Data cannot replace business logic

Focus on intelligence data mining

Yesterday we mentioned the use of big data tools. On the other hand, intelligence data is also what e-commerce companies should really pay attention to.

For the processing of intelligence data, from the daily work, the collection of data and intelligence takes up most of the time. This includes communication with upstream and downstream supply chains and across departments. For example, a purchasing staff should go to the production line to analyze the quality level of each supplier, the production capacity, the difference in the production cycle between the excellent factory and the second-line factory, where the raw material purchase price is the lowest, the transportation time is short, and the transportation cost is stable. In general, such an intelligence can be used for one to three years.

Although the data is not strong, the value of these intelligences is very high. Talking about data mining is worse than intelligence mining. Intelligence mining can provide real productivity-level support for e-commerce companies. If intelligence mining is not done well, it is necessary to digitize and quantify it.

To give an exaggerated example, when a brand has 200,000 manufacturers who have no choice, in order to find a production company that matches the needs, it is necessary to establish a big data model for screening. Now, only intelligence is needed first, and data mining technology is used when the scale reaches a certain level and it is difficult to make decisions.

Indeed, the application of big data has to penetrate into China's e-commerce companies, and there is still a long way to go.

The marketing field is completely different. The data model of marketing has matured, and the Internet has brought enough information sources to e-commerce companies. The application of big data can directly provide advice to decision makers.

Take Taobao's original women's brand as an example. They will spend 500~1000 yuan per day for intelligence mining. The cost is already quite high, even similar to the daily consumption of some companies for Baidu. They have specialized intelligence collectors who analyze data based on Data Cube, Quantum Hengdao, and CRM systems, and then combine this information to assist with the most basic business decisions. Consider the next new product model, based on the analysis and needs of the old members. Analysis, whether it is necessary to expand new categories and so on.

For example, when it has 50 items and 1 million in cash, how should it be arranged for production? The intelligence miners will remind the decision-making level that there are 2 explosions, 6 long tails, 2 slow-moving items, and even suggestions for replenishment and clearance of various products, so as to adjust the price and adjust the position of the page. A series of subsequent actions such as push adjustment. It is not difficult to get the required data from the system, but the data needs to be further spliced, and then think about the causal link between the data.

It is common to understand that intelligence analysis in the business world is business logic.

"Intellectual support is an understanding of business logic, and data supports the ability to process business intelligence." We believe that we must first do intelligence mining and then do data mining. If the intelligence is not done well, it is equivalent to the understanding of the business logic is not up to standard, counting on the data to directly explain the business logic, some idiotic dreams.

Data cannot replace business logic

Big data needs to be based on quantitative data, plus business logic, to help e-commerce companies make global and systematic decisions. Excluding a series of uncontrollable factors and stripping the conclusions from the actual situation, the model in an ideal state is only the conclusion given by the mathematics expert.

The core of big data is the integration of business logic. In the business logic, we must first understand the market, understand the changes in the real demands of consumers in a certain field; secondly, understand the industry, including the characteristics, requirements and rules of the industry; finally, understand the operation of the enterprise, and have multiple support modules and resources. Integrate sequentially to create value together.

In the case of all of these, the quantitative data is used to assist decision-making, and under the leadership of business logic, the role of quantitative data is truly exerted.

"The lack of this business logic, the quantitative data is something that is unconstrained." We see business logic as a real problem that needs to be solved. Business logic will change because of different industries, different companies, different categories, and different timings. This is a dynamic and balanced art and philosophy. Although data cannot replace business logic, data can modify and adjust business logic. "The production of a decision depends on part of the data, part of the business experience, and part of the business intuition."

When it comes to this, it involves data stratification. Judging from experience, the more the data is on the macro-strategic level, the higher the practicality, the more micro-microscopic data, the higher the uncertainty. Because macroeconomic decisions are large, large to small effects can't work, and microscopic decisions are just the opposite.

For example, in the transformation process of "de-papering" to "online subscription" in 2014, the New York Times faced the pricing of different devices, different media, different time periods, and the free channels and free quantities given to cultivate user habits.

For example, the size of the entire industry, how the market growth power, itself is a multi-sample comprehensive data, the impact of each sample is only a part. Once at the micro level, such as the color of the advertisement, the strength of the discount, the amount of the full reduction, the data of a certain item will play a decisive role, so most businesses today believe that A/Btesting and other forms, to believe in data research and The combination of its own business logic.

We believe that "the macro level is more about the data, supplemented by experience; the micro level talks about experience and supplemented with data", and the combined content is valuable to e-commerce companies.

Returning to the essence of business, data is only a by-product of business, and the business system is good. Under normal circumstances, the data system will not be too bad. If the cart is upside down, the data system is good but the business system is poor. As a result, it will be found that the data content generated by the data system cannot be effectively accumulated and used.

This is not to say that the data is not important, but please don't be superstitious, because the risk of data uncertainty, such as the MACD gold divergence in the stock market, may also fall. The risk of business decision-making is unaffordable for most companies. Enterprises need to return to business logic and make corporate decisions through rich data dimensions and data content.

In short, through big data to explore, display their own business value, and make decisions, it requires scientific and rigorous planning and storage, rather than follow the trend of the big flow, after all, time and effort is also the cost of the enterprise, this is not a joke.