Join 3,494 readers in helping fund MetaFilter (Hide)


What drives ticket site pricing mechanisms?
February 21, 2011 5:52 PM   Subscribe

Do airplane ticket sites (Expedia, Travelocity, Kayak, etc.) increase prices between successive browsing sessions, as well as manipulate prices so that they are in agreement with each other?

If I browse for a plane ticket from site S1 at time T1, I can buy a ticket at price P1.

If I browse site S2 at time T1 + five minutes, I can still buy a ticket at price P1. The price doesn't change between sites S1 and S2 (or Sn).

If I browse sites S1, S2, ..., Sn at time T1 + an hour, I can now buy a ticket at price P2, which is often substantially more than P1 (roughly 20% or more of P1).

If I come back a day later, the prices are still hovering around P2 — or higher. If I come back a week later, the prices are, again, still around P2 or higher. I never again observe pricing equal to or lower than P1, unless I browse from a different IP address (such as a work computer).

Is it standard practice to do price-fixing between ticket sites, which shares the visitor's IP address or other identifying information not only to fix prices between sites, but also raises and maintains higher prices if the visitor returns to any ticket site after a period of time?
posted by Blazecock Pileon to Society & Culture (19 answers total) 20 users marked this as a favorite
 
It's a standard practice to have a certain number of seats of a certain fare class allocated at a certain price.

You aren't seeing a conspiracy, you are seeing a seat get sold.
posted by bensherman at 5:58 PM on February 21, 2011 [4 favorites]


My sister travels *all the time* and is *always* on these sites (like weekly)... and she insists the same that you do, that prices get manipulated based on browsing sessions. I now close the browser and clear cookies before logging onto a new travel site. True? Who knows but now you're the second person I've heard mention this.
posted by braemar at 6:04 PM on February 21, 2011


I use Bing to search for tickets. It shows ticket pricing histories and trends with recommendations to buy or wait. These trends have never appeared to depend on my browsing habits.
posted by mnemonic at 6:12 PM on February 21, 2011


According to my sister-in-law, who has been a travel agent for 25 years, you are being misled by coincidence (the seats are getting sold). The SABRE system is huge and unwieldy, and if it could be that responsive to cookies, it could do the things it's actually designed to do much better and faster.
posted by Sidhedevil at 6:14 PM on February 21, 2011 [3 favorites]


Addendum to my post: my experiences include searching on Bing for tickets from different IP's (I'll search from my home computer and work computer and see the same results).
posted by mnemonic at 6:16 PM on February 21, 2011


I'll one up you on this theory. Not only do I experience this but AFTER I buy the tickets, I'll then go back and check prices and inevitably and without fail the prices will have dropped.
posted by Sassyfras at 6:19 PM on February 21, 2011


Consumer Reports looked into these claims. And they did some testing. It's worth reading, but the bottom line is: The airlines can claim it's all sorts of stuff, but people with experience say it's worth clearing out your cookies frequently while you search for tickets.
posted by circular at 6:27 PM on February 21, 2011 [7 favorites]


Dynamic pricing. I'm a bit of a flight-bargains hunter and I have noticed this phenomenon. I remember seeing it mentioned on AirfareWatchdog as well.
A quick Google yielded these links specifically referencing buying airplane tickets:
Smarter Travel: Do airlines hike fares when a route is searched often?
Consumer Reports WebWatch: E-Commerce Vendors Use Online Technology to Create 'Dynamic Pricing'
Consumer Reports: Before you click "buy" again, read this!

I get around this by checking prices on ITA first to confirm the route/pricing, before going on ticketing sites.
posted by flex at 6:27 PM on February 21, 2011 [10 favorites]


also, many sites are just UI fronts (like kayak) for the same underlying engine (like ITA home)
posted by victors at 6:52 PM on February 21, 2011


I won't believe the conspiracy until I see results of two people sitting next to each other doing the same things and getting different results. Seems to easy for confirmation bias to creep in- even CR did it- "oh, the prices were different last week when we checked".

It seems awfully complicated and not worth the effort to price this way. *Maybe* some kind of "first page view buy now discount" is possible, but anything else seems unlikely.

Dynamic pricing applying equally to everyone, however, makes sense. As the supply runs out, you want to raise the price to meet the demand. (Presuming that a seller would rather have stock in something than have to say we're all out.) Changing prices based on how many people are searching for a thing? Maybe. If a site is suddenly getting an abnormal rate of non-buys (surely they know that X views usually = Y sales), then maybe they drop the price. Or raise it if the opposite is true.

Especially with airline seats and hotel rooms. As noted above, they slot certain #s of seats in certain price groups. When the cheap ones run out and the plane is nearing full, the price goes up so that only the buyers who NEED to be on THAT flight will buy. Other buyers will choose other less full flights. Keeps supply and demand even. And the converse is true- as flight time approaches and the flight isn't full, they will drop the price to fill it up. At that point, they just want to get some money flowing in their direction, as the costs of flying an empty airplane and a full airplane are roughly the same.
posted by gjc at 7:45 PM on February 21, 2011


Another voice in the "I don't have scientific proof but it sure seems this way to me too" crowd.

Also, from a technical perspective, dynamic pricing is not complicated. Just sayin'.
posted by hapax_legomenon at 8:07 PM on February 21, 2011


Why don't we test this right now? Choose a round trip on expedia April 22-29. Report the 2 flight numbers and the price(s).
posted by cashman at 8:19 PM on February 21, 2011 [1 favorite]


I won't believe the conspiracy until I see results of two people sitting next to each other doing the same things and getting different results. Seems to easy for confirmation bias to creep in- even CR did it- "oh, the prices were different last week when we checked".

My partner and I did this after hearing an NPR story about the phenomenon. We each used three different browsers (two of which we never or rarely use) and two laptops. We even searched the airline sites directly. Prices for the chosen flights were always the same.
I'm still not totally convinced, but I'm much less skeptical.
posted by swingbraid at 8:38 PM on February 21, 2011


Er-- I mean, I'm not totally convinced that it doesn't matter how often you search for seat on a particular flight.
posted by swingbraid at 8:40 PM on February 21, 2011


Airline ticketing, scheduling and pricing at the level they do these days is far too consuming to worry about whether the person getting a ticket price is the same as some other user in the past, even if you're trying to tell them so. It's too complicated to take non-buying return shoppers into account.
posted by rhizome at 9:13 PM on February 21, 2011


I think it's the cookies! This happened to a colleague of mine while we were in some downtime during a meeting. He was booking a flight for his next trip, browsing, and bam! Price went up by a few hundred bucks. I'd heard the cookie rumor, so we decided to test it. Cleared the cookies and another bam! Price was back to where it was for the same flight & seat class. So that's a data point.
posted by smirkette at 10:12 PM on February 21, 2011


I used to work in the airline business, including on dynamic pricing issues, though that was back before the growth of these particular booking sites.

The fundamental thing that is going on is that the airline is juggling two competing desires, a) Filling as many seats as possible, and b) Not selling a seat to person X when there is a healthy chance that some person Y would be willing to pay more for that seat at a later time.

The upshot if this is that usually prices rise as the plane fills up. But they can also fall if the plane is filling up slower than predicted.

There is no particular reason for the airline to care about what price they showed you an hour or a day ago, esp given that you chose not to buy at that price point. (Think about it... does any other org you decide not to buy from come back to you with a higher asking price after you start to walk away?)

If you think you can guess what the airline systems are doing from a few data points, think again. There are forecasting and optimization processes running as far as a year ahead of the flight, with reforecasting and re-optimization happening ever more frequently as departure nears.

The basics of what's going on is described in this PDF from a Dartmouth MBA course. But that is just an intro for students, the reality is far more complex and sophisticated.
posted by philipy at 12:58 PM on February 22, 2011 [3 favorites]


I don't use Expedia, Priceline, etc, but I do use Kayak a LOT (multiple times daily, sometimes, for several days at a time, up to every day a month), and this does not happen to me. I keep spreadsheets of prices/routes/times-of-flight, too, so I'm quite sure.

Occasionally, I'll get a result I'm not expecting, and it has to do with a specific airport in a region getting checked/unchecked, nonstop getting checked, or some other change that was made to the search parameters (sometimes by me accidentally, sometimes it just seems to load funny). Of course, clearing the cookies would fix that, but so would unchecking the box or reloading the home page, and it's certainly not some sort of price-raising strategy. Just a UI bug.

Of course, I'm seeing the results from the airlines/SABRE and not from a third-party ticket seller, so I guess that could change things. It seems weird, though, if Expedia et al really do that and the airlines themselves don't. It seems like it would take away any advantage of getting tickets through a third party in the first place.
posted by wending my way at 1:44 PM on February 22, 2011


I don't use Expedia, Priceline, etc, but I do use Kayak a LOT (multiple times daily, sometimes, for several days at a time, up to every day a month), and this does not happen to me. I keep spreadsheets of prices/routes/times-of-flight, too, so I'm quite sure.

Do you clear your cookies too? Because that's the only way you'd know -- otherwise the prices will always be consistent with your spreadsheets, because that's always the prices you're getting shown.

RyanAir do this pretty obviously -- price goes up on day 2, clear cookies and it goes back down.

This:
There is no particular reason for the airline to care about what price they showed you an hour or a day ago, esp given that you chose not to buy at that price point. (Think about it... does any other org you decide not to buy from come back to you with a higher asking price after you start to walk away?)

is not true in their case. Booking a flight isn't a single decision for most people. They'll price up flights over a couple of days while making their other plans, talking with friends/family/colleagues. So then they check Ryanair again, and oop! price is going up. Better buy now, now, now before it gets any higher. Putting up the price lets them force dithering potentials into a decision.
posted by bonaldi at 5:29 PM on April 5, 2011


« Older Dreaming of a maritime summer ...   |  Seeking long-lost interactive ... Newer »
This thread is closed to new comments.