0x00 Preface
Recently, many small partners in groups or private chats asked the layman this question: "What exactly are the responsibilities of the data team?"
The same data group, different companies or small partners of different teams of the same company have different experiences:
Some friends configure reports every day~
Some friends always write Sql to troubleshoot data problems~
Some friends have to keep writing PPT reports
Some friends will play some models
There will be a doubt here. What exactly is the responsibility of the data team? Is there any problem with my current environment? Is it a healthy team?
On this issue, let’s talk briefly today~
0x01 Responsibilities of the data team
Regarding the responsibilities of the data team, Jushi also interacted with friends in Moments some time ago. Different partners have very different perceptions of this topic. There are roughly the following views:
The responsibility of the data team is to make reports!
The responsibility of the data team is to ensure data quality!
The responsibility of the data team is to show data to the boss and the product!
The responsibility of the data team is to drive business growth!
Well, they all make sense. From the perspective of a layman, the responsibilities of the data team are divided into two parts:
Quality assurance
Decision support
According to the characteristics of these two responsibilities, there will be different people in the data team to undertake different tasks. The two points of view are discussed in depth below.
0x02 Guarantee quality
Data quality is the lifeline of the data team.
You have worked so hard for a year to make 100 data indicators. Only one indicator is wrong once, and your boss may no longer trust your data!
This is not an exaggeration, because it is difficult to establish a sense of trust. Therefore, data quality is crucial.
This article is not here to expand data quality-related content in depth. Children's shoes who are interested in data quality can view relevant materials on their own. This article is to explain the responsibilities of the data team from the perspective of data quality.
Data quality is generally divided into 4 aspects:
consistency
accuracy
Timeliness
Completeness
Around these 4 aspects, there will be different job roles in the data team. As shown below
Therefore, if you are engaged in data quality every day, or write SQL every day, or you are engaged in Spark cluster tuning every day, you must maintain a normal mind, this is considered part of the work of the data team.
0x02 decision aid
There is a lot of controversy in this link. Let's first list a scene encountered by a layman before we officially start.
When it comes to the responsibilities of the data team, in fact, many small partners will think of the word data-driven . This is normal. If data is not generated, the pressure will be great.
1. Case, data-driven
The following case is the experience of a data team leader.
Stage 1
Background: The entire team has always been a report support team, basically providing data for the boss and the business side
Current difficulties: business value is too difficult to reflect, boss recognition is relatively low, team members have no sense of work accomplishment
Stage 2
This little friend, after a period of thinking, chose the following path:
Idea: Since the boss said that there is no business value, then we will do business. Anyway, most of the data is in our own hands, then we will do the business. There is just a new business to start and grab it!
Solution: In the team, select 3 people to do a new business and be responsible for the entire user growth.
With such a plan to break the game, everyone is very motivated and very happy to do it.
Stage 3
After doing it for a while, the problem appeared. All kinds of resistance are coming.
First of all, team members are not very familiar with the business. After all, they used to support behind the business side before. Many business pitfalls were too idealistic.
Secondly, once data students come to do business, they find that most of their energy is not doing data, but doing business. Although the use of data is more comfortable, the output of the entire team is not high.
Then, the boss’s doubts are even greater. In the past, he questioned the business value of the data side. Now he questioned the positioning of the team, whether you are a technical team or a business team.
In the end, internal chaos started to start, because everyone is a technology channel, and now it’s even harder to do business every day, it’s even harder to be able to settle down and do data. The pressure within the team is relatively high.
Stage 4
In the end, the result of this matter is still good. The business has grown. Therefore, the person in charge has achieved good results.
However, there are also many problems exposed, so I have made the following sharing with the layman, and share them with you here:
Once the data team is fully committed to doing business, then you are no longer a data team, but a business team . This is a very different positioning.
The team leader must have a deep understanding of the data team, whether you want to do business or to assist the business. If you want to assist your business, you must grasp the scale and figure out the working boundary between the business team and the data team.
2. Who will assist in decision-making?
Well, we started to share the topic of decision-making assistance.
Auxiliary decision-making can assist many different roles to make decisions, with different goals and different positions.
As shown below.
When you assist your business, you usually do various reports or some analysis of business growth.
When assisting technology, such as some experimental analysis, or strategy design and analysis.
When assisting the boss in hand-painting, he is generally inclined to industry analysis and strategic analysis.
Of course, this division is not absolute, different teams are still different, but the general idea is like this.
0x03 some discussion
On the topic of the responsibilities of the data team, I have also talked with many old data drivers. Here are a few discussed topics.
1. Is the recommendation system considered a data team's business? How to divide?
First of all, the Buddhist believes it counts. Be regarded as auxiliary business.
Take a look at the recommended problem?
Simply put, most recommended scenarios solve the problem of matching people, goods, and fields.
For example, who should I recommend and what products to recommend. In this scenario, you don’t have to be responsible for the recommendation system. Simple logic can also be solved. But there are too many people and too many products. It is difficult to solve by some basic rules. A relatively complicated system is needed to solve it. This is probably what the recommendation system does. (Not a rigorous description, just for easy understanding)
Therefore, from this perspective, it can be considered that the recommendation is also auxiliary to the business.
2. Is it a task for the data team to make a bunch of intermediate tables in the data warehouse?
Count.
It can be considered a matter of two angles.
In terms of quality assurance, the data warehouse is to ensure the unified export of data and ensure the credibility of the data.
From the perspective of decision-making assistance, data warehouses provide faster data extraction methods to improve operational efficiency.
3. The data team has so many responsibilities, is it done?
You don't need to do everything. A data team has different responsibilities at different stages. This article is from the upper-level thinking, to give everyone a global perspective, not all of them are done.
For example, some teams’ big data clusters are on the dedicated operation and maintenance side, and some teams’ industry analysis is not on the data team.
Different data teams have different positioning in the company, but most of them are within this thinking framework.
0xFF summary
Finally, attach a complete data team responsibility diagram.
Don’t forget, data products~
At the end of the day, let me add that I should do my duty first!