By now you've probably heard about the fact that Orbitz users accessing the site via Macs are seeing more expensive hotel options when they search. But it seems worth clearing up a couple of fallacies.
First, it's not as if the same hotel room is being offered at a higher prices to Mac users. (So no, using Windows to access Orbitz won't get you a better deal.) Orbitz offers the same search results to Windows and Mac users, but Mac users may see more expensive hotels higher earlier in the search results.
And the reason why this is happening isn't segmentation, as some have claimed. From what I remember of the Orbitz presentation at the R user conference a couple of years ago, all sorts of signals go into the R model to try and predict which hotels a given Orbitz user might be interested in: WebTrends data, search history, purchase history, planned travel dates ... and the user's browser and operating system. It would seem that the browser (perhaps Safari / iOS) and/or the operating system is a strong signal that the user is likely to purchase a higher-priced hotel room. The Orbitz system (as I understand it) is designed to rank search results showing the results the specific user is most likely to be interested in at the top of the list. If Mac users, on average, spend more on hotel rooms than other Orbitz visitors, then it's not surprising that more expensive results would appear at the top of the list.
But it's not as if Mac users being singled out for special treatment. Segmentation is a marketing operation (and a conscious decision), whereas the use of signals like these is part of a predictive analytics operation (driven by data, not people). It's important to understand the distinction.
You can read more about how Orbitz uses R and Hadoop to drive its search at the link below.
Revolutions: How Orbitz uses Hadoop and R to optimize hotel search
Not surprising at all, if you're willing to spend the extra cash to buy an overpriced computer, you've probably got extra to spend on a nicer hotel room.
Posted by: fm | June 28, 2012 at 02:57
You need a dose of econ 101: it is just such "signals" which implement "segmentation". Do you think segmentation happens spontaneously, the way people thought flies appeared before science told them otherwise? The seller ideates sections of the demand curve, and poof!, there they are, money clutched in their hot little hands?
No... It requires data to split the demand curve on characteristics which place buyers in each's particular point. So, yes Orbitz (whether using R or not), reads "Mac, more middle class, ergo expensive room". You're attempting to make a distinction that doesn't exist. Not sure why, either? Corporations exploit asymmetric information (because they can afford it, Wall Street is the poster child for exploiting financial advantage) as a matter of course; why defend Orbitz, in particular? While, in this case, the segmentation may be harmless (Mac owners may well prefer to see more expensive lodging than lowly Acer users; I wonder what linux users get, Motel 6?), it is done on purpose.
Posted by: Robert Young | June 28, 2012 at 06:49
Yes, you can use predictive analytics to define segments, but as I understand the term from a marketing point of view, "segmentation" means targeting a well-defined sub-population of your market. This can be via a specialized marketing campaign, or a special offer. That's not the case here: Mac users aren't treated any differently than an other segment. Each *individual* user (again, as I understand the process from the Orbitz presentation) gets a different ranking of search results based on the predictive model. It's even possible for a Mac user to have lower-cost hotels ranked first, if other signals in the model indicate that would represent the most likely purchase for her (a Super 8 regular for example). That's what I meant when I said this isn't segmentation, because "Mac users" is not a defined segment.
Posted by: David Smith | June 28, 2012 at 09:17
I actually tend to agree with Robert, though with a much more cordial tone :-) But "mac users" actually is a very well defined segment, especially since Orbitz as a website knows which OS (or browser, at least) each user is using. More importantly, I disagree that segmentation refers to targeting well-defined groups. Really, segmentation is any kind of differentiated product offering, messaging, marketing, or just about anything based on characteristics of the customer. Back in the day, targeting wasn't nearly as precise as it is today, especially on the web, so you had to break your customers into a few key groups and say "we're going to go after group A this way, group B that way, and group C we're not going to target." The consulting firm I was with out of college made a boatload doing exatly that. But with a site like Orbitz, there's no need to be that discrete and the whole thing happens on the fly, so the distinction between signal and segmentation becomes very unclear.
Posted by: Mike Nute | June 28, 2012 at 12:18
Have a look at this Wiki article: http://en.wikipedia.org/wiki/Consumer_surplus The purpose is to extract as much of consumer surplus as the law allows; market segmentation is the surgical instrument used. I didn't mean to be uncordial, but market manipulation, globally, is a pet peeve. What Orbitz does could be (again, could) seen as of a piece with phishing. The linked piece on price discrimination provides the gory details: http://en.wikipedia.org/wiki/Price_discrimination , there's even a segment on travel industry (a well known offender).
Extracting consumer surplus is more aggressive than just differential marketing, it's predatory commerce, and depends upon asymmetric information to succeed.
It explains what's going on, from an econ point of view. One point, though, is that the supply/demand curves refer to different amounts of a homogeneous product, so what Orbitz is doing might be a tad different. If they're showing Mac users "the same" accommodation from the same lodging house at higher prices, that's the classic case. OTOH, if (and more likely) Orbitz is finding those who'll pay a lot more for a little better, the analysis of what they're doing gets a tad murky.
Posted by: Robert Young | June 28, 2012 at 13:07
Turns out InformationWeek has written up a more nuanced piece (just found it): http://www.informationweek.com/news/internet/ebusiness/240002760
Posted by: Robert Young | June 28, 2012 at 13:44
SAS has a function called CleanMACsignal(). They should have used SAS.
Posted by: SASlove | June 28, 2012 at 14:58
@Robert, not sure I'd call that piece exactly nuanced, given the lede: "Imagine walking into a retail outlet and asking how much a certain item costs and instead of an answer you get a question: How much do you have?". As pointed out above, Mac users and Windows users are offered the same price for the same property. 10 paragraphs down it does get into "dynamic pricing [vs] dynamic presentation", but too many commentators have promoted the "shakedown" fallacy.
BTW, a former Orbitz employee confirmed that OS doesn't affect search results, so it's actually the browser, not the fact that it's a Mac that counts.
Posted by: David Smith | June 28, 2012 at 15:09