In the company, how can data analysts help the company generate revenue?

In recent two years, the concepts of data analysis and data mining are very popular. Teachers in various educational institutions like to demonstrate the value of data analysis by using XX credit card company in the United States or the old old terrier of beer and diapers. I have been working in the circle for a long time, and I know that there is only one way for data analysis to help the company directly generate income: to help Party B generate income! Because only Party B will sell data analysis, data mining, data products and data consultation as commodities. In this way, there are three ways to help companies generate income by doing data analysis: making products. For example, BI companies, big data companies, public opinion companies and credit reporting companies sell a whole set of data products. Data analysts actually play the role of product producers in these companies, so they directly help the companies to generate income and provide services. For example, some consulting companies, new media companies and big data provide data mining services, data collection and report writing services. These services are aimed at the needs of Party A's brand, promotion, marketing and other departments, so they can be sold. Here, the data analyst is actually a product producer, but the output is not a specific product, but a service composed of reports, excel, ppt, codes, conferences and so on. Do pre-sale Quite a few software companies and consulting companies will hire a data analyst to do pre-sales, because it is not competitive to talk about how good my plan is when fooling customers. We need someone who understands data and can analyze problems to make a quantifiable plan to convince customers. Here, the data analyst actually plays the role of sales, but this sales is selling knowledge, and impressing customers depends on professionalism rather than kickbacks. So you will find that the high salary of data analysis is basically Party B, or Party B's department in Party A's enterprise (such as Ali Data Bank and Intelligent Customer Service, which are Ali's projects, but still provide services to other companies as Party B) because data analysis is the direct productivity here. In party a? On the issue of income generation, data analysis has been ranked last. For example, Party A's father wants to produce a new product to increase his income. What does he need to do? Design products, production products, sales channels, brand promotion products, promotion of logistics and subsequent data analysis, and see how the results are ... Yes, you will find that the other six steps can be done without data analysis; Only data analysis does not have the first six steps, data analysis is a piece of waste paper, which is the embarrassment of data analysis in Party A. Some students will say: that data analysis can help enterprises design suitable products! But in fact, product designers can still design products without looking at the data. They have been doing this for more than 100 years, which gave birth to Joe's classic sentence: I never watch any market research! This kind of embarrassment is the inherent limitation of data analysis. Data analysis needs data to analyze, which is a post-processing process. The core of similar product design is creativity; The core of product sales is the ability and motivation of business team. The initiative of these people is the prior action made by the enterprise performance, not calculated, so the data analysis of income generation is actually very weak.

There is only one scenario data analysis that may be useful for income, that is, a business unit +B42 is really badly done, and something is wrong. At this time, if some benefits can be improved through analysis, they will not be happy. This is why many mature data mining projects are aimed at customer service outbound, SMS sending and EDM. Because in these places, the natural conversion rate is horribly low, and the copywriting, products and advertisements of business departments do not play much role. At the same time, these channels are point-to-point push, and the data accumulation and modeling environment is relatively closed. The data model can improve the natural conversion rate from 1% to 2%, and the business department has been thankful.

In fact, data analysis is helpful to enterprises, which is more reflected in posts, such as performance evaluation, result assessment and result optimization. Interestingly, many practitioners themselves don't want to understand this. For example, Fan Ruan also has the answer to this question. You can have a look. The examples given in it are all about how to cut costs instead of increasing income.

However, Fan Ruan's answer itself is very professional. Because cutting costs is easier to reflect the credit of data analysis than increasing income. Let's review the process of increasing revenue by the above new products. If the data analysis says that I did this performance, at least six departments will take credit with you. But if the data analysis says that there is a product here that is rubbish and can be cut down, then at most one department (the department that designed this product) will be offended, and the remaining five departments will still support you (because there is no need to waste time). Therefore, intelligent data analysis always proves value from the perspective of internal control, not from the perspective of external income increase.

However, this leads to the second embarrassing place, that is, should I do this for the last data product of wool? Even wool, I have to hire a data analyst to do this? Because the data of invoicing is also in ERP, theoretically I want to know which product has poor efficiency as long as a programmer who knows SQL runs from ERP! Therefore, if the value of data analysis is only linked to internal control, then the importance and professionalism of data analysis will be very low. The bosses of all departments will analyze it themselves. Do you know anything about sql? What do you care if you don't understand business?

At this time, further packaging is needed to reflect the value of data analysis. The core is the final product! Just like the concubines in the harem, they will please the emperor for a while when they are young and beautiful, but in the long run, they still have to have a child. Having a child will ensure your position. For example, sales can use paper bills, why use pos system? That is, the pos system is online and the business process is running, so he has no reason to stop. The child has been born and still has to be raised.

There are several kinds of children who are familiar with data analysis: management-oriented dashboards and boss-oriented data products suitable for scientific management theory. It may be a recommendation system, an accurate marketing model, a business assistant or a data mart. In short, it is a link that must be used in the daily work of business departments. Packaging, packaging with data, packaging into product-oriented marketing reminder tools and operational data guides. Let the sales staff have a look every day, it will be uncomfortable not to look. Let the operators have to look at the heat ranking before writing the copy, not at the bottom of their hearts. I won't go into details. How to attract the attention of the boss, how to win over the business department, and how to let the front line use it, it is enough to write a book. After consulting for so many years and contacting a large number of Party A and Party B, all intelligent data people finally embarked on the road of doing internal control → attracting management's attention → launching products → cooperating with business departments → expanding organizational structure. And those who claim that the big data system can make a profit of XXX in the end are basically bad.

In the past two years, the concepts of big data and artificial intelligence have been on fire. The position of data analysis has been favored by the bosses of major enterprises like young and beautiful concubines, and countless students have poured into this field. Therefore, I sincerely remind you that we can have many methods and complicated concepts ourselves, but whether the enterprise finally makes money from us is the capital for our long-term settlement. If you only play an auxiliary role, export a product around a specific business scene as soon as possible and combine it closely with the business, so that your position will be stable. Finally, for example, we should pay attention to the distinction between algorithms, because algorithms can be applied to both production systems (such as photo recognition, material distribution, route planning and process control) and analysis systems (such as recommendation, prediction and BI). If applied to the production system, their position is relatively stable, because the production line will not be completely replaced, but will be continuously optimized. However, if it is applied to the analysis system, there will be too much water. We should carefully look at what this algorithm is for before making a decision. As early as 20 13 "big data era" became popular, there was a wave of "big data analysis". As a result, I shouted to my boss at that time: "We can use big data XXXX to analyze and improve our performance." Now it is estimated that the grave grass is as tall as my baby ... As a senior, I have an obligation to tell you the truth of this industry. The value of data can be varied, which does not necessarily directly increase income. The data is really useful, but it doesn't mean that the bosses recognize this use, nor does it mean that we can get a promotion and a raise from here. In addition to technology, how to create value may require the assistance of something other than code and algorithm. * * * with everyone.