The Myth of Automated Front-End Booking Systems

Article Published in Travel Weekly, Observer Column- October 9, 1997

Labor generally represents close to half of an agency's costs. To reduce these costs, many agencies are offering electronic booking solutions to customers. Software is made available through a corporate network, on Internet or Intranet, or loaded on each user's computer for direct dial access. The traveler or travel buyer uses the software to create an itinerary directly in a CRS. The effect of this process is to pass a portion of the agency labor cost to the buyer. Such transactions do little to actually reduce the cost of the booking transaction - and, in fact, may increase it!

The marketing myth sold to corporate customers is that self-booking provides a base for cost reduction ... in which the customer can participate. Undoubtedly, there are cases where this may be true. But over a sustained business-booking platform and given the present distribution structure, it is truly a myth.

The core of this myth lies in the fact that all of the automated self-booking systems use the existing CRS or airline host systems. These systems, with the exception of one or two new entrant airlines, use 30-year-old special function inventory systems that can only accept command language instructions as queries to the data structures. Interactive booking tools act as "front-end" interpreters between the human user and the archaic command language of the inventory/distribution hosts. Developers of these front-end interpreters have, to date, focused on providing users with a "user friendly" tool. "User friendly" is a necessary ingredient for selling the concept and products.

These developers rarely, if ever, address the design issues of "interface" to the host system. Accordingly, these booking tools are inefficient in their links to the hosts. Few use any sort of "fuzzy logic" to understand a pending booking query from a user. Without such "fuzzy logic, it is difficult to incorporate any of the faster or more efficient command language instructions that are a part of a travel agent's skill set. Because simplicity of design structure, speed of development, and cost controls, are the major "drivers" in development of a link to any of the host systems, these systems typically repeat the same queries with minor changes to reflect carrier or market specific needs.

Even when used by trained agents, the redundant nature of the simple query designs put increased demand on the CRS and airline host systems as a function the CPU time needed to process the queries (often referred to as "hits" on a host). When used by untrained agents (i.e., the self-taught, corporate travelers and/or their secretaries, etc.), the "hits" burden is compounded an estimated 10 or more times. Thus, while booking time may be only twice as long, the "hits" burden is significantly greater. While an agency offering such tools may have reduced its local labor costs, but the airlines have shouldered increased CPU processing costs.

Both airlines and CRSs have been impacted by these costs. Under the current method of CRS billing, a major portion of the CRSs cost is passed to the airlines in the form of increased booking or other service fees. The airlines are reacting. Airlines like Continental and Northwest have limited commission revenue on Internet generated bookings - well below the regular agency commission capped levels announced last year. American, followed by United, now pay a flat fixed fee ($15 and $10 respectively) per Internet booking ... effectively making Internet bookings nett fares, less the fee!

Perhaps more important in the mix, however, is the literal cost of the labor shift.

If the buyer self-books the travel, the labor and time spent to do this work becomes that of the user ... not an agent cost. Depending on location in the United States, agents unburdened salaries are between $20,000 and $30,000 per year. Most corporate executives that travel, earn significantly more. Typically, staff assistants that work for those executives that travel also earn more. Yet these corporate executives (or assistants) have neither the travel experience nor training of an agent, nor the benefit of access to the more efficient expanded CRS command language (which assumes, that these corporate users have the skills to use the command language). Studies indicate that the time to book a travel itinerary using "front-end" solutions is more than double that of an agent - among the most efficient "front-end" users. Thus, a self-book buyer is not only paying more for the time to book, but is taking more time in the process.

This example is, perhaps, simplistic. But it demonstrates the issues. Further, agents still most play a role in the booking process ... ticketing, submitting audit coupons, fixing problems, etc. On one side, there are those that would argue that self-booking makes "more work" for the agent -- but that's speculative. On the other side, there are those that would argue that self-booking eliminates the dialog of two people (buyer and agent) on the phone at the same time. That too is speculative. These are both factors of the buyer and agent relationship. The realities lie in the shifted and displaced costs.

Finally, as Internet and other self-booking users get smarter (i.e., move up the technological learning curve), the cost benefit for both agent and buyer increases. But due to the command language structure of the CRSs -- and continued dependency on audit coupons for ARC settlement -- there is no significant reduction in the total cost of distribution of an airline ticket. Even at the point where "front-end" technology can match the efficiency of an agent (and that time will come), self-booking tools will still only shift costs. These solutions are unlikely to offer significant cost advantages without a structural change in the travel distribution channel.

Such a change is pending. Virtually all of the CRSs are building Internet/Intranet platforms to serve those who do buy/sell travel. CRSs, airlines, and hotels, and car rental firms, and tour operators, etc. are all building new solutions. For example, each of the CRSs are building data-streaming interfaces to smooth and speed the links between booking tools and the inventory host systems. Concurrently, all of the CRSs are implementing strategies to compete with the direct booking solutions that the airlines, themselves, are building. The key point is that ... these tools are being built. These are tools of the future - but not tools of today.

Thus, its not where you buy/sell ... or how you buy/sell ... or what you buy/sell that is the issue. What counts is the measurable "value added" service that the intermediary tool is able to offer. Agencies and corporations alike should measure these choices very closely - not because of the "hype" that surrounds new technologies or the fear of "missing the wave" ... but to fully comprehend the real bottom line costs of the travel booking process. There may be, in some instances, a cost valid reason for implementing "front end" booking tools. But the basis for that decision needs very careful study ... not just a "knee-jerk" responses "hype salesmen" talking "the Internet myth".