Shangpai CEO Li Jongwei: How to grasp the big data opportunity of electronic commerce

Li Jongwei, CEO of Shang-faction company

Li Jongwei: Good afternoon, everyone! Thank you for "business value" for this opportunity to share our experience of digital mining in the area of E-commerce. In fact, E-commerce has a relatively large characteristics of its data accumulation and database compared to other industries congenital will be a little simpler, the resulting application is very much, but in practice, if there is no good method then your accumulation of data is strictly a disaster, and we will discuss this in the following part.

What is the big data that the merchant understands? As an electrical dealer itself under the drive of it, any process, all of which we can understand as data nodes, but in what way do you save and in what number slice to see it? For the entire electrical business, if you don't understand this area, your data will not be available in the future. Commodity is an E-commerce software company, our current cumulative service customers total more than 1.33 million enterprises, the average daily production of daily orders of 600,000 single, double more than 10 million single size. He would probably produce such a data accumulation, and if it's a very small one, even if it's on a global shift today, it's a huge challenge to the computing power of an enterprise.

These data we break it down into attributes, for our e-commerce, the system itself is accumulated by the application itself, the most natural form of data slicing and even data storage is not easy to be applied, which is likely to lead to late data availability of the recession is very severe. Big Data The first thing is to build your slice dimensions, such as browsing behavior, transaction behavior, payment behavior, communication behavior, social behavior, feedback behavior, and all the necessary technical indicators, an application system is likely to be at multiple scales at the same time response to data, so the understanding of data slicing, This is the core idea that you should build first when you have big data applications.

When we're doing big data applications, basically set up the thought dimension is relatively simple, the first data collection and storage, the second data organization and management, the third data analysis and presentation, especially to be emphasized in the statement just now, we are more also Shangpai in history paid tuition, When you don't have a clear definition of this, or if there are some flaws in the definition, the historical data is often unavailable, data mining and usability are very difficult, including the original number definition, the minimum granularity, etc., if done poorly, the data is almost no usability. Essentially, in a social environment, whether it is a payment system or an external open system, or the entire logistics system, an enterprise is essentially able to obtain a series of such data accumulation.

The development of the Internet brings the characteristics of user behavior data and even the entire exchange of interactive data also has a great accumulation, so today we understand the data has been completely not from within the enterprise simple business behavior or internal accumulation of their own enterprises, but the whole social environment in the accumulation of data. Another advantage, in these years, the technology of large data mining is relatively mature, when we look at the big data, when we look at the large data, we will understand that the Shangpai of the application of the data and the focus of the consumer are inconsistent, from the dimension of the end user use, What we can now see in the field of E-commerce is that it relies on data-directed operations to drive his business in a visual environment at this stage.

We can now see that common in a data analysis dimension, customers may exchange order acquisition, two sales, page return, billing cost, etc. conversion ratio, which is usually some operational indicators, their visualization can help enterprises improve business capabilities. The data source for these operational metrics is the need to slice the numbers in front of them. So we have introduced a number of application systems, such as ECAE, Betternow Visual Dashboard, now basically serving two directions, customer system stability and operational efficiency, through the constant monitoring of data to create a visual environment, At the same time in this environment to further the consumer-oriented level to enhance the price of traders and products recommended precision, especially in the commodity data, often this data does not have a single dimension of the match, because it is relatively difficult to make decisions, in a considerable number of times, we are in the consumer trade match, We must take into account a wide range of external supply systems, at present, usually a product in the market suppliers, such as a mobile phone products you can easily find thousands of suppliers, with the analogy of the quality of the supply to find 3-5, in such an environment how to protect your trading conversion rate, this should be done through constant monitoring.

We mentioned it in our conversation just now, in today's society environment, the enterprise's most complete data source is not the most original internal production of data, many are large data system integration, traditional data such as the internal accumulation of financial data, supply chain circulation data, CRM close customer contact data, And the behavior data of the whole customer, including the social data of the customer in the non sales environment, the combination of these information systems is we see the big data in the future will change, it is not in a static, closed environment operation. We will see that when your data dimension is like this, there will be a very big difference between the numbers, from the future of the enterprise operational capabilities, his access to the data and even the entire business operation of the full range of visual needs, we will see relatively strong direction. One of the problems we want to talk to you about here is that in the internet age, especially socialization, you are relatively discriminating against consumers, which in fact did not exist before, and we just gave a definition to the CRM within the enterprise, which is biased towards historical data, itself does not necessarily reflect the behavior of the entire consumer, this is what we are doing at the present stage, it mainly uses matching and tracking methods to help enterprises build dimensions.

Beyond this dimension, from the current point of view, we will see that when the data is accurate, direct-link response will be more effective, in turn, there will be a larger problem, such as the two customers we serve, we try to use large data in the way of the potential customer tracking, when your data slicing to do fine, Your whole preparation and preparation is done in a small section of your sample population, even in the data dimension, all of your coverage users have a high rate of feedback, social communication is strictly a double-edged sword, once the use of bad, customers will be your failure to spread the delay, this is a more troublesome thing. The series of dimensions that we set up on the data, at present, the first step of the commercial run, we are in the information and commodity level run more, before the electronic commerce up, we know that many traditional enterprises in the market may buy the whole market data, at present our customers can do the content, competition brand, Price of goods, description, sales and inventory fluctuations, to achieve near-punctual monitoring, currently due to a series of system problems, including the entire enterprise system pressure, we are placed in the previous, the entire time data is very accurate, in double one we can realize in the small household electrical appliances field Double 11 promotion system, More rely on a wide range of sales channels rather than direct channels to complete, and as cosmetics although the sales network is very broad, but the overall sales performance is not mainstream, so that the matching data behavior and the rebound in user behavior, you will find that there are some things are decisive factors. These are some of the directions we see.

Further orders, at present, we know the order of the harvest address can be in the entire logistics industry and even distribution throughout the entire process played an important role, including the choice of distribution network, at that time also includes payments, customers and so on, overall speaking, we said that the direction of the entire e-business industry will gradually promote our application. Our industry is a better industry, from the outside to talk about big data is the whole world hotter topic, from the E-commerce business itself, we currently have a better chance to direct access to these data, these data than previously collected more convenient.

But we are in communication, in the entire electrical business industry, data acquisition, collection, analysis is the need for consumer service providers and ultimately the data side must be cautious and serious topic, either side from our current view, Malicious or in the context of the lack of adequate preparation of the data tracking and use of the entire data environment disturbance is very serious. For example, we have launched a level of commodity radar, can help enterprises achieve commodity alignment, but there are some problems. In this context, we hope that in the future with all data suppliers, data partners including technical partners, including a wide range of open platforms can be stable and long-term cooperation, and ultimately to the entire industry to carry out solid data services, in the future to create real meaningful data value to customers, In order to help enterprises and even decision-makers in a visual digital environment to make the right decisions and judgments to guide the enterprise forward, thank you!

(Responsible editor: The good of the Legacy)