Take off the big data coat and get out of the big data dilemma

"Enterprise Observer" reporter Jianwen

If you want to choose the top ten hot words that appear in the media in the last two years, big data will be chosen without any surprises.

However, it seems that everyone is aware of the power of large data but cannot find a way to use large data. Whether it's a business, a government or any other community organization, big data is so close, but so far away.

"Big data has become a hot word for people to talk about, but few mention the ability that people need to have in this huge change, and countless businesses and individuals forget to ' count '." ”

As the vice president of Alibaba Group's Business Intelligence Department, the forerunner of domestic big data practice and data observer, with more than 10 years of rich experience of data combat, perhaps the domestic for how to use large data this question the most right person.

"The future is the era of big data, the future competition is the data competition, perhaps, we should forget those flashy noise, let big data really from ' see ' to ' use ', really live." ”

In order to let more people learn how to use large data, car products feel writing "Decisive battle data" book. In the book, he tries to show the nature of the big data he understands and how to get companies out of the trouble of "using" Big data.

"As a person who has dealt with data for more than 10 years, I know deeply that it is a complex process, from ' looking ' to ' using ' and ' using ' to ' keeping ' data." What we should do at the moment is to forget the concept of big data. ”

"I want to start with a practical perspective of the ' fog ' of big data and tell everyone what the specifics of the big data should be, what we want is not the amount of data, but the quantity of ' quality ', which is the important purpose of writing this book. ”

Take off the big data coat

Research, practice of large data for many years of car, for large data has its own unique views. In his view, the big data is like the King's New clothes: everyone is saying beautiful words, but no one saw the dress.

"There are a lot of people on the web talking about big data, but they just talk and they don't." The so-called big data experts have never been done. ”

For the reason why there is no good way to do, the car product sense "big battle data" a book said: The people with data do not know where the big data come from, do not know how the data to use large data.

"The people who use it are afraid to use it because of the authenticity of large data; People do not know how to use it because of the complexity of large data." The result of this problem is that the volume of data becomes more and more large and is increasingly unusable. ”

"Big data is never a free lunch, and with the advent of the big data boom, there will be a lot of questions about big data--big data mixed with false information; The source of large data is multiple channels, bias, random error always exists ..."

Apart from the above problems, the fault of the talents is the most serious problem faced by the big data.

"Now, the people who collect the data do not know what to do with the people who use the data in the future, which is the big life of the current data," he said. ”

Cheping explains that when using large data, our usual practice is to collect the data first, because the data may be useful in the future. However, "the future may be useful" is doomed to raise a problem-the people who collect data do not know what to do with the data in the future.

"At this point, if you ask the data collector How to better collect the data, then the use of the data will fall into a dead loop." ”

In the book, the car's sense of the search for cold medicine to support his assertion, he wrote: For example, I found in the search engine somewhere to search for "cold medicine" very high frequency, so I decided that the place may have influenza. But is this the right way to use large data?

The answer is no. Therefore, Cheping says, the person who produces the data does not give the user some parameters for this data, and the user will face loss if the effect is not good after use. And such information is asymmetrical, and it is the people who use the data that are ultimately compromised.

The question is the answer.

In the sense of car products, if you want to really use large data, you need to see how big data can bring the value of the enterprise, and its essence is what.

"When you return to this problem, if you do not understand it, then even if you have more data, you will only be unprepared." ”

So, what kind of value can big data bring to the enterprise? What is the nature of big data? In the "Big Battle Data" book, the car products feel gives its own understanding: large data can allow enterprises to allocate resources reasonably, as well as to bring a better user experience, and the essence of large data is to restore the real needs of users.

"Data collection realizes the rational allocation of enterprise resources, for example, the recommendation system to do better, so that more users can have a higher probability to find the goods they want, so as to improve the purchase rate of goods, to create greater economic benefits for the enterprise; Data collection is also a promotion of customer experience, for example, Convenient search engines make it easier for customers to find the goods they need and create a better shopping experience for them. ”

As for the judgment of the nature of the large data, the vehicle perception in the book makes the following assertion: "The essence of the data is to restore, this is the key method of collecting metadata, without this concept, you do not know what data you need in the future, even less understand what is important data, to the end will only produce more and more unrecognized data. ”

In the sense of the vehicle, this reduction of data is like a sight, aiming at the consumer's shopping behavior, and recording it in the form of data.

"How to better understand the different needs of users in different scenarios, how to better understand the value of data fusion, will be a future business in every enterprise must consider the issue." ”

The Secret of Alibaba

As Vice president of Alibaba Group and Chairman of the Data Committee, vehicle quality awareness is one of the few people who have a say in Alibaba's transition from IT strategy to DT strategy (Big Data Strategy). In the "Big Battle data" book, he summed up Alibaba's experience in the implementation of the DT strategy to share to the successor.

"When it comes to the data operations of Alibaba, the first thing I think about is ' people ', and we spend too much time talking about what we should do, but we rarely turn to the idea that if the data operation falls to the ground, it starts with ' people '. ”

Therefore, in "decisive battle Big Data" a book, Cheping first with the reader share, is the Alibaba data operation's internal strength-"mixes", "Pass", "bask" this axes.

The so-called "mix" refers to the data analyst to the management of the business unit to mingle. Cheping that, as a data analyst, if you don't mix with the business, you don't know what the business unit is doing, "the business sensitivity is ' mixed ' and it doesn't appear in front of you." ”

"Only commercially sensitive data analysts will know what data to use to drive companies to achieve their business goals, and now in the company, I don't want to see data analysis teachers sitting in their positions, and I'd rather see them blending in with the business people." ”

And the so-called "pass", it refers to the enterprise to determine whether the data is a valuable ability, if done this, it has done "tong".

"Sticking with business issues to observe the data or to observe the business with data, both sensitive, is to achieve the ' pass ', some people in a very short time to determine whether the data is valuable, is because of ' mixed '." ”

But it's not easy to get through the data. In the book, the car's sense raises three key points to get through the data: the first is to do a good job of data security to ensure that employees in different positions within the company can view different data, and then unify the data standards of different departments, so that the company's internal data have a unified interface to avoid confusion; Finally, correlate data from different departments, Create opportunities for data operations to spread beyond the sector.

After the data, Alibaba is doing is "drying" data. "On the ' sun ' data level, it's often the data that answers these questions: good or bad business, how data changes can make business better, how to use data to help businesses discover opportunities, and even generate new business value." ”

"Mixed", "Tong", "" This axes, is Alibaba in the implementation of large data strategy, the core magic weapon, Cheping said: "They are in fact with the data methodology and people's cultivation, can do to borrow things, so that people with data growth in the data, step-by-step to let everyone become data analyst.

Of course, this axes is just the kernel part of Alibaba's big data strategy. In the book, it also mentions the extension strategy needed to implement the big Data strategy, and the best way to realize the big data is to read the big data of decisive battle directly.

5 Great value of data

01. Recognition and Series value

The data that distinguishes relationships and identities is the most important, how much of this data should be stored and never give up. In the large data age, the more can restore user real identity and real behavior of data, the more can let the enterprise in large data competition to maintain strategic advantage.

02. Descriptive value

Description of the business data including turnover, number of transactions, site traffic, web site details of the traffic, the number of sellers, etc., we can through the data on the business description to observe whether the transaction activity is normal.

03. Time Value

The time value of the data is the most direct embodiment of the large data application, and through the analysis of time, it can well sum up a user's preference for a scene.

04. Predictive Value

The forecast value of the data is divided into two parts, the first one is the prediction of a single product, the second is the forecast of the operating condition, that is, the forecast of the whole operation of the company, and can guide the company's business strategy with the conclusion of the forecast.

05. Value of Output data

In terms of the value of data, many of the data do not have a particular meaning in themselves, but after several data are grouped together or some data is integrated, new values are created.