Shulou(Shulou.com)06/03 Report--

Only the data best understand the pain points of the company and guide the decision-making direction of the enterprise.

At the 2018 China Big Data Summit Forum," rel="nofollow"> CEO Feng Jian shared his understanding of data capitalization, the construction of big data platform, challenges in the process of big data landing, and opportunities and challenges for data application in the future. The full text is extracted as follows:

1. What is data capitalization

"Data capitalization is something that Shuolan has always adhered to and continues to do."

Feng Jian once saw many application scenarios when he was in charge of Alibaba Group's big data business. He feels that most people don't know enough about data at present, nor can they understand where data comes from, what value it has, and where it supports future business.

For example, if we are a mobile phone factory, we have data in various fields and types, and we produce data every day, but in addition to producing data, the enterprise also wants data to be directly applied to the business field to optimize business results. This is the difference in cognition, that is, what value data can bring. The process of turning original data into data that can be used in business is data assetization.

Of all the assets in the world, only data assets are more valuable the more they are used. Because it's so close to business, if you don't use it, it becomes a bunch of numbers, and if you use it often, it becomes more and more valuable.

We have two key technical points in our own practice process: the first point is to integrate the data. Many enterprises have dozens of business systems provided by different manufacturers. How to connect and get through these data, including mail, video and voice, is a problem that the whole industry is solving at present.

The second point is to make these data really effective after the data is opened up, which can be seen, identified and used in the business. This is a particularly big challenge facing the whole industry.

We call the process from data opening to data organization to data labeling and then to internal systematization of data assetalization.

Second, the construction and application of big data platform

If you look at data assetization from another angle, you can see how platforms are built.

Now shopping malls have detection technology, when customers come will know who this customer is, these data is very valuable, but they are only data assets, not data assets. What is data capitalization? When someone enters and exits, the relationship between people, things and scenes of such behavior is matched, pulled to the dimension of history, and depicted to every time node of history. This is an event based on space-time dimension. In all space-time dimensions, this process itself is assetization, and the result it can bring is that all behaviors of a person appear. For example, this person used to go to shopping mall to eat, watch movies and go shopping alone. Suddenly one day it became two people going in and out together, and another day this person went to the mother and baby store after watching the movie. This is the process of data capitalization under a scene. It can be seen that there are only three native data, but the data assets brought will be enriched to hundreds of thousands of dimensions, and even a person can be characterized as shopping psychology of shopping around or hesitating. From the perspective of data, the value is huge and boundless, which is data capitalization.

We have a series of methodologies to support the completion of data assets. The first point is to string these data based on scenarios and space-time dimensions, otherwise there will be no reference value.

The second point is that our big data platform can process, develop and model customer data. From the business platform level, it is data data-mapping. How to make good mapping data into profile is to make a full-system portrait of the whole dimension of detecting data to people. This is data development.

The process of data development does not need to change the customer's knowledge structure, nor does it need to change the previous data storage structure. The data platform is a full-system and full-dimensional data development platform. This is not enough, there is also a lot of data to be made into a label system, but how to quickly generate applications is still a problem that needs to be solved.

For example, marketing, or perhaps consumer insight portraits, we call these data applications. We need to make a user portrait to see whether these people are bad or good, using personal data assets plus data technology, and then through algorithms and model calculations, draw this person out, this is the user portrait ability. We encapsulate this capability into an entity called a data engine. Data plus data technology constitutes the data engine, and then presents various data engines in the platform, which is particularly convenient for customers to use.

After having the development platform, data engine and own data assets, it is necessary to build a data application platform and service platform, which includes three core capabilities, including data-mapping, data-profile and data-service. If you combine these three points, a data platform is basically formed. The construction of our whole platform is also based on this theory of ShuLan's own. We can put in the judged data, connect time and space and scenes, make it into a set of profiles based on the development platform, and then divide it into some data services based on a set of ShuLan's own data technology. Whether it is to B for physical stores or enterprises or to C for consumers, there are suitable application scenarios, and finally make the data live and use through the data platform. Of course, this is the ideal we have always wanted to achieve.

Third, the time reversal theory of big data

The challenge here is cognitive data-mapping, how the data is mapped, and actually the cognition of the data.

When we do data services in the real estate industry, we find how to use big data to improve owner satisfaction, which is a process that requires data recognition.

In everyone's imagination, it is very simple to improve the owner's satisfaction. It can be analyzed through the data such as whether the owner has complaints and whether there is repair. However, traditional satisfaction analysis uses only a single data, which brings limited value to the improvement of owner satisfaction. When the real estate company found us, we hoped to use big data to improve the satisfaction of owners. At that time, we put forward a hypothesis that the whole real estate company had five service bodies: owners, property companies, suppliers, contractors and service providers. If the relationship between these five subjects can be constructed, then the improvement of satisfaction can find an entry point, such as taking the whole family to see a house, a total of five people to see, three people satisfied, two people dissatisfied, whether satisfied or dissatisfied, there is no quantitative standard, there is no right or wrong distinction. But when I construct the entity between these five relationships, find the relevant data for any event, where the intersection of this event starts, then we can understand. And that's our ability to drive this event--our ability to recognize data.

For example, if I buy a new house and I'm not happy with it, I'll call and complain about a lot of problems. The owner will describe all the problems clearly in a phone call. This process will generate a lot of unstructured data, such as whether the owner has family members, children, wives, old people, etc., which faucet leaks, and then who produces the faucet, who serves, who is the property company, service cycle, supply cycle, service quality, etc. After all the data are strung together, you will know whether this is a batch problem or a personalized problem.

For example, if the owner says,"You have to move the toilet five centimeters to the left," the property company will say that the house is finely decorated and delivered, all floors, even the whole community. But why did the owners move? He'd say,"my kid's bathtub won't fit. Do you want to move it? "The property may say that I will move it to you, but this project is particularly huge. The information given here is that he has children and needs to put a bathtub to bathe them. This is the root of the problem. Data must be deeply and multi-dimensional insight in order to approach the essence of facts. Sampling analysis alone cannot find such problems.

I call this ability event-based mapping, and this is often the case with large amounts of data. Continuously moving the events of each time node onto this Timeline and extracting the common data of such events, we call this method, summarized as time reversal theory.

4. Only the company with the most data understands you

Many businesses face the problem of not knowing where their data is worth and how it is used. This is a product application problem of data, many enterprises will take it for granted to make a product and then tell customers what pain points this product can solve. I told the employees of Shulan that any of us talked to customers and tried not to tell customers that our products could solve his business pain points, because we didn't understand his business. Real estate customers have been doing it for decades, and the pain points of their business are not something we can quickly understand. But data can, data can best understand the company's pain points, and even judge the company's subsequent decision-making direction.

Just like the example I gave earlier, when doing real estate, after I get through their complaints and work order information, we can use big data to improve the satisfaction of owners. As a result, customers find that they can still do many things, such as fund management, supplier capability management, service provider management, etc. Owner satisfaction is not satisfied, only the data can know, we call this process based on data asset scenario drive. We like to reason with customers and tell them that I don't understand your business, but your data can understand your business. If you give me the opportunity to turn your data into assets, you will find that his value will be very high when you do data business based on this data asset.

We first reverse to the customer's business as traction, all the customer's data assets delivered to the customer to meet the customer's business scenario. This kind of application is particularly widespread. I give him a data asset with a label prompt. He can do countless data asset applications to analyze where the traffic of this store has gone, how much it contributes to me, which ones do assets and which ones do customer analysis. In fact, it is still a cognitive problem here.

I estimate that in the next two to three years, the biggest obstacle to most big data applications in China is the cognitive problem of data and scenario.

V. Data application entrepreneurship needs to focus

At present, Shulan will still tend to traditional companies, because after decades and hundreds of years of traditional companies, they know very well where their shortcomings are and what they should pay. When we are guests, we only do traditional companies, real estate, automobiles, retail, chemicals, we are involved.

In the wave mouth of the data age, there are many traditional companies catching up. They have completely missed a lot of things in the Internet age, and they don't want to miss the data age in the big data age. Unfocused companies can be OK at the beginning, but they will lose competitiveness later. We are currently focusing on retail and real estate, but maybe next year we will have a finance division.

VI. The data age is scene-driven

Recently, I have discussed this topic with many people. We understand it this way. We will now be more cautious in raising big data. We call this era the data era. In the IT era, IBM or Oracle, all the core of that era was driven around requirements, but the data era was driven by scenarios.

Judging whether a company is a big data company is a standard, whether the product you provide is demand-driven or scene-driven, and if you sell a mature product, it is a typical demand-driven, which can only meet part of the customer's needs. However, the arrival of the data age is particularly strange. It is driven by scenarios. Can your data products provide support for me? We are also cautious when choosing. One conclusion we have discussed these two days is that consulting service is a particularly important link in the early stage of the data age, that is, a set of theories and methodologies are constantly summarized and discussed, and constantly told that he should do so. Only then can we slowly move to the level of data operating system.

We believe that there may be an inflection point in the next two to three years. Many references are dominated by big data consulting business. No matter how well business products are done, they are all beyond single-point business, including AI products. More extensive consultation is still the mainstay in the next two to three years. We have 3331 plan, in the next three to five years there will be a big wave of primary, that is, general data products will come out, it may drive the application of data in an inflection point way, but we don't know what this data is like, we have been exploring this data form.

About Wind Sword

The founder, chairman and CEO of Shulan Technology, a top data application scientist, an advocate of international leading big data concepts, and a pioneer in domestic big data application practice.

Gan Yunfeng (Feng Jian), Chinese nationality, founder, chairman and CEO of Shulan Technology. He has worked as a data architect, data scientist, and data business leader in Huawei, Kingdee, and Alibaba. In 2012, he began to be responsible for creating the core data interconnection technology of Alibaba Group (ID-MAPPING), data asset system (TCIF), data value marketing system (DMP), urban intelligent brain (AI project), etc.; precipitate international leading big data ideas and cutting-edge technology research, lay the basic theoretical framework for big data application construction, and build the first systematic and comprehensive big data application platform in China on this basis; In 2015, as one of the first batch of big data application practitioners in China who flexibly applied big data concepts and big data technical capabilities to various professional fields, especially traditional industries, it was recognized and favored by customers and capital markets in more than 20 industry fields.

Small partners who want to know more about big data can join the digital community yo, here are the most complete data platform/visualization/Hadoop and other related articles ~

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.