Core indicators: uv, conversion rate, customer unit price, gross margin, promotion roi, repeat purchase rate.

Based on the core indicators, gradually make detailed indicators for media, users, commodities, marketing and other objects; at the same time, monitor the internal operation performance: customer service, commodities, warehousing and logistics.

There are two levels of data analysis:

First, website data analysis, for the product.

On how to run around the product, do the analysis of the closed path. Derive whether the product's click is smooth, whether the function display is perfect.

At the same time, collect and analyze the current sales of the largest proportion of several products of the conversion rate, traffic, inventory, replenishment cycle, price, and discounts and so on.

Second, the focus of the study of the customer's visit, to tap the potential needs of customers.

If it is a transaction-oriented e-commerce site, it is to study how to efficiently facilitate the transaction, whether or not a joint order can occur! Magic is a big data modeling platform, is a service bus and distributed cloud computing based on two major technical architectures of a data analysis, mining tools platform, which uses a distributed file system for data storage, support for massive data processing. It adopts a variety of data collection technologies to support the collection of structured data and unstructured data. Through the graphical model building tool, it supports process-oriented model configuration. Through the third-party plug-in technology, it is easy to integrate other tools and services into the platform. Data analysis and research platform is the collection of massive information, data model construction, data mining, analysis and finally form knowledge to serve in the field, the process of decision-making, the platform mainly includes the data collection part, the model configuration part, the model implementation part and the results display part. Slowly buy big data

Data collection and processing:

Using distributed crawler technology, search engine technology, natural language processing technology, hierarchical and layered data mining and labeling, to establish a link between the data and the business, and multi-dimensional analysis of the data model. I monitor the data around the industry, brand, sales, SKU data, and analyze the commodity data through the dimensions of comments, activities, and prices.

In addition to Jingdong, we cover a number of e-commerce platforms such as Tmall, Gome Online, Suning.com, Vipshop, NetEase.com and others.

Visualized data analysis: simple interface, visual and intuitive data query, analysis, prediction functions to help enterprises understand the market environment of the product crisis, timely grasp of competitors' channel operations, historical sales and historical price adjustment, promotions and so on. To help enterprises tap the potential market demand, enhance market competitiveness.