Question 2: What are the data sources of data mining commonly used in e-commerce? 1. Traffic 1. Search for vehicles: search for diagnostic assistants.
A- Basic conditions: If it is not illegal, you can check it in Seller's Workbench-Search Diagnostic Assistant-Baby Diagnosis.
B- relevance: category attribute relevance and title keyword relevance. C- popularity rating: whether to recommend in the window, whether to add consumer protection, DSR rating, Alipay usage rate, Want Want effect speed, and the time difference from auction to delivery.
Figure D: Many sellers often ignore the optimization of pictures when optimizing the main search traffic. However, the difference of click-through rate directly affects the final search traffic. Buyers don't search directly, but are attracted by pictures, so it is very important to optimize pictures. It is suggested that the through train can be used to test the pictures (the method will be introduced below).
E price and sales volume: products with similar sales volume, the higher the price, the more opportunities for display; Products with the same price have high sales volume and many exhibition opportunities. And checking this indicator is mainly to check the gap between yourself and your direct competitors, especially the gap in sales for 7 days, so as to make adjustments.
F-Title optimization: use long tail words when the sales volume is relatively low, and use common nouns and prefixes when the sales volume is high. After repeated tests, the maximum search conversion rate of search traffic is obtained.
2. Paid means of transportation: data report and store inquiry of all paid means.
-Taoke: The diagnosis of Taoke only depends on the difference in sales and commission between itself and its competitors.
Second, the conversion rate 1, conversion rate tool: shop inspection.
A- inner page: first look at the sales volume, then look at the evaluation quality, and then look at the single product conversion rate, page residence time and inquiry rate. If there is no basic sales volume, the evaluation is poor and the conversion rate cannot be good. The two premises have been solved, and then it depends on whether the single product conversion rate, page residence time and inquiry rate are not lower than the industry average (or the good treasures sold in the store). If it is lower than USP, optimize USP selling points, logical order (whether it is all around USP), display content and display methods.
B- Depth of visit: Since 80% of customers enter the store from the inner page, we mainly optimize the diversion position of the inner page, that is, store tactics, baby page association, baby page sidebar and store tail. Then optimize the home page.
C- contribution rate: whether it exceeds 80%.
D- Marketing activities: Regular marketing activities can improve the conversion rate.
Electronic customer service query conversion rate: whether it has reached the industry average at least. View tools: third-party tools such as store search. Optimization method: establish a standard answer for every question of customers. 2.DSR tool: Taobao DSR score calculator. Optimization method: a, upgrade Taobao's original service (7 days is no reason to upgrade to 30 days, 3 days to 24 hours delivery, etc. ); B, Taobao has no service innovation (innovation around the contact points between customers and businesses, such as SNS and games). 3.CRM CRM mainly depends on the proportion of old customers, the conversion rate of old customers, the secondary purchase rate, and the ROI of customer grouping short color emails. Tools: Seller Workbench-Member Relationship Management, Shu Yun, Customer Service and other third-party software. Optimization method: establish old customer groups and create different old customer rights according to the groups. The more advanced customers have the more advanced privileges.
Question 3: Why is data mining high? Now is the era of big data, and it is necessary to mine the relationship between data, so as to draw some rules. For example, when you shop online, Taobao discovers your behavior preferences through mining technology. When you browse Taobao-related web pages, it will give you your favorite products.
Question 4: The difference between big data and data mining Data mining needs a lot of interdisciplinary knowledge such as artificial intelligence, database, machine language and statistical analysis knowledge. Furthermore, the emergence of data mining needs conditions, the first condition: massive data; The second condition: the processing ability of computer technology to big data; The third condition: the storage and operation ability of the computer; The fourth condition: the development of interdisciplinary subjects.
Big data is only a condition of data mining.
Question 5: What do data mining engineers usually do? Job responsibilities:
1, according to your own understanding of the industry and the company's business, independently undertake complex analysis tasks and form analysis reports;
2. Relevant analysis directions include: user behavior analysis, advertisement click analysis, business logic correlation and competitive environment correlation;
3. According to the change of business logic, design the corresponding analysis model to support the development of business analysis.
Job requirements:
1, more than 2 years experience in industry modeling;
2. Bachelor degree or above, major in mathematics, statistics, computer, physics and other related fields;
3. Proficient in basic theory and data mining technology, especially regression model and decision tree model.
4. Skillfully use various data analysis tools such as SPSS Clementine /SAS EM, and be able to issue professional analysis reports;
5. Have practical data mining project experience in a certain industry of finance, communication or Internet, and have a deep understanding of the business of this industry;
6. Enthusiasm for the Internet field, strong learning and interpersonal skills, influence persuasion ability, and like challenging work.
Question 6: Which is more promising, big data or data mining? Big data includes data mining, which is one of the branches and the foundation of big data. If we study the direction of BI, data mining is the foundation. The two are closely related. The concept of data mining appeared earlier. We should know that data mining has been applied in early data warehouse modeling, and big data is quite popular in recent years, with a good trend. The future is the era of big data. At present, many large enterprises are doing big data (such as solution providers: IBM, ORACLE, SAP, EMC, Huawei, etc.). ); Self-research: Taobao, Tencent, etc. Party A: mobile, telecommunications, etc. The job search prospect is still very good, and the content of big data is very rich, including hadoop, streaming, distributed, NAS/SAN and so on. It will be of great help to your future development. My suggestion is big data. Hope to adopt.
Question 7: How to use big data mining to correspond to e-commerce data mining can discover e-commerce customers' personality and personality knowledge, inevitable and accidental knowledge, independent and related knowledge, reality and prediction knowledge. All this knowledge can be analyzed, and the consumer behavior such as customer's psychology, ability, motivation, demand and potential can be statistically and correctly analyzed, so as to provide decision-making basis for managers. The specific application is as follows:
The application of 1. classification prediction method in e-commerce
Classification is a very important task and the most widely used technology in e-commerce activities. The purpose of classification is to construct a classification function or classification model, which is usually called classifier. The construction methods of classifier usually include statistical method, machine learning method, neural network method and so on. These methods can map the data in the database to a given category for prediction, that is, using historical data records, the generalized description of the given data can be automatically deduced, so as to predict the future data.
2. Application of clustering method in e-commerce.
Clustering is to divide a group of individuals into several categories according to the similarity principle. For e-commerce, customer clustering can provide strong support for market segmentation theory. The purpose of market segmentation is to make the distance between individuals belonging to the same category as small as possible, while the distance between individuals belonging to different categories as large as possible. By extracting clustered customer characteristics, e-commerce websites can provide personalized services for customers.
3. Application of data extraction method in e-commerce
The purpose of data extraction is to condense data and give its compact description, such as sum, average and variance. , or by histogram, pie chart and other graphical means. More importantly, he discussed data aggregation from the perspective of data generalization. Data generalization is the process of abstracting the most primitive and basic information data from low level to high level. Multidimensional data analysis method and attribute-oriented induction method can be used. In e-commerce activities, for the customer data warehouse in e-commerce activities, the method of dimension data analysis is used to extract data. Aggregation operations such as summation, summation, average, maximum value and minimum value are often used in data analysis, and the calculation amount of such operations is particularly large. The results of aggregation operations can be pre-calculated and stored for use by decision support systems.
4. Application of association rules in e-commerce
The management department can collect and store a lot of sales data and customer information, analyze these historical data and discover association rules. For example, analyzing the buying behavior of online customers can help managers plan the market and determine the type, price and quality of goods. There are usually two kinds of association rules: meaningful association rules and generalized association rules, and meaningful association rules, that is, rules that satisfy minimum support and minimum credibility. The minimum support degree refers to the minimum degree that a group of objects need to meet statistically, such as the number of customers, customers' consumption ability and consumption mode in e-commerce activities. The latter is the least reliable association rule specified by users. The second is the generalization rule, which is more practical, because there is a hierarchical relationship between the research objects. For example, bread and cake belong to West Point and West Point belongs to food. With the hierarchical relationship, we can find more meaningful laws.
5. Optimize enterprise resources
Cost saving is the key to enterprise profit. Based on data mining technology, enterprise resource information can be grasped in real time, comprehensively and accurately. Through the analysis of historical financial data, inventory data and transaction data, the key points of enterprise resource consumption and the input-output ratio of main activities are found, thus providing decision-making basis for the optimal allocation of enterprise resources, such as reducing inventory, improving inventory turnover rate and improving capital utilization rate. By mining Web data, business information can be quickly extracted, so that enterprises can accurately grasp the market dynamics, greatly improve their ability to respond to market changes and innovate, maximize the use of human resources, material resources and information resources, rationally coordinate the relationship between internal and external resources, and produce the best economic benefits. Promote the scientific, informational and intelligent development of enterprises.
6. Manage customer data
With the "customer-centered" business philosophy deeply rooted in people's hearts, analyzing customers, understanding customers and guiding customers' needs have become an important topic in enterprise management. Based on data mining technology, enterprises will make maximum use of customer resources to analyze and predict customer behavior. & gt
Question 8: How much is the R programming data mining service? I found the name and price of the "Big Data Tribe" store in Taobao, and the evaluation was quite good. Judging from the difficulty and workload of data service, what kind of treasure buyer do you want to send specific requirements to, and he will judge you. Generally, such a price on any treasure is a unit of measurement, which is actually a multiple of 20 yuan.
Question 9: What is the promotion system of Taobao shop operation? Promotion system of Taobao shop operation;
First, e-commerce strategic planning
Based on data mining, through the 360-degree insight analysis of the market, competitors, consumers and enterprises themselves, the overall e-commerce model, overall strategic objectives, development stage steps, investment and expected benefits of enterprises are planned, and the ideas and directions are clear.
The functions of the project are decomposed to form a Gantt chart of project progress control, and the detailed strategic implementation plan is subdivided into executable, supervised and controllable.
Second, the store planning and decoration
On the basis of overall analysis and planning, first-class Taobao shop planners and first-class UI designers are established, and the overall architecture, column division, process experience and visual style of the online store are integrated and planned, highlighting the brand temperament style of the store and the customer shopping experience.
Third, product planning.
Adopt a comprehensive system of USP (unique sales proposition) planning +FABE mode+brand planning. Combining the characteristics of the industry and the cultural characteristics of Taobao Shopping Network, through the organic combination of perceptual and rational ideas, the most marketable product baby page is planned and designed, which effectively improves the product conversion rate.
On the basis of data mining, through the matrix planning and pricing system planning of star hot-selling products, Taurus profitable products and sniper products, a complete product width and product portfolio are formed to achieve the balance and unity of hot-selling and profitability, and solve the problem of online and offline channel conflicts.
Four, commodity promotion operation
Use Taobao's various promotional activities, plan various theme activities of creative shops, related sales, cross-selling and other means to realize the vividness of goods, enhance user stickiness, increase customer unit price, create explosive products, and finally realize sales leap.
Verb (abbreviation of verb) promotion and operation
The BRICS Taobao promotion and operation system takes the introduction of target traffic as the core, and adopts free promotion in Taobao station, advertising promotion of tools in Taobao station, and auxiliary promotion of the whole network. , systematically solved the traffic problem of Taobao stores and brought a large number of effective target buyers to the stores. Under the guidance of strategy, we will achieve the maximum promotion effect with the minimum investment, never blindly follow the traffic, let alone promote the invalid traffic, and achieve the dual effects of sales and brand promotion.
Sixth, customer service sales
Customer service sales is the key link to realize sales and has a core position. The BRICS will carry out standardized and systematic operation from four levels: business, culture, management and training, so as to realize the process and reproducibility of the sales customer service system.
Seven. data analysis
Data mining and analysis are the most obvious differences between e-commerce and traditional offline commerce. The data of e-commerce is accurate and real-time, and the foundation of Taobao operation system of BRICS e-commerce is data mining and analysis.
Through the horizontal, vertical and cross analysis of all kinds of data, we can formulate strategies to improve the promotion effect and the conversion rate of the store, so as to improve the ROI of the whole store and maximize the profit of the enterprise.
The above views on Taobao project operation are only a brief analysis of my personal suggestions from the institutional level. BRICS believes that Taobao's e-commerce operation should be based on data mining, with improving the conversion rate of stores as the core, starting from strategic planning, online store planning, product planning, product promotion, Taobao promotion, customer service sales, data analysis and other aspects, and system construction can win!
Question 10: Is data analysis a "needle in a haystack"? Guide: How does big data generate value? Is big data everything? Where is the application boundary? It seems that everyone has a vague concept about these questions, but there has never been a unified answer. Today, the discussion about "big data" has reached a peak, and data is the center of the new strategic development of Internet companies in the future. What is big data, how does big data generate value, is big data omnipotent, and where is the application boundary? It seems that everyone has a vague concept about these questions, but there has never been a unified answer. When it comes to big data, Ali should be the first to bear the brunt. It has been working around the data ocean for a long time and has derived financial lending business. Ma Yun merged the two core businesses of Ali Financial and Alipay into the group to form Ali Microfinance, and arranged Peng Lei, the successor with the highest voice before, to take the helm of Ali Microfinance. Ma Yun's emphasis on the future data battlefield can be seen. As the data platform of Ali Microfinance Services Group in preparation, Feng Chunpei, the person in charge, also has unique opinions on the data. He told the author that the discussion on big data in China focuses more on the technical direction, that is, "how to precipitate data", and pays less attention to the application of data. How does data generate value? This needs to start with the essence of big data. Online data is big data. To figure out what big data is, we must first know what kind of data is useful. According to Feng Chunpei's understanding, any behavior itself will produce data, but only online data can be precipitated and utilized. "For example, if you don't go through Taobao, people's trading behavior also generates data offline, but this trading behavior is private. No one knows my trading behavior except the buyer and the seller, and both sides are anonymous. From the nature of data, it cannot be precipitated, and there is no way to effectively collect it from the source. " What is big data? Feng Chunpei's understanding seems to be closer to the essence: "The essence of having data is that you have a more comprehensive and clear understanding of the world, these people, these enterprises and this era. You can understand the needs of these people and you can understand any changes in the world. " You can understand it this way. If you are a deep user of Ali (such as Taobao seller), and they have enough data about you, your credit evaluation will be more comprehensive. These data can not only play a role in the financial field, such as helping you get small loans in Ali, but also reflect your credit status in life. "For example, blind date, how do you prove your income? You take out the bill of Alipay, and the girl spends 1 10,000 yuan a year. You said that your credit is good, and your credit card repays on time every month. Much more useful than you? " Data is the means of production. If the data is only used as auxiliary reference information, it is necessary to invest so much energy. What is the role of data in factors of production? Feng Chunpei's definition is "means of production". "The name of our department is the Business Intelligence Department, and the data is more like an auxiliary decision of business. As a "staff officer", we must gradually integrate these data into the process of our business and products. Data and business are like two gears that can be locked together. When we dig and understand data more and more strongly, the final data can not only generate value, but also directly give birth to products, such as some data of Ali Finance, which we define as means of production. " This is what Ali will do in the future, turning data into means of production. Different from the traditional means of production, data can be used indefinitely, and the more you use it, the richer it will be. Recently, Alibaba has made frequent efforts in the mobile Internet market. In the future, it may be possible to fuse data, and all kinds of information of users can be presented in a panoramic view. Even in a completely strange city, with this service, you can know which store nearby supports Alipay payment and which netizen in Weibo has just stopped at a nearby coffee shop. Data analysis is "looking for a needle in a haystack", just like most Internet products. The data generated by the Internet may be forged, but it is also disordered and fragmented. In this regard, Feng Chunpei has made no secret of it. "It is of course possible to falsify data in a short period of time, and it is entirely possible to falsify data in a specific dimension. However, because our business is based on long-term data for tracking and analysis, and the dimensions adopted are wider, the cost and difficulty of data fraud will become greater and greater. According to our current credit model, the income from forging data is unlikely to cover the cost, so we can basically judge that the data ... >; & gt