What is COST, or Untapped niches in the online travel market

Last year, McKinsey shared its vision of the future, more precisely with new types of companies that will grow rapidly on the world stage in the next 20 years.



These are a variety of businesses: from integrated solutions of the smart home, which will combine security, utility and media services (in the terminology of the study - HOUSE, HOme Unified SErvices) to new methods of high-quality water treatment for drinking and agriculture (WADE, WAter DElight) . What unites these new types of companies (and there are twelve , according to McKinsey) is that their creation becomes possible only now, when the necessary infrastructure, technology and demand have come together at one point.



Do not forget in this list of companies and the evolution of the travel industry. Significant role in it should begin to play services, which in McKinsey called COST. The abbreviation stands for Convenient Organizer Service for Travel, that is, a convenient service for organizing travel. It is assumed that such a service will offer the user a journey as a finished finished product, taking on an effective assessment and selection of all its components: flights / transfers, accommodation, plans for visiting sights, etc.

')

OTA, meta, COST

To better understand how COST fits into the current landscape of the travel market and how this service differs from the existing ones, let's remember how the industry developed. The market is based on the ultimate service providers: airlines (for example, Aeroflot and Lufthansa) and hotels (for example, Marriott and Hilton).



For a long time before the advent of the Internet and a little after the only "superstructure" over these companies were tour operators. The person had a choice: spend a lot of time searching for a profitable flight and accommodation, studying the offers of dozens of airlines and hotels, or delegating the organization to the tour operator and incurring additional costs. At the same time, the choice was rather conditional: tour operators work with certain areas and with a limited set of hotels and airlines.



The process of finding and buying tickets on a pre-planned route has become easier with the advent of online travel agencies (OTA). These include, for example, Expedia, Kayak, Anywayanyday or OneTwoTrip. Their task is to provide an opportunity to choose between offers of the same service from different suppliers.



The next stage of the travel market development was metasearch systems, such as Mamondo, Skyscanner or Aviasales. They aggregate the offers of different OTAs in order to select a cheap / convenient option from a larger number of offers.

Each year, according to eMarketer, the share of OTA and metasearch systems on the market grows by 2-3%, that is, customers gradually move away from direct service providers to these companies.

Travel as a whole product

Thus, in recent years, the industry has come a long way to make life more comfortable for travelers, simplifying the process of finding profitable ticket offers and hotels for the desired date. However, another key and time-consuming stage, travel planning, remained virtually unchanged.



Often we do not know when and for how many days it is better to go on vacation in terms of the optimal total cost of flights, accommodation and rest. If, for example, you can fly away on any September day, then manually sorting through all the options is almost unreal. The task is many times more complicated if you want to visit several cities or countries at a time.



In addition, the journey - it is not only the road and housing. The completed plan should include car rentals and personal recommendations (or orders) for places of interest, such as attractions or cultural events and ready-made offers for restaurants and cafes.



Thus, if the task of buying a separate type of service (hotel, flight, etc.) is already quite qualitatively solved, then the first decisions in the field of travel planning are wholly rare and fragmentary. And to COST can be attributed to companies that will deal with this in the coming years.



The solution to this problem is worth it. The global online travel market today is estimated at $ 408 billion. Let us assume that “travel as a product” is needed first and foremost for advanced travelers, for whom the problem of travel planning is particularly acute. Surely, they will be the first users of the new services, the so-called early adopters. Complicated travel today accounts for 5% of the market, that is, about $ 20 billion a year.

If the service earns commissions from all service providers that are in the proposed plans (flights, hotels, car rentals, excursions, etc.), even with a commission of 2% of the pie for COST, today it is $ 400 million.

Forecast instead of requests

Actually, it is for this market that we want to fight. Our IQPlanner service must overcome a number of obstacles, some of which rest against existing ticket reservation services.



This is, first of all, a relatively long processing of a request to the global reservation systems (GDS) and the difficulty of evaluating the results obtained.



For example, we have a plan to go by plane from Moscow to Paris and back, stay there at one 4-star hotel for 5-7 days depending on the budget, which should not exceed 60 thousand rubles. And, of course, 7 days would be ideal. We are ready to go on a journey any day during September, that is, thirty days.



To make such a plan, you need to make 90 requests for hotel search: 3 requests for each day of September, but with different length of stay: 5, 6 and 7 days. Add to this 90 flight ticket search requests. We receive 180 requests, which even in the case of processing by a system that consistently makes on request to the GDS, would be processed for more than 15 minutes. Then the obtained data still needs to be analyzed to find the optimal combination. To do it manually is an extremely difficult task.



Our team of mathematicians and developers at IQPlanner has managed to make significant progress in this direction. For the last two years we have been working on creating our own price forecasting function, which would allow us to reliably predict the cost of a ticket over long periods of time. The function is based on machine learning algorithms; accordingly, the accuracy of its results will increase as the number of users grows. Using this function, we can reduce the processing time of the request, making an assessment and selection as quickly as possible (from a few seconds to a couple of minutes) for the user.

Opportunities

Such “overcoming the temporary barrier” opens up qualitatively new conditions for automated travel planning.



The most obvious example is the possibility to determine the ideal day to start the journey and / or duration (for example, from 10 to 14 days), specifying intervals of up to a month or more as possible departure dates. IQPlanner will forecast the change in prices for tickets and hotels during this time period and will determine the most profitable day in terms of total costs.



To achieve this result manually, I would have to write prices for the cheapest air tickets for all possible combinations of round-trip dates. Then it would be necessary in the same way to go through the proposals of hotels - find and record the minimum cost of accommodation in hotels in the city for each arrival date and possible stay. And, finally, the third step is to sum up the flight and hotel options for each day / stay option, to finally find out when and how much it is more profitable to fly.



And this is the simplest application of IQPlanner. There are tasks and more difficult, which is almost impossible to handle on their own. For example, with the service you can determine in what order it is more convenient and more profitable to visit the planned cities, how long to stay in each of them, taking into account the changing prices for tickets and hotels, and where it is better to use which transport.











The result of IQPlanner is a ready-made travel plan, as well as prepared booking requests to our partners. We are already working with Rome2Rio, Anywayanyday and Ostrovok.ru and are planning to connect a number of new partners who will cover other travel segments, such as car rental and excursion booking. By the way, if you work in such a company, or just have a great idea, we are open to suggestions and discussions.



This week began beta testing our service. We will be grateful if you try to make a plan of your dream trip in IQPlanner and share with us your ideas, comments, or suddenly not caught bugs here in the comments or on our forum .



In addition to our huge gratitude, all participants in beta testing can count on a special status - we are already preparing various bonuses.



Become a beta tester IQPlanner »