Micro Car CEO Lei: Big Data 70% is a pseudo proposition

The four-wheeled car box has become a space for people to act second. In this space, people are relatively still, but have been moving. Can we find useful and effective data in this static and moving process?

On the forum of the "Smart Car Salon" held by shift last Saturday, the CEO of Lei made a keynote speech. In the automotive internet age, he argues, software-level entrepreneurship may have more opportunities than hardware.

Sharon, Lei first introduced the micro-car is what.

In one word, a micro-car is an app that extends a range of services based on distance from the car. Its biggest characteristic is that the user does not need to open actively, the program will run automatically in the background. function, the micro-car has now achieved violations, limit the line, the push-back function, covering all the streets of Beijing stickers map data, road conditions, such as the forecast function.

He says nearly 10 million users are now using micro-car apps that produce billions of dollars a day. Micro-car on the one hand can use the user's mobile phone itself to identify the various sensors are currently in the car, where the current driving, the current use of what kind of driving behavior in the car, on the other hand can be in front of the user every day, by identifying the user's usual route, combined with the current Predict how long it will take to get to the destination when you go out this road today.

In addition to displaying some of the features in front of the user, Lei said that if the anatomy of the micro-car to do things, can be summed up to three keywords: vertical, just need, data-micro-car based on the automotive vertical field, based on data to meet the rigid needs of users. The key point is that the micro-vehicle's product threshold is the data, the micro-vehicle positioning itself in the data based service engine. For the user is a vehicle owner Service portal.

Next, Lei shares his views on big data, and what big Data the micro-car cares about, and what big numbers make sense.

70% Large data mining is a pseudo proposition

In Lei's view, 70% of the large data are pseudo propositions. For example, Baidu is not a big data company, in fact, not. It is just an information company that brings together reptiles from around the world and then gets the data from these reptiles to the platform for users to search. "In China, there are two types of industries or companies that have big data," lei.

One is the operator industry. Operators can know where you are every day, and who is on the phone. But there are two restrictions on operators, one is not to do too much data mining, do not dare to use this huge data; second, the user's ability to arrive is particularly weak, now effective way to arrive is a text message, but either by the user intercepted or ignored, conversion rate is particularly low, can not commercialize the data.

Second, the electric business enterprise. Electrical business is a real big data, it can know when you shopping, what you buy, sent to where, gave who, these are data. This is also Ali, Jingdong, Amazon and other electric platform of the core value.

So what's the data? Lei that in general, the big data industry should have three features:

Scale: It is not possible to abstract data models to make data analogies without large enough scale.

Mapping: That is, how to relate the data to people, and on this basis, through statistics, deduction, induction and other means, can be the data conformation.

Arrival: The data reaches the user to play its value, if can not reach the user, even if the huge data is meaningless.

How to make big data about cars

Since 70% of the big data are pseudo propositions, how can the automotive sector make real big data? Lei that the automotive sector, the data can be roughly divided into three categories: cars, roads, people. In-car data is now OBD products are concerned about the fuel consumption of the data, it can remind owners to pay attention to vehicle status, can also be used as the basis for maintenance; The road data include traffic, road condition information, municipal construction and so on; People's data is the daily behavior data of people.

Since it is large data, the first is to have a large enough user base to generate data. "If an app has 2 million active users. Take this number for more than 100,000 4S stores nationwide, if each user to two times a year 4S shop, then 2 million users can give each 4S shop in the country on average only 40 users recommended. If this data is negotiated with the 4S store, it doesn't make sense because every 4S store has more than 40 customers a day, which means that the amount of data is small, and this business model doesn't. "Lei said.

"For the auto industry, if you want to do O2O, affect the business of the 4S shop and get revenue from it, you have to affect at least 10% of the size of the business." Take group buying, its influence to the new open shop can reach 20–30%, this explains its business model is feasible. That's why you need a large enough group of users. ”

The second premise is that enough data is generated. We all know that mobile internet still relies heavily on advertising revenue, but the mobile end has a characteristic is that although the user's fragmentation time is more and more, stays in the app the time is getting longer, but compares with the traditional PC connected, its display area is too small, the information quantity is too few, therefore the advertisement coverage certainly is very few. Therefore, it is necessary to have enough data for users to concentrate enough valuable content, in order to timely push advertising, improve conversion rate.

What a micro-car wants to do is big data, and it's inevitable to talk about what the business model is. Lei said that the main focus of the micro-vehicle is human data, when the master of large enough user groups and user data, and then based on this to help users on the micro-vehicle platform for consumer decision-making, this is the micro-car business model.

The boundary of the vehicle data is the user's life boundary.

Cars produce a lot of data every moment of the day, but what data makes sense. Lei said that the micro car is only concerned about one kind of data--and the user's life radius is closely related to data. For example, the micro-car can recommend insurance companies for users, but in fact, it is not a cheap recommended which one, but which compensation easy to recommend which, of course, the premise is that the price can not be too much difference. This recommendation is based on the user's life radius data, such as how many kilometers of one months to open the highway, the user's life radius how many insurance company maintenance points, claims points, and easy to obtain claims, and then can recommend which insurance company for users most suitable.

However, these data mining, if only to do a shallow level of recommendation, it is far from the value of large data to play. Lei describes how to use data in a deeper way by using two examples of automotive industry--auto insurance and post-market.

First of all, the current factors that can affect the price of insurance, in addition to models, mainly the owner last year, the Chuxian data. Some of the current driving habits such as data and insurance are rarely linked. At present, micro-vehicles in the context of data collection, but also in the thinking of how to combine with insurance companies, such as illegal data may be another factor in insurance. This also means that when you make big data, you need to consider whether your data is relevant to the insurance company.

After the market, the online data will still be linked to the 4S shop below the line. Scattered across the country's 4S shop is located in the radius of radiation, such as the average radiation radius of 4S stores 10 km. So this will be difficult, on the one hand micro-car to use the Internet perspective to expand users, on the one hand, will be considered to get the data and offline combination. For example, there have been on-site maintenance, on-site maintenance, such as new service mode, but can not be large-scale still doubt. Because these modes of service are mainly user-centered, the model that is the core of the traditional 4S store is still contradictory. Now the number of micro-car users and enough data, how to import these users and data into a reasonable platform is still in the groping stage.