Wednesday, 14 July 2010

Google buys ITA Software (Part 3: What's it mean for travellers?)

Monday in Part 1 I discussed the origins of ITA Software, the air travel pricing and reservation software company bought this month by Google for US$700 million. Yesterday in Part 2 I tried to explain what ITA Software actually does (and doesn't) do. Today in Part 3, below, I'll get to the meat of the matter and the question readers have been asking: How is this deal likely to affect travellers?

As I've been saying, "Google's purchase of ITA Software is likely to be a bad thing for travellers." For the why's and how's, read on:

As I discussed in Parts 1 and 2 earlier this week, ITA Software has eschewed trying to model the heuristics of human airfare experts or the analysis of fare structures, and has tried to reduce the problem to one of "search". ITA Software deals with fares and their rules (as distinct from prices for specific tickets) only to the minimum extent necessary to assign prices to results obtained from systematic and periodic queries of availability on all possible flights.

That's an approach that Google, as ITA Software's new owner, is likely to share and want to continue. It mirrors Google's approach to search (crawl periodically, cache locally, index the cache, generate responses to user queries on the basis of the local index). And it mirrors Google's disinterest in the messiness of human indexing, classification, or selection; in anything that can't be completely automated; or in anything that relies on human input or expertise.

But that approach hangs up in two places, which continue to cause problems for ITA Software: the need to continually query external CRS's and airline sources of availability and the inevitably continual staleness of the cached availability data (which could be eliminated if ITA Software itself hosted the airline's reservations), and the need to parse and apply often-ambiguous fare rules to try to determine the price to cache for each availability response (which could be eliminated if seats had individual price tags, rather than being priced according to a separate system of fares and rules).

To ITA Software, the separation of fares and availability is presumably one of the primitive elements of "legacy" reservations systems that they expect to eliminate as soon as possible. And ITA Software claims to be ready, if any airline is willing to be the first to try it, to offer a reservation hosting system engineered from the ground up with as little concern as possible for the prior art.

[One of Google's largest categories of "search" advertising is travel. But everyone in the travel distribution and technology industry has long assumed that Google would prefer -- i.e. would capture a larger share of the advertising and intermediary revenue through -- greater vertical integration, so that a Google search for prices for a particular trip would directly return links to those prices, rather than a mix of "natural" search results and ads linking to third-party Web pages that might or might not have the specific pricing information being searched for.]

So here's the essence of the situation:

  • To move forward on the basis of its current technical strategy, ITA Software's core imperative is to eliminate the separation of fares and availability and the need to query external availability systems, and transform the availability landscape into a flat local database of available tickets.

  • Google's core business is brokering advertising, and Google's core competency is in individualized, automated, real-time micromanagement of ad placement and pricing.

Put them together, and what do you get? The most obvious way for both companies to combine their strengths and overcome their weaknesses would be for them to eliminate or bypass the concept of "fares" and a "tariff" entirely for airlines hosted by ITA Software, and offer tickets dynamically priced for each customer by Google's personalized-pricing algorithms.

What's wrong with that?

The problems with personalized pricing of airline tickets relate to its inherent potential for invidious price discrimination and its inherent opacity (I have no way of knowing what price is being offered to anyone else), which precludes meaningful comparison shopping and precludes government or any other external oversight of pricing (and price advertising) practices.

Personalized airline ticket pricing is currently illegal. Airlines are required by law in the USA and most other countries, and under almost all bilateral and multilateral aviation treaties, to operate as "common carriers". As such, they are required to accept all would-be customers (a common carrier can't "reserve the right to refuse service"), to publish a tariff of fares, and to charge the same fare to everyone complying with the same set of rules and conditions in that tariff.

There are good historical consumer protection reasons (discussed here and here and here) for these rules. In the USA, the Airline Deregulation Act of 1978 -- which authorized certified domestic airlines to fly any routes and publish any fares they wanted for travel within the USA -- quite properly retained, for purposes of transparency, fairness, and oversight, the requirements that airlines operate as common carriers and that they publish their fares in a tariff whose rules are equally applicable to all would-be passengers.

Historically, off-tariff airline ticket pricing was largely limited to international airline ticket consolidators, who were allowed to operate in a quasi-legal gray area of overt non-enforcement, at least in the USA and UK, as a way for airline to offer lower prices despite the unwillingness of foreign governments to approve lower (published) tariffs. But that consumer benefit and justification disappear in deregulated markets -- including domestic US airfares and international airfares on the growing number of routes subject to "open skies" treaties. There is no need, in such markets, to accept the opacity of consolidator pricing as a trade-off for lower prices. If an airline wants to lower its prices in a deregulated market, it can publish lower fares.

Airlines, however, prefer the opacity of off-tariff pricing, both to conceal their pricing from their competitors and to give them a bargaining edge against their customers.

Under the previous several US administrations, the argument that over-regulation would stifle Internet innovation and economic progress, combined with a lack of government resources and expertise related to online travel distribution, allowed airlines to experiment with off-tariff online ticket sales with little risk of enforcement or sanctions. More and more newcomers to the travel industry, even airline staff who ought to be trained to know better, are unaware of the concepts of "common carrier" and "tariff", and assume that they can set their prices and pick and choose their customers the same way any other business does.

Google's acquisition of ITA Software is a significant step towards the eventual emergence of an industry push -- in the name of "market efficiency" -- for repeal, or at least more general and overt non-enforcement, of the "common carrier" and "published fare tariff" requirements.

But in capitalist theory, an efficient market depends not just on competition (regulations governing places of public accommodation and common carriers were developed in significant part to check the local monopolies of the only hotel in town, or the only railroad offering transportation to a particular farm or factory) but on perfect transparency of prices to both buyers and sellers.

Personalized prices are one of the most extreme possible case of information bias against the consumer: The seller knows exactly how many seats have been sold at what prices, how many remain, how much each other customer has paid, and what price is being offered to each other potential customer. Each consumer knows nothing about available inventory (is this the last seat, or is the plane empty?), how much anyone else has paid, or what price anyone else is being offered. Completely personalized prices would tilt the playing field as far against the consumer, and in favor of the airlines, as is conceivably possible.

There are also issues of oversight and enforcement of anti-discrimination and truth-in-advertising rules. If each potential customer is offered a different personalized price, how is it possible for consumers, watchdogs, or government oversight or enforcement agencies to tell whether members of disfavored groups are being offered systematically higher prices, or whether anyone is actually being offered tickets at the prices in airline advertisements? Today, the USA Department of Transportation is overly lax in its enforcement of truth-in-advertising rules against airlines. But personalized pricing would effectively preclude enforcement even if the DOT wanted to act.

We aren't there yet. But Google's acquisition of ITA Software clearly heads us in the wrong direction, toward changes with potentially serious negative implications for consumers.

[Update: There's more background on how airline ticket pricing works in my books, and more comments on Slashdot.]

[Follow-up: Government puts antitrust conditions on Google-ITA Software deal -- but does nothing to protect consumers (8 April 2011)]

Link | Posted by Edward on Wednesday, 14 July 2010, 12:39 (12:39 PM) | TrackBack (1)

A very nice analysis, one that I agree with.

In a similar vein, I am pretty sure that the main reason for the excitement in the industry about mobile advertising is the potential for individual pricing, something I wrote about in my 2009 paper "Network Neutrality, Search Neutrality, and the Never-ending Conflict between Efficiency and Fairness in Markets" in the Review of Network Economics,

in the section on how Google might come to be perceived as "evil."

P.S. You might also be interested in another paper on the general topic of price discrimination and how society is likely to react to it, namely "Privacy, economics, and price discrimination on the Internet," which appeared in "ICEC2003: Fifth International Conference on Electronic Commerce," N. Sadeh, ed., ACM, 2003, pp. 355-366

and has been reprinted in a couple of places.

There is a short follow-up, "Privacy and the clandestine evolution of ecommerce," Proceedings ICEC2007: Ninth International Conference on Electronic Commerce, ACM, 2007,

Posted by: Andrew Odlyzko, 19 July 2010, 14:26 ( 2:26 PM)

Your conclusion presumes that airlines will always charge higher prices when given additional personalized pricing power. The reality is that additional flexibility will create downward pricing pressure just as often, and on balance there should be no change. In other words, airlines would do well to charge more for certain inventory situations, but doing so will lead to just as many situations where pricing will be logical at lower levels. Some consumers will benefit, others will pay more, and on balance, it will be a wash.

Posted by: Peebus, 20 July 2010, 11:19 (11:19 AM)

Thanks for the informative series of posts. I'm sceptical about your fears of per-customer pricing; I'm not aware of Google charging advertisers more based on the searcher, only on the keywords in the search. Perhaps they do, they certainly could. But couldn't any other travel agent site do the same?

Is there any chance that displaying more accurate fares could increase competition between those setting fares?

Posted by: Douglas, 20 July 2010, 11:38 (11:38 AM)

At present "Revenue Enhancement" programs do personalize availability of seats. They do this by getting your whole itinerary from the CRS, before accepting a passenger. They calculate the portion of ticket price they will receive from their leg of the flight (using IATA rules) and if it is too small then no seat will be available. They pay for this full itinerary information from the CRS. The problem is that IATA rules split the money between multiple carriers by mileage, not by costs (airport landing fees, etc.) E.G. the short haul flight provided by a local airline to the final destination will not be profitable because the long haul part of the flight (greater mileage) will take all the money. This analysis is for full fare IATA "businessmen's" flights.

Posted by: Jim Bennett, 20 July 2010, 14:17 ( 2:17 PM)

Interesting analysis, but hasn't the boat already sailed? I find it impossible these days to find a human being in a travel agency to book a ticket. The remaining ones have become tour operators selling package holidays. The only people using humans to book airline tickets are corporate agents servicing large organisations.

Meanwhile, the rest of the world buying tickets for personal use are completely dependent on the Expedia and Travelocity's of the world, or going directly to the airline's site. And into the world of personalised, opaque pricing that you lament. Further, all these places make it difficult, if not impossible, to book anything but the simplest of itineraries. But that's what the world is moving to, and I dont see that changing unfortunately. Google's acquisition of ITA is only one step in this process.

Posted by: AS, 20 July 2010, 17:05 ( 5:05 PM)

The interesting contrast is Google's favor of "net neutrality".

Posted by: Thomas Whaples, 20 July 2010, 17:37 ( 5:37 PM)

I find your analysis to be fairly sounds but you may have started with an unlikely premise. "The most obvious way for both companies to combine their strengths and overcome their weaknesses" Airlines are unlikely to switch from there current legacy CRS systems however cumbersome they may be to a hosted one for the simple reason that they do not see the cost benefit in doing so. Most large airlines like KLM Air France or American Airlines have already fully integrated their internal procedures and pricing systems into their CRSs. To change this would create enormous costs that do not have a direct return for that particular airline. Google must know this since this is a common problem is software migration and Google is essential a software company. This means that until Google can somehow lower the cost to the airlines for migrating the status quo will likely remain.

As far as personalized pricing is concern, I do not think that this will be Google's strategy. They are likely to take an approach similar to their "Froogle" service now called "Google Base" which gives an overview of choices to the buyer. Google's core goal was/is to index the world's data, if they follow that idea they may simply provide a overview of tariffs with some form of ranking, which like their web page raking system is independent of the individual providers' wishes (with ads on the side for hotels and car rentals). This may not be a paradigm shift but is likely to increase Google searches. Most Google services are designed by idealistic software engineers with little governing by management as to what direction they should take, so a tweak-able ranking system like that seen on most on-line travel agency websites will likely be included.

As far as the acual booking of the ticket and the payments there is very little inovation possible in this area other them a Google Checkout integration to handle transactions.

Posted by: Chris M, 21 July 2010, 06:34 ( 6:34 AM)

From the perspective of someone who grew up reading OAGs for a living, this is quite interesting.

But I have a better idea for Google: use the $700 million to provide a $1 million, incentive to each Congressman and Senator to re-regulate the airlines.

In the days when the F-Y-K prices to go from here to NYC were the same for all airlines and published in a book that almost anyone could look at, we had efficient service, decent seats, reliable, safe and friendly pilots and crew. The idea of nickel-and-diming passengers was the farthest thing from the minds of airline executives. Southwest alone among US carriers can still make some of those claims, and is notably absent from ITA's website.

Today, we have chaos and poor service at a price yet to be announced. People used to love to fly; now I can't find anyone willing to do so except under duress.

Posted by: MichaelM, 21 July 2010, 07:18 ( 7:18 AM)

@Chris M. - You suggest that airlines will be willing to change hosting providers only if a new provider is cheaper. But hosting providers don;t just provide automation services; they also provide pricing and other optimization services. Improvements in revenue optimization may generate enough revenue to justify and otherwise-costly hosting chnage, as I discussed earlier this year in relation to Jetblue's switch to Sabre:

Posted by: Edward Hasbrouck, 21 July 2010, 08:56 ( 8:56 AM)

I'm sure you understand this far more deeply than I, but just based on my understanding of your analysis I am not sure I agree with the summary "Google's purchase of ITA Software is likely to be a bad thing for travellers."

It sounds like from your analysis that Google is most likely to maintain the status quo. While the status quo would not be an improvement for travelers that doesn't mean it'd be necessarily worse either. Perhaps "Google's purchase of ITA Software unlikely to benefit travelers" would be more accurate?

You primary argument for it being worse is Google's targeted advertising, however taking Google search as an example Google has, IMHO, done a very good job of keeping paid search results separated from and to a minimum in relation to natural search results. This behavior is similar for Google Shopping, therefore I'd watch carefully, but expect the same approach to travel pricing related search.

In summary while Google does personalize their targeted advertising they do not personalize their natural search results and do a good job of making the distinction between the two.

Going a little off topic, I find ALL the existing fare search engines lacking. None that I've seen offers pricing based on what you're actually buying. I want on a single page to see what the baggage restrictions and costs are, what the change fees are, what in in flight food is and may cost, what the seat dimensions are, if there are laptop power jacks on each seat, etc... Too many consumers are choosing flights strictly on price without enough detail as to what they are actually paying for. Such price centric search engines are a deservice to both the customers (who buy blind) and the airlines (who are forced to reduce quality in the pursuit of lowest cost).

Posted by: Jon Brown, 21 July 2010, 13:58 ( 1:58 PM)

I have found these articles to be very interesting and the comments equally so. Some were so interesting that I feel that I must add to the comments.

There are private fares out there that are done between airlines and corporations and airlines and travel agencies.

Someone said that they didn't think that personalized private fares would drive the price up, I must disagree. What keeps the fares down right now is the knowledge and ability to search for lower fares. If fares are personalized, consumers will have lost the ability to compare fares and thus find lower fares. Having worked in the travel industry for over 25 years, I can safely say that the airlines would love to have personalized fares so that they could raise fares.

The individual who says that there aren't agencies that will do airline tickets, hasn't shopped around very much. I can find lots of corporate agencies that airline tickets and almost nothing else, other than hotels and cars. If you go to a travel agency that specializes in vacation travel or cruises, you won't find an airline ticketing expert, because that's not what they do. That would be like walking into a GM Truck dealer and expecting to find a Prius for sale. Some of the largest, in terms of revenue, agencies are those that do just business travel and according to the annual studies done by Topaz Inc. out of Portland, they do a better job on finding lower airfare than Orbitz, Expedia, etc. and they provide better service.

And to those who call for reregulating the airline industry, I would point out that the only winners in that would be the legacy carriers. Everyone else would suffer. Consumers would pay more because airlines wouldn't have to be worry about controling costs. New airlines would be at a disadvantage because the regulatory barriers would add even higher hurtles to creating a successful new airline. This means less competition, less innovation and a return to the days when air travel was for the well to do and everyone else either stayed home, drove, took the bus or the train. Deregulation has made air travel cost effective and opened up the skies to the general public.

Thanks again for the interesting read.

Posted by: John M, 21 July 2010, 17:17 ( 5:17 PM)

For a non-expert, airline pricing has been opaque for some time. When I look up airfares, they'll vary by $100 depending on what day I look them up, for the same itinerary. I'll allow that there are hidden regulatory and technological mechanisms that keep it from being worse, but the only thing I can think of that would be worse is if prices were higher (which seems to happen anyway). If you make a "transparent" system complicated enough, it might as well be opaque.

"If each potential customer is offered a different personalized price, how is it possible for consumers, watchdogs, or government oversight or enforcement agencies to tell whether members of disfavored groups are being offered systematically higher prices, or whether anyone is actually being offered tickets at the prices in airline advertisements?"

I already thought each customer got a personalized price; that's certainly how it appears, and I have no way to judge otherwise.

If it takes years of experience and training to understand and navigate this "transparent" system, the airlines *already* have all the information. You're worried about a tilted playing field between airlines and consumers; can you explain how we're not already there?

Posted by: Chris D, 31 July 2010, 11:49 (11:49 AM)
Post a comment

Save personal info as cookie?