The "money" conjecture of airline large data

At present, in the domestic aviation market, the vast majority of services are free, airlines are gradually seeking services to charge the method. However, at this stage of the pay service is not "hit", only to find the most wanted services, is the demand.

To face the data and use the data, the core of the large data commercialization is to develop different marketing and service strategies for different passenger groups. Large data can help airlines to achieve business model transformation.

"Big Data" is not "big database"

Currently, the domestic airline's data "warehouse" is only a "big database", not "big data". Many data items, storage and management are messy, can not achieve the data sharing between different departments, it is difficult to make large data collation analysis of the conclusions of a wide range of applications at the company level. For example, the number of airline tickets sold per day and the amount of fares is the basic data provided to the financial settlement system, few airlines think of using this data to analyze their own forward flights on the passenger seat trend, as well as the changes in the number of agents sales, so as to guide their sales staff to conduct marketing activities.

In fact, airlines are the easiest to achieve large data applications of the enterprise. Because of its daily production operation for the business process and time to match the precision requirements are very high, airlines are often introduced and developed a series of computer systems to intervene in the production operation of various links, thus providing airlines with a large number of data records.

In addition, because the data "warehouse" contains too much data, so before using, it engineers often do some data cleaning work, through the correlation between the data to modify, and ultimately provide users with the collation and modification of the data after the refinement.

For example, when airlines are counting passenger data for a temporary upgrade at the airport, they will find that some passengers have to refund the tickets before buying. This is actually a first-line ticket staff in order to save time, directly handle the passenger economy class ticket refund, and then sell passengers a first-class ticket. As a result, the operation of a standard passenger lift becomes the operation of ticket refund and repurchase. Generally after data cleansing, in accordance with the principle of departure data, this part of the passengers will be removed from the upgrade data, and finally these passengers have become the airport temporary ticket buyers. This data cleansing modifies the market reality: The passenger's will is to ascend rather than to buy the ticket temporarily, in the statistics temporary purchase the number of passengers will produce errors, ignoring the demand for passenger lift, and exaggerated the demand for temporary tickets to the airport.

Let the data fully show the market reality

The application of large data, we should not ask "why", but let the data tell us "what". The nature of large data should be a collection of integrity, confounding, and correlation.

Integrity, which means that large data is used for all data, not random sample data. The confounding of large data, when collecting and analyzing large data, does not require one-sided pursuit of the accuracy of the data, but can put the mixed data together. For example, some airlines first start to count the number of passengers passing through the hub (O&D), using the industry's standard statistical rules: The journey of multiple flights are in the same booking record, in the transit airport stay no more than 4 hours belongs to the relay. After such statistical classification, domestic airlines in their own base city carrier transit passengers accounted for the proportion of all outbound passengers will not exceed 5%. Later, as the number of data records more and more, the recording project is more and more detailed, the airline began to use the passenger identification number to match, no longer require passengers to travel in the same record, and the transfer to the time to increase to 24 hours. In this way, the number of transit passengers is 5 times times higher than before, and the analysis of the demand behavior of these passengers is more accurate.

Moreover, when the airlines collect more and more relevant data, for example, from the railway department to obtain passenger information on high-speed rail, from travel agencies to get the information of the bus passengers, and then put together these mixed data, instead, can get a more complete and comprehensive transit passenger database, Especially for the airline hub construction and transit product design is helpful.

The relevance of large data, you can use the "stone of his mountain, can attack jade" method to deduce the conclusion. Taking the hub of the airline as an example, an airline wants to attract passengers by giving bus tickets to hubs around the town, but it is impossible to estimate how many passengers will enjoy such free tickets, and it is difficult to count the cost of the product and the number of passengers it will attract. Later, in a promotional campaign, they learned from the local trade and industry bureau that the small towns that wanted to promote promotions had a chamber of commerce in the Hub city, which covered 80% of the town's entrepreneurs, and that the Chamber of Commerce had begun to regularly count the economic and trade activities and travel needs of its member companies. So, from then on, the offices of the town chambers of commerce in the hub cities became the customer groups regularly visited by the airline market staff. Using these Chamber members ' travel needs, airlines can deduce the number of "Ticket + ticket" products.

Use data to find travelers who are willing to pay

The core of large data is the huge amount of data, all the data can be collected should be statistical analysis. For example, airlines in the terminal will only count the number of passengers entering the VIP lounge, and never consider the number of passengers entering the smoking room, the number of children in the children's play area, and the number of passengers taking ferry batteries. In fact, these data for the airline marketing are meaningful: Smoking more passengers, indicating that passengers for more than 2 hours of the voyage endurance than poor, in the flight process is easy to irritability; more children, the cabin environment will be noisy, the flight attendant service will improve the difficulty; more passengers, Reflecting the distance from the gate to the checkpoint, or the late arrival of the passenger at the airport, this can be improved by optimizing the downtime between different flights.

The reuse of data can be used to realize its potential value. For example, airlines flight passenger data, in the event of flight delays, the data are passed to the call center, ground services, crew, catering and other departments, to provide passengers with delay notification, follow-up change, service changes and a series of security work. After that, as the traveler finally makes the trip, the data will only remain on the summary of the statistics on the business statement, the specific passenger information data are no longer used.

However, by analyzing the passenger data that have already been affected by the flight delay, the company that designed the flight delay insurance can calculate the amount of the insurance policy is more suitable, the airline can calculate how much cost to pay for the passengers to buy insurance more valuable. and specifically targeted at the most frequent delays in the number of passengers to provide some compensation services, can retain passengers, improve their brand loyalty to the airline, and even airlines can be published by third-party companies to obtain business value of some data.

Most of the foreign airlines put resources in customer relationship management, and strive to achieve high-end passenger value. However, for domestic airlines, improving passenger satisfaction should be the primary direction. This is because many foreign passenger service is already a charging project, looking for high value passengers willing to pay high fares, is the key point of the airline competition. And in the domestic, most of the services are free, airlines are gradually seeking services to charge the way to find the most wanted services.

For domestic airlines large data applications, the first should be to expand passenger services to improve passenger satisfaction. As a result, airlines are measuring and measuring whether they need to put in more service resources to improve passenger satisfaction.

From the 90 's projection television, to the current LCD screen, handheld tablet computer, etc., the entertainment equipment on the machine has undergone many upgrades, but the complaints of its passengers is still high. This is because the composition and preferences of travellers have also changed, and young people have shifted from their obsession with television to the Internet. The analysis data obtained from Apple's iOS platform and Android platform will be more helpful to airline companies.

At the same time, airlines can use large data analysis to achieve precision marketing. The high frequency passenger groups with the highest number of flights each year, their ticketing behavior is very distinct, like to choose a fixed channel to buy tickets, time to buy tickets is very close to travel time, generally do not buy tickets in advance, and rarely advance processing value machine procedures, the majority of these passengers often choose to take off the last 30 minutes before the check-in procedures. These passengers do not care about the price of tickets, more concerned about saving time, to provide them with "easy boarding" products more to meet their needs.

However, for airlines, data relating to the privacy of travellers should be a restricted area in use, and data conclusions should be prohibited from pinpointing the information of a single passenger.