In the class of drip product evolution camp in March, Mr. Wu minghui, founder and chairman of milliner data and second hand system, explained to the students in depth what he had learned about entrepreneurship over the years:1. Why not blindly believe in big data?2. How did two To B enterprises founded by wu minghui -- millinglho and second hand grow into unicorns?3. Big data industry, where are the opportunities in the future?Today's notes amount to 3,454 words, and are shown to you throughout:The value of data comes from the decision value of subsequent behaviors after data is obtainedIn the context of the era of big data, many people with a keen sense of business have started to think about a question: can we do business with data? The answer is unfortunate. Conceptually, data and business are very different things. Data is the solution to information asymmetry, the nature of business is to use information asymmetry. How to make a business by connecting the two sides of the conflict?That said, there are plenty of people who make money with data. For example, if you think of data as a low-level thing, you can actually say that most people make money out of it. Just make money means is different, the asymmetric information that business USES can be divided into two kinds: one-time information asymmetry and continuously changing information asymmetry. Generally speaking, it is possible to make money from data only if it is continuously flowing. A one-time information asymmetry is at best a dull fortune.How to turn data into a product, how to express its value, is crucial. Data is the friend of time, because time is the most valuable, so the value of data products lies in helping others to save time and improve efficiency. We can further understand the value of data from two directions:The first direction is to create trust with data, which is also the motto of second hand company. Reducing decision-making cost is a very important value of data. For example, when advertisers are hesitant to put in advertisements, the second hand with sufficient data analysis, can promote the decision to happen. Data creation trust, in fact, the data itself is not the most critical right, trust is its core. Trust reduces the core cost of the whole society, improves social efficiency and endows data with value.Psychologically, why do data generate trust? Because the decision-making process of human is that your sensory system translates external signals into symbols. The second step is to process the symbols by logical reasoning. The decision-making process generates the final decision and puts it into action. That's the value of the data.The second direction is to reduce the cost of trial and error, which is a very important value of data. Why can data be falsified instead of verified? Anything that you think is true is true by strict logical reasoning, logical reasoning is about syllogism, the first paragraph is about premises, assumptions, why are conditions true? If you want to prove that condition to be true, it has the premise that if you push it down one layer at a time, to the bottom layer that is the part of the axiom, is the axiom necessarily true? Non-euclidean geometry says that two parallel lines can intersect. The data is too complicated, but there's nothing you can do about it?The world is complex, but there is a phenomenon that the world is mostly continuous. It's hard to predict the long term in the face of complexity, but it's accurate in the short term. For example, in the case of stock trading, programmed trading, high-frequency trading is actually short-term predictability, which is one of the things that complex systems talk about, a lot of things change continuously, you can make money by trading very quickly.Data analysis can help you find best practices by comparing time and space, which may save you the trial and error costs of making decisions in your business. You're going to try it out, but you're going to find some directions that you don't have to try, and that's the value of the data, that's the value of the information, and you can cash it in.The data itself isn't valuable, it's valuable because of the decisions you make after you look at the data. The value of data, or of a data industry, depends entirely on the scenarios in which it is applied. For example, in the digital advertising business, the second hand helps customers achieve accurate marketing and magnifies their profits by providing them with advertising monitoring data centering on the target audience. This convenient and efficient service experience naturally makes the second hand produce great product value.Many companies take a business plan and say that their model is great and that the big data generated by the free model is valuable. If you think you can make money with data, it's definitely a hypothesis.Data products: monopoly scenarios are more importantMany product managers know an important formula: product value equals new experience minus old experience minus replacement cost. Going back and focusing specifically on the data product, what should be its value formula? Decision makers should know the change ahead of time after using your data product. The value generated by the data is to let you know the change ahead of time.Data product value = the decision maker understands in advance the cost savings and added value of change - alternative costsThis is the new experience of data products minus the old experience, minus the replacement cost of course. The replacement cost may sometimes be negative. Sometimes the cost of data is reduced in the data industry. A new method may cost less than the original method, while the replacement cost increases. Our methodology for continuously optimizing the value of data products is very simple. We should either increase the number in front, or decrease the number in the back, preferably negative number, so this is the product formula of data products.The goal of data products is to accelerate the realization and amplification of decision-makers' benefits in information asymmetry. Acceleration is also important. The sooner you learn this information, the better.Cybernetics is a very important theory in modern science. In the process of making a decision or a system moving forward, there are actually stages of perception, understanding and decision, and then action. In cybernetics, there is a very important model called perceptual response model, which has three corresponding elements when applied to data commercialization, namely, data source, people and application scene.• data source: in the perceptual response model, the perceived result is the data received by the sensor., people, policy makers requirements: many Internet companies never buy second hand products before, but they are all bought you today, because they have become big business, drops, Meituan, headlines are bought a second hand service, their more and more money, please, more professional, once the professional will use professional tools for analysis, before it is small when no one can do analysis, so it's very important to people.• scenarios - data usage scenarios: mainly information asymmetry scenarios. For example, the police intelligence research and analysis of the relationship between mining and analysis, this is certainly not as good as the machine.If these three things are not done well, then the data cannot be commercialized.Therefore, the opportunity of data business must be a great change of one or two of these three elements, and then there will be new innovation opportunities. Why is that? All business is continuous, others do well in this industry originally, why do you suddenly kill out today? Why did you come and turn them upside down? There has to be a big change in one element, two elements, three elements of the industry, and that's when you have the opportunity to innovate, to have the opportunity for us to do unicorns.The value of a data company is determined by the value of the industry it serves. How big is the industry you're serving, and multiplying that by a percentage should give you the value of your company, and that's data commercialization.The selection of scenarios is the core of data commercialization, and there should be a core decision problem in each scenario, which requires some data for decision-making. As long as you choose the right one and the decision is valuable, you can go to commercialization.So how do you make money with data?You have to have a monopoly in any market segment, you don't want to have a monopoly you don't want to have an opportunity to make a profit, why does hardware have an opportunity to make a profit? Because its marginal effect is not zero, its hardware itself has a material cost, and this cost is not transparent in many cases, so you have the opportunity to make money in it.So, if you want a monopoly, what about a data source monopoly? Or do you have a monopoly on data? Or is it monopoly? There is no profit without a monopoly. What is a monopoly? I want to tell you my own view, I've been doing data products for so many years, and I think monopoly scenarios are more important and easier, and it's almost impossible to monopolize data, and even less possible to monopolize people.Scenarios can be monopolized because all scenarios have fixed budgets. Let me give you an example. This money cannot be copied. If it is given to service provider A, it cannot be given to service provider b. it is impossible for everyone to buy it. If you buy more good data, the scene can be monopolized, because money can be monopolized, because the amount of money that can be used to buy data in the hands of the person who USES the data is limited, and this money cannot be copied, so if you give A, you cannot give B.A data product, an information product, not monopoly has no chance to win. Because this brand trusts and chooses you, after the brand is established, data products have the opportunity to monopolize a scene. This is a feature of data products.What's the other opportunity? One is a major historical change in many scenarios, and the other is a sudden change in the history of your data costs, which can also create opportunities for data businesses. Data sources are often subject to change, and many new data sources will be born.We're actually generating data faster, more accurate data, more and more data, more and more good data, faster iteration, more and more valuable data.A lot of companies make good products but they die and customers don't use themThe previous part is about my own experience of using data to do business, and finally, a little interpretation of my two companies' business. To sum up today, there's a reason the second hand makes money. For example, in the development process of more than ten years, second hand also tried to do small B business, but always failed in the end. I just think the scene itself is too small. As I said earlier, the value of a data product varies with the size and value of the scenarios it faces.Software products, information products, data products, once full competition, because your marginal cost is almost zero, so eventually will be locked in a price war, you end up with no money to earn. And for that reason we can't do it, we have to do it in a different direction.In other directions, we found a big problem: how to help our customers use this data is more valuable? For example, I chose the public security industry, not only to help him make a platform to save these Numbers, but also to help them really use these Numbers to solve crimes, so later we made this system to help him solve crimes.Therefore, it should be said that the commercialization process of data products is very painful, because it is useless for you to have data, and it is useless for you to have scenarios. Many, many companies are dying in the end. When you make a good product, no one will use it, because you will use it, but it does not mean that your customers will use it. Here we also take the public security system as an example. We just started to provide the public security research and judgment system. Only very professional police can use it. This is a product acceptability problem. Limited use of the product or 'users will not use it' will directly affect the popularity of the product. Only by adjusting products for human reasons and gradually forming a closed loop through constant iterations and attempts, can the continuous growth of company value be realized.Similarly, millmillers' enterprise-class Siri xiaoming is also designed to solve human-computer interaction problems by lowering the threshold of using artificial intelligence products in the form of natural language q&a. For example, a rookie policeman said to xiaoming, 'xiaoming, help me analyze the case clues.' Xiaoming feedback on the key words of the case, and analysis. This will make it easy for police without computer background, risk control managers and subway owners to use, so that customers will be willing to buy and use your products.