Mining business opportunities with large data three iron rules to follow

In Japan, in addition to big companies such as human resources giants recruit and Kao, H.I.S (a large comprehensive travel agent), DeNA (Mobile internet) and Mitsubishi Heavy Industries, including some local backbone enterprises and ordinary small and medium enterprises are also unwilling to do, have begun to use large data to create business opportunities. To be sure, in the era of large data competition, many companies will compete to use large data to enter new business areas, creating a business model that rivals cannot emulate.

However, it is impossible to find the treasure if you simply collect a large amount of data and analyze it aimlessly. Must have clear goals and actions to be able to discover their commercial value from large data.

Mining business opportunities with large data

Has nearly 300 years of history of Japan's Kyoto Yu-Tea shop-Kito right-back door, due to the production of excellent Japanese tea and famous. In the January 2013, the famous old shop, which has been making tea for nearly 180 years, began to manufacture and sell wine (Fig. 1), which mixes tea with Japanese wine and plum Wine to develop a new wine market. Although this is a no one has entered the field, but the business planning Department of the Minister of rang---very confident.

The maxleaf of the Kito Right-back Gate is the manufacture and sale of tea and tea-making materials. No matter how to analyze the sales data and customer data, there is no definite conclusion--should explore the new wine market. So why is the company able to have the answers?

Large data drives enterprises to develop new markets

The company began to participate in the wine business from the local winery and the development of a combination of tea and Japanese wine mixed with the new products-"night of green wine." May 2012 trial, half a year to sell more than 5000 bottles, become a very popular commodity.

Kito right-back Door Business Planning Department minister rang and "night of green Wine"

But even so, it is not possible to "formally invest in the cause (rang)". In order to establish the new field of "tea-drinking", it is necessary to expand the commodity line. At the same time, in order to improve the awareness of the need to increase investment in advertising and publicity. No matter how popular, if it is a temporary fashion products, it can not be officially put into production.

In the end, the big data eliminates this unease and makes the company determined to challenge new areas. The company uses WINGARC's commodity intelligence software--dr.sum EA to analyze the sales Web site's access logs and sales performance, POS sales terminals and membership attributes, with the aim of detecting the sales tendency of each customer attribute (whether it is a new customer or a regular customer). And the cross selling rate between the goods and the effect of the promotional activities.

The analysis turned out to be unexpected, "the original forecast is the vast majority of new customers ((Guang-rang), but in fact about 80% of customers have bought the company's products, and the high number of people." The company estimates that only by sending a letter to the old customers can also get the corresponding sales, so the final decision, decided to formally put into production.

Kito Right-back door for the analysis of the data is about 600,000 member data and 2 million sales data, as well as the annual 30 million access logs. Of course, the size of the data and the general so-called "big Data" is far from the difference, but as a local backbone enterprises can collect so much data and analysis is rare. And more than that, the company has challenged "opening up new markets", even for large enterprises.

At present, such as Kito right-back door Such use of large data enterprises, not the size of their enterprises and regions, has been emerging. These enterprises have been robbed of other enterprises in the front, occupy a competitive advantage to seize business opportunities. In fact, internet companies, which lead the market in the forefront of big data, have started to create new business models for the big Data age.

Network Enterprises Open Physical stores

The human resource giant recruit, which provides network information services, also operates the medieval vehicle information website, "Car Sensors (carsensor). Net", the company recently entered the sale of the real business of medieval cars, the name of the service is "cars and counters."

January 2012, the company in the Sendai Shopping center opened a physical shop, where the medieval car professionals in accordance with the needs of customers in the network to choose the car, but also for the cooperation of the middle car vendors signed contracts and other business. The operation of the shop is handled by the group's Bei Guan Dong Marketing department.

Although the company operates the medieval car information site, but in the sale of medieval cars belong to the successor. If only the sales agent, compared with the traditional services, because of the inability to provide differentiated services, so there is no advantage. As a result, the company decided to make full use of large data to create a new business model that other companies could not follow-the same quality of medieval cars (the type of car, the year of production and the type and distance of the car and the maintenance records, etc.) and sold at the same price.

In fact, there was no clear standard for the price of medieval cars in the past. People who buy in medieval car stores, though, refer to the price of the auction market, but ultimately rely on "feeling and experience" to determine the price of the sale. Therefore, even the same quality of medieval cars, the prices of different shops are not the same, the prices vary greatly in different regions. and recruit effective use of large data, collect millions of car price information, analysis of factors determining price. Final realization-the same quality, the same price (the same quality of car sold at the same price).

Recruit Group IT Solutions Department of large data in the head of the Chrysanthemum original extension in talking about the purpose of using large data, said: "Through data analysis to customers show the rationality of medieval car pricing, eliminating the customer's distrust of price." ”

Using Hadoop to realize service difference dissimilation

Supporting the new business model of the large data is the company's medieval car information service "car sensors" published data. Based on more than 30 pricing factors (such as vehicle type, automobile production year and model and travel distance, exhaust quantity, etc.) and 3.4 million market price information per month, a unified pricing inference system is established.

In fact, the company had this idea long ago, but the need to consolidate the basic data is too large, and in order to derive price calculation reasoning, must be the annual car production and model and travel distance, such as the characteristics of the middle vehicle and the market price and other data control, to be able to determine the price factor Even if it takes a few days to do cluster processing, it takes years to establish a business model and therefore has to give up the idea.

The change in this situation is distributed processing software "Hadoop". Using 5 servers to build an experiment based on Hadoop and verify the results of cluster processing, data that took days to process was processed in 1.5 hours. In other words, large data technology solves the business model problem.

Spend 3 months of time, and repeated cluster processing and verification, the successful construction of the best price calculation inference program, the cause of the finally have a prospect. At present, due to hardware performance enhancements, the total is only about 30 minutes. and the monthly price calculation of the reasoning to upgrade, improve the "same quality, the same price" accuracy.

Opening up the way to expand overseas markets

At present, the big data also becomes the Japanese enterprise expands the overseas market the pushing hand. Mitsubishi Heavy Industry, for example, has been involved in urban transport since May 2012 in United Arab Emirates Abu Dhabi Dall.

Mitsubishi Heavy Industry uses the traffic simulator with large data, proposes the plan of popularizing electric vehicle (EV) and electric bus in Stall City, and calculates the economic ripple effect.

Popularization of electric vehicles and electric buses requires investment in setting up charging stations. The cost of promoting the electric drive of passenger cars and buses is different. Mitsubishi uses traffic simulators, by changing the type and number of electric and electric buses, to speculate on the necessary power consumption (walking and air-conditioning), while calculating the cost of popularizing electric cars and electric buses.

To produce the simulator, Mitsubishi Heavy Industries analyzed various data such as road slopes and EV acceleration, and the deterioration of the battery. And, once used 10 ev "i-miev" "in Stall City for actual test, collect walking data and do analysis. "The ability to predict results correctly will be a powerful weapon in the future of our company's participation in building smart city plans around the world," said the head of Mitsubishi Heavy industry with confidence. ”

Three principles of iron to challenge large data

The common characteristic of the three companies mentioned above is that they have grasped the business opportunities of the new business through the analysis of large data. Everyone said that "big data is a Baoshan", but whether there is really hidden treasure is not at a glance.

So how should companies use the weapons of large data to challenge new ventures? The author thinks that enterprises that use large data need to abide by the principle of three iron, namely "on-the-spot feeling is the key of analysis"; In addition to the company's internal data, but also to target external data ", effective use of tools to make" analysis easier and faster.

According to the Nomura Research report, about 60% of companies with sales of more than 1 trillion yen are currently studying ways to make use of big data. Less than 50 billion yen enterprises, four companies have 1 in the collection of information. So what are the new initiatives for our competitors? This article will continue to introduce a number of advanced examples for reference.