh/articles/%E6%95%B0%E6%8D%AE%E5%88%86%E6%9E%90%E5%B8%88-%E6%95%B0%E5%AD%97%E8%A7%A3%E7%A0%81%E5%99%A8-f4e9c38d.md"
attribution: "New Distribution — https://xinjignxiao.com/en/articles/data-analyst-the-decoder-of-data-f4e9c38d/"
usage_policy: "https://xinjignxiao.com/ai-policy.txt"
---

# Data Analyst: The Decoder of Data

> Data analysts are the decoders of data's commercial value. They are scarce and crucial for digital transformation, requiring a blend of business and technical skills. The article outlines the purposes of data analysis, types of analysis, and the essential qualities of a data analyst.

**Scan the QR code in the image to register**
**-01-**
Data doesn't think; people do.
Algorithms are just programming based on human thought logic. Even AI's algorithmic logic is human. Artificial intelligence is [human intelligence + machine computation], where machines excel at computation.
**Data analysts are the decoders of data's commercial value.** Che Pinjue, former vice president of Alibaba and president of the Data Committee, said that data analysts are the "strategists" of business.
A boss who implemented digital operations years ago told me that data analysts are his treasures. According to him, cultivating a qualified data analyst takes several years, which is more difficult than training a marketing manager.
**-02-**
What is the purpose of data analysis? There are roughly three types:
## **First, data insight: discovering opportunities**
This is the need of middle and senior management for digitalization. Opportunities lie outside; discovering them requires going to the front line. In the past, discovering opportunities required senior management to go to the front line and use their insight.
Digitalization gives enterprises another ability to gain insight: **data insight.**
How to do data insight? This is a challenge. Any insight is hard to imitate and hard to program. What is programmed is not insight.
Insight cannot be replaced, even by big data insight. Now big data can discover correlations between previously unrelated factors, but it's not yet at the level of insight.
## **Second: judging and grasping the overall business situation**
Every level and every person needs data to grasp the overall situation, including bosses and frontline salespeople. In the past, this was done through frontline visits and statistical data. Visits take time, and statistics lag.
**Digitalization allows real-time data to judge and grasp the overall situation, reducing unnecessary visits.** For example, frontline salespeople can use real-time data to judge which terminals have problems and then make targeted visits. Originally, a salesperson could only manage 150-200 terminals; now they can manage multiples of that.
Digitalization has now achieved "visualization," which facilitates grasping the overall business state.
## **Third: improving business processes**
Traditional marketing had no 2C, only 2B, so business processes were relatively simple. 2C business processes are more complex, usually represented by AARRR. Acquisition, activation, retention, revenue, referral—each step has data to rely on, making sales policies targeted.
**-03-**
**There are three types of data-driven business analysis:**
**Type 1: Model programming, automatic distribution.** According to a defined business analysis model, a fixed program is compiled. Real-time response in 2C is like this. Every time you click on an e-commerce platform, Douyin, or Toutiao, the system immediately follows the algorithm, showing different content for each person and each time. This push must be "instantaneous" and imperceptible to the user. This process is called marketing automation.
**Type 2: Routine analysis, manual judgment.** The system provides analysis templates, and data is analyzed according to the template, but final judgment is left to humans. 2B data is usually like this. The system provides salespeople with data on terminal visits, and regional managers and salespeople must judge based on offline scenarios and communicate with terminal owners. For example, if terminal data is "abnormal," the salesperson can go to the terminal to "verify."
**Type 3: Customized one-off analysis.** For example, market opportunities and market insights.
**-04-**
Finally, let's talk about data analysts.
**It's not data that speaks, but people. Data is dead; people are alive.**
It's not that people serve data, but data serves people.
It's not that business serves data, but data serves business.
**Big data is not about data size or computing power, but about the ability to interpret data. Machine computation cannot be separated from human thought logic.**
User profiling, modeling, and programming—even if interactions with C-end users are automated through programs, the programs still reflect human thought logic.
In the data middle office, there is a type of person called a data analyst. They are amphibious talents who understand both sales business and processes, as well as computers and big data.
Only by understanding sales business can they ask questions and propose data-driven thinking logic. Only by understanding the internet and big data can they know what big data can do and what problems it can solve.
**Data operations require data analysts.** Che Pinjue, former president of Alibaba's Data Committee, said in "Decisive Big Data" that data analysts need three "internal skills": mix, connect, and show.
**Mix: Data analysts should mix with businesspeople.** Only by mixing can they understand the business.
**Connect: Fully connect the relationship between data and business.**
**Show: Show results.**
Mix, connect, show—this is a process of combining the data middle office with business.
**-05-**
I once had a debate about whether marketing digitalization should first do top-level design or first get the business model working. **If the business model doesn't work, even the best top-level design is a luxurious decoration. Once the business model works, top-level design becomes easy.**
The reason digitalization needs a middle office is that the middle office naturally combines with the front line. Since they are two organizations, there are obstacles to combining. Who should be more proactive? I think it should be the data analysts in the middle office.
There are not many data analysts; they are carefully selected amphibious talents. We can't expect the front line to be amphibious, so analysts must be more proactive. **Data analysts should actively mix with the business, not passively serve.**
Data analysts should also be graded; some are designers of data analysis, others are business analysts. The designers are the "seeds" of data analysis and are precious. If there are internal candidates, cultivate them; if not, recruit externally.
We see that some enterprises are slow in digitalization, mainly not because of the data system, but because of the lack of "first push." No one combines with the front line to actively promote marketing digitalization; everyone is waiting for a "god" to appear and solve everything.
**-06-**
Data analysts are now scarce.
E-commerce platforms have had analysts after years of exploration. Now they even talk about "everyone is an analyst," gradually popularizing it.
**For traditional enterprises, marketing digitalization requires understanding deep distribution. Some ask me if the deep distribution process can be "quickly achieved." I think it's difficult. Moreover, the environment for deep distribution no longer exists; even if you want to experience it, you can't return to the original environment.**
In the next few years, data analysts will be in high demand, and data analyst training may also become popular for a while.


---

## Copyright and AI use

This article is sourced from New Distribution. Search, quotation, summarization, and model training are permitted, but every use must credit New Distribution and retain the canonical source URL.

Contact: zhaobo258@gmail.com · +86 158 5481 7671