Hacker News new | comments | show | ask | jobs | submit login
Google Flight Search (google.com)
516 points by revorad on Sept 13, 2011 | hide | past | web | favorite | 278 comments

It’s rather powerful: as long as the origin airport is one of those shown on the map, and the destination airport is reachable via a domestic flight on one of the major airlines (AA, AS, B6, CO, DL, F9, UA, US), the search results come back instantaneously. Multiple origins and destinations (up to five of each) are supported as well:


And the results still come back immediately! Including the three-month chart of prices. This leads me to believe that Google/ITA has precomputed all of these results, and is simply serving these results out of a cache of some sort. That would explain why they are only offering a limited set of origin airports at this time: it probably takes an incredible amount of computing power and storage space to pre-calculate all of the possible results.

Colour me impressed.

I would think the limiting factor is availability. You can pre-compute valid fares between cities super-fast with sufficient computing power, but you're still (somewhat) limited by the rate at which you can update availability, which _still_ has to be polled from airline systems (to some degree) and can't give you answers in <400ms :)

One of the bits ITA/QPX got really right was smart availability caching (http://hasbrouck.org/blog/archives/001880.html) -

I would think they can do the fare search/assembly in realtime with their infrastructure, but they're relying on cached availability, hence the limit to "popular US cities" at the moment, to keep caches fresh enough within the practical throughput limits.

Expanding on this, one of the things ITA/QPX comes with is a dirty big availability cache that's constantly being primed (and paid for) by their customers (ie: Kayak, Orbitz, et al), who are now ironically not getting any traffic from Google :)

And they have this: http://itasoftware.com/pdf/BrochurePDF_DACS.pdf

Interesting. That makes me wonder: would that really be the optimal approach? Google is able to return search results for previously unsearched queries nearly instantaneously too, so I'm wondering whether it's really a cache that's serving those queries. Would some sort of graph-based index be possible in this situation?

Yes, using a shortest path tree [1], it is possible to efficiently (for some definition of "efficient") calculate the best path from a single origin node to all destination nodes in the graph. Dijkstra's algorithm is the standard implementation, but is not very amenable to parallel processing.

Perhaps unsurprisingly, Google has a framework called Pregel [2] which does parallel graph processing. Their research paper outlines how it can be used to calculate a single-source shortest path tree (in section 5.2) in reasonable time.

[1] http://en.wikipedia.org/wiki/Shortest_path_tree

[2] http://www-bd.lip6.fr/ens/grbd2011/extra/SIGMOD10_pregel.pdf

Here's the classic presentation people usually refer to when talking about flight pricing complexity.


I'd love to know how Google's doing it - it's a really hard problem.

What's interesting is that you linked to an ITA Software presentation, and Google own ITA Software. So it's almost certain that some of the people who put that document together worked on this pre-computed solution.

If I remember correctly, it's mentioned in Artificial Intelligence, A Modern Approach that this problem is not only hard but in fact formally undecidable in the general case.

Well, if one narrows down the options a bit, it becomes far more tractable. Considering that Google's flight search only works for domestic return trips of 15 days or less (it didn't seem to give me results for anything longer than that?) and doesn't allow stopovers or open jaws, it's entirely possible that the complexity is much reduced.

Lol. If I were Google, I'd probably have bought all the tickets in advance ;)

Thing that gets me is, do they really send their employees SFO to LA on United?! Southwest's like a third the price... And why isn't SW listed, anyway? Did Herb take issue with something here?

This information is a decade old, but my wife told me that Southwest were the only American airline to make a profit. That could have something to do with it...

EDIT: This recent news article states "Of the five biggest U.S. airline companies, analysts expected only Southwest to report a first-quarter profit":


Some of Southwest's profit comes from fuel price speculation, which could backfire on them much more quickly and easily than their core business:


From what I understand fuel contract hedging is a pretty standard part of airline business. Obviously hedging has potential to go the other way but it's unlikely to erode profit unless competitors are able to significantly compete on price and steal customers...

Probably for the same reason you don't see Southwest flights on any travel website. Southwest doesn't allow their flights to be booked outside of their website.

Seems they are only pre-computing for next month or so..


gives: Your search – New York to Chicago – did not match any results.



1347 of 1347 roundtrips shown

Trip length duration is currently limited to 14 days.

It's odd that the limits are just in the data it seems, not in actual limit checks. So when you hit a limit, it's indistinguishable from there being no available flights... which seems very bad for UX.

Published airline fares change only a couple of times a day (10:30am, noon and 8pm), and only once a day on the weekends. Airlines (like American) are trying to push travel sites to use more dynamic pricing. The data is cached and outdated (including availability).

It doesn't take a tremendous amount of computing power -- remember that most airlines are still running ancient mainframe reservation systems.

The multiple destinations functionality was from ITA -- it's been around for years. Check out flyertalk.com for discussions -- you can input multiple layovers to have it compute cheap fares that maximize miles/segments. They keep a basic backend open at: matrix1.itasoftware.com

Internally, this has been around for a little while (disclaimer: I work for Google but not on anything related to this).

What continues to impress me about Google is:

1. Just how quickly this was built (really, it was quick); and

2. Google wants you to use our services because they're compelling not because we don't give you any other choice (ie "Don't be evil").

Sure there are limits to what it currently does but I think you'll see it rapidly iterate.

Let's be honest. Google wants us to use their services because the better they understand us, the better they can advertise to us. That's the profit model at Google. I don't begrudge them for it either. You've got to turn a profit if you want to keep the lights on.

I don't mean any offense, but the "Don't be evil" mantra is growing a bit tired. Evil is a strong word, and "not being evil" is a pretty low bar, IMO.

If you drove past a restaurant that had a big sign outside that said "no cockroaches here", what would you associate with it?

I have always been amazed that some people would consider "don't be evil" to be a motto with any power to motivate or lead people, since it ventures no strong opinion, and points in no clear direction.

"Do good" is far more costly, less likely to be politically correct, and more likely to invite ridicule. At the same time, this kind of heart-on-sleeve sincerity has more power to inspire. Sincerity, humility, innocence tend to achieve more than cynicism and jaded worldly wisdom.

> If you drove past a restaurant that had a big sign outside that said "no cockroaches here", what would you associate with it?

In a town in which I expected from experience to encounter a few cockroaches during any visit to any restaurant, I would find that a reasonable, positive, encouraging message to advertise with.

Think of the Golden Rule:

Should one treat others as one would like others to treat oneself? (the positive form)

Should one not treat others in ways that one would not like to be treated? (the negative form)

I consider the maxim “Don't be evil” to be more actionable, useful, and meaningful than “Be good.”

I think Don't be Evil is just easier, since you could make the argument that you could theoretically actually enumerate a list of things that most people would consider truly "evil". Do Good... well, good luck making that list.

So, uh... in hindsight I guess I'm just agreeing with you :-P

Sadly, the set of "evil" things is also uncountably infinite.


I thought about that in the wrong way.

What I really meant is it's just easier to not do evil. You can look at something and say "that's evil, so I won't do it". Sure, sometimes that can cause you hardship, but you deal with it. Maybe evil is an uncountable, but you don't need to count it. You just passively avoid options that are evil.

Doing good is harder. You have to actively seek out and choose good things to do from a huge set. That means attempting to enumerate at least part of that set. You have to figure out how to actually do those things. As a public corporation, you have to figure out how to do things and spin it as being in the company's best interests.

Yes, it's also harder because it requires leaders with strength of character, willing to be misunderstood, willing to go against the grain, willing to stand alone at times, even in the face of opposition.

Weak leaders who don't do evil, but who also fail to do good, or fail to actively advocate and define doing good, are as much part of the problem because they fail in their responsibility of teaching, rebuking, correcting and training in matters of service.

You would only need to "spin it as being in the company's best interests" if the employer truly believes that business on any basis other than service is the most profitable.

Or, you know, they could just be taking a cut for connecting users with flights better and faster than their competitors. But please, keep thinking that google is hoping to populate some evil database with precious, precious data to make their advertising 0.000001% better. There's no way they are considering the obvious and profitable notion of getting serious about travel search.

No probably not 0.000001% better, but if they make it 0.1% better that's $28 million per year. That 0.1% is probably doable optimizing ad placements after long haul flights.

Oh, but Google is very serious about ads. They will make this product as good as it needs to be to sell ads, and no better.

Making money from selling things and leaving the user's data alone is just not in the DNA of this company.

"Blub"--The nerve!

I strongly disagree. Google wants us to use its services because at its core is a bunch of engineers that simply want to make the world better.

What Google "wants" as an entity is different from what the individual component humans want.

(i.e. the agent 'Google' is distinct from the agents that it is composed of)

Google is strongly guided by the agents that shepherd it.

Are you aware that several Googlers resigned after a petition signed by 10% of the entire engineering workforce was ignored by management?

See https://plus.google.com/100010618263697835406/posts/azTW8hvw... for details.

Would your complaint sound strange if it was worded, "Are you aware that several Googlers resigned after a petition ignored by 90% of the workforce was ignored by management?"

Ultimately, some policy on some random social networking site nobody uses is not much of a big deal. If you don't like the policy, don't use Google+. Easy.

"It seems that the G+ team is currently overwhelmed with technical issues of the current system -- that, combined with Vic's attitude, means that nymwars is not going to be addressed anytime soon. If engineers can scrape together 20% time to implement possible solutions and manage to convince decision-makers to let them roll out changes, we might see something."

It sounds less like it was ignored and more like it isn't a priority for a resource stretched team.

You mean agents like Eric Schmidt, that declare online privacy as dead?

Online privacy never existed. If it is online it can be leaked, stolen, or shared.

People still believe in a benevolent google after the App Engine scandal? Let me ask you something; what would Google have to do before you finally believe they're just a publicly traded company?

The App Engine pricing issues are anything but a scandal.

Really? Getting a bunch of developers locked into a platform and then raising prices sometimes 100 times more than it was? What would they have to do for it to be a scandal for you?

If Apple suddenly came out and said "right, from today on we want 75% of your sales instead of 30%" would you have the same reaction?

Terrible comparison. A dick move maybe but not a scandal.

Google App Engine:

* Not the dominant hosting platform. Not even close to being the dominant.

* Maybe a million dollar product. Certainly not a billion dollar product.

* In beta/preview until just now with developers knowing pricing had not been set or could be changed.

* Optimizing your application properly can negate most if not all of the pricing increase.

* Open source alternative framework available if you want to keep using the same APIs.

Apple App Store:

* Dominant market position with only the Android Market coming close.

* Billion dollar product.

* Existed as fully launched for several years.

* Set percentage. You can't change it other then making your app free.

* No open source alternative that supports the same APIs.

Beta/preview products can and will change.

the first motivation of any for profit corporation, especially one as large and successful as Google, is profit. That is their entire reason for existence, don't fool yourself into believing otherwise.

People like to say this, maybe believing it fulfills some deep-seated need for cynicism, or makes the speaker feel wise and worldly. But what is your actual argument in support of this? Just saying it does not make it true.

If profit is the only motivation for decisions at Google, why did it pull out of China? I'm sure you're cynical and wise about that too, but here's what I can tell you. I saw Sergey up on stage saying the same things he said in this interview (http://bits.blogs.nytimes.com/2010/03/22/interview-sergey-br...) and he got fiery and emotional about it. This is a guy who grew up in the Soviet Union until he was 6 and for him it is a very personal and ideological issue.

This corporations-as-profit-seeking-automata meme is old and tired. At the end of the day it's people who make decisions, and just as in every aspect of life people can have complex motivations for the decisions that they make.

I never said profit was their only motivation, but it is the reason they are a corporation, and not a charity.

I refer you to the comment above by jey

  What Google "wants" as an entity is different from what the individual 
  component humans want.
Although Sergey is an obviously important influence at Google, he is not Google.

You said profit was the "first" motivation. You haven't even justified that. Saying that a corporation has to at least break even to avoid being a charity is a far cry from your previous grandstanding. Just because a corporation has to make money doesn't mean that money is the "first" motivation for every decision. It doesn't explain why Google (not Sergey, Google) pulled out of China.

Um...they chose to go into China in the first place, and they didn't do it for charity work. A lot of people thought that was pretty evil.

In any case, the parent doesn't have to justify anything, because he's stating a relatively well-known argument: in the US, there's an obligation for corporate executives to maximize shareholder value. This point is debated (http://www.linkedin.com/answers/law-legal/corporate-law/corp...), but not to the extent that you're claiming.

They pulled out of China because the government was hacking them! Do you think they had some knight-in-shining-armour reason? That was all just PR.

I never said profit was their only motivation, but it is the reason they are a corporation, and not a charity.

Not necessarily. To parrot one of the other posters here: saying it doesn't make it true.

You could also argue that being a charity would be a very slow and inefficient means to achieving many of the things Google has set out to do. Being a for-profit corp gets them there faster and with fewer distractions not core to their purpose. I'd even venture to guess that it wouldn't be possible for Google to do what it does as a non-profit.

Sure, profit is a motivation, but it's certainly not the only (or even necessarily primary) motivation.

An interesting take on the subject by pg:


"That was a surprising realization. Companies often claim to be benevolent, but it was surprising to realize there were purely benevolent projects that had to be embodied as companies to work."

>why did it pull out of China?

Do you mean "why did they make a big rant about pulling out of China and then not actually do it until they were literally cyber-attacked by the government"? Google was never all that big in China anyway. I'm sure their whole strategy around this was to get gullible people to buy into the "Don't do evil" nonsense.

>Sergey up on stage saying the same things he said in this interview and he got fiery and emotional about it.

Yea? So do politicians. Very passionate about what they're talking about until it comes time to vote. Then they follow their wallets.

>This is a guy who grew up in the Soviet Union until he was 6 and for him it is a very personal and ideological issue.

I know plenty of people who grew up in the Soviet Union. What exactly do you think goes on there that makes the "lucky few" who escape want to change the world? Maybe it did have some profound effect on him but it sounds more likely to be simple positioning. Getting people to think he really cares will cause them to defend him every time his company betrays that trust.

>This corporations-as-profit-seeking-automata meme is old and tired. At the end of the day it's people who make decisions, and just as in every aspect of life people can have complex motivations for the decisions that they make.

It may well be old and tired but it's how the world works. The only thing a CEO can get in trouble for not doing is increasing share holder value.

You've been conned.

There is a lot of truth in this, but what impresses me most about Google is the way they have aligned their business model with their users interests and how that has enabled them to sacrifice short-term gains for long-term profits... it's a very Buffett-esque strategy that is difficult to pull-off.

True, but maybe they think the best way to achieve profit in the long term is by building cool things and not being evil.

Profit is a large motivation, for sure. There are some folks at Google (I'm guessing particularly on the business side), who probably have that as their primary motivation.

I think the core of Google is Larry, Sergey, Urs, Craig, and the other dozens of engineers I've read about (never met), and from their actions and words over a course of a decade, I think they've proved that a huge motivator for them all is to simple make the world better.

They've all got lots of money. I think the long term evidence shows that they care about doing good. I think their motivation for earning money is to expand their own ability to do good.

If someone disagrees, instead of responding with more argument I'd ask you to make a table and list everything Google done has done in one column and whether or not it seems "good" or "evil" in another. Then we could debate further.

Haha, if you want to disagree with my rose-colored review of why Google are so benevolent then please do a bunch of work so I can just sit here nitpick with nearly no effort on my part.

Google has a dual-class stock structure which gives class B holders (Schmidt, Page, Brin) 10 votes per share v.s. 1 vote per share for the rest.

This is relevant to your statement, because while it's pretty much a requirement for a publicly held company to pursue profit at any cost, it's not necessarily the case for a company run by the people that have a disproportionately large number of votes on major issues.

I'm pretty sure they still have over 50% of the votes, so together they can decide how Google will handle tough moral issues without any support from other shareholders.

I would argue that it's entirely plausible that a couple altruistic billionaires will sacrifice a few dollars in pursuit of non-evil (their opinion of non-evil mind you).

I saw a documentary about advertising agencies. They interviewed someone working at the marketing division of nike. She said her job was to get people to play sports. That of course is not true. Her job is to get people to buy shoes. I think your comment is misguided as well. Google really just wants people to click on ads.

Yes, Google wants to make more money. No question about it. But why are you so eager to reject any other possible motivation just because profit-making is involved? There is satisfaction in making the world a better place; why wouldn't Sergey & Larry want to get that satisfaction?

Exactly! I can't agree more on that. I can get a job which will give me trise as much as I make now with my startup+freelance jobs. But that wouldnt give me satisfaction. I suppose even seeing how people react on what you're up to on scale that Google operates has to be cool.

I was responding to this:

"I strongly disagree. Google wants us to use its services because at its core is a bunch of engineers that simply want to make the world better."

I don't doubt that the people working at google want to make the world better and work on cool software. I don't think that is the primary motivation of google as a company.

I can't tell if this is sarcastic or not.

Wow, wtff? I don't understand how people can get so up in arms about Apple fanboys so long as google has fanboys like this.

Are you serious or trolling? I would expect if someone had such an ignorant/naive opinion they would at least be smart enough to not let other people know.

Here's what I see from google:

Search - a product that helps them sell ads

Gmail - a product that helps them sell ads

Google App engine - A platform where they used the hype of MapReduce et al to get people to join and then ramped up pricing drastically after a lot of people had been on the service (a service for which switching away almost certainly means a big rewrite) for years

Android - phone OS to help them sell ads.

I'm not seeing a lot of "make the world better" in there. Unless you mean "for Google executives".

Nope, I don't think this one is about advertising. My guess is they want to move into the hotel referral sales game (it's not secret they've been trying to move into this space for a while). Before you can move into hotels though, you have to have a good flight search engine.

I am reminded time and time again how much "Don't be evil" mantra still holds for Google whenever other evil corporations do things like suing companies for looks of a product, primarily to hamper competition, and when they take cut out of sale made on their platform without adding any value whatsoever to the payment platform.

When was the last time Google abused its set of patents to screw competition? There are so many ways in which Google can be evil, but it consciously chooses not to be.

Edit: Grammar

Stupid lawsuits about product looks are almost irrelevant to the society and individual.

A centralized database of individual's profiles is a much more dangerous tool. Google's efforts to destroy the idea of online privacy are much more dangerous to society.

I think, google is the one protecting the idea of online privacy by not having stupid restrictions on exporting your own social data and by allowing the users to decide what gets shared with whom (circles).

And stupid lawsuits are not irrelevant to the society, they hamper competition which in turn hampers technological progress. If you think Samsung Galaxy Tab and Ipad look similar, just look again at their screens side by side the difference between PLS and IPS is day and night, it is unfortunate that the court does not deem that as differentiating enough.

Patents are just one way to be evil. Another way to be evil might be to let Canadian companies advertise illegally to sell prescription drugs in the US.

From http://www.huffingtonpost.com/2011/08/24/google-settles-phar...

"Investigators snared Google's ad system by creating seven undercover websites offering prescription drugs to be sold without a prescription or the completion of an online medical questionnaire, Martin-Weis said. An undercover investigator informed Google employees creating the advertising for the products that they were manufactured overseas and did not require customers to have a valid prescription, she said. "In each instance, despite this knowledge, Google employees created a full advertising campaign for each of the undercover websites," Martin-Weis said."

Google is not law enforcement and thank god for that. Unlike apple which strives to be the moral leader of the world that nobody needs.

What google did here was illegal and it paid the price for it, but I don't see anybody getting directly harmed by google's decision, it was not google selling the drugs themselves or that the drugs were not upto the standards.

Besides by the same reasoning google should block all the illegal torrent sites, or wikileaks for that matter. Always remember your truth is not everybody else's truth, and it is better that the world's leading search engine does not take sides for you.

So let me get this straight. Protecting your IP with legal patents is evil. Willfully and knowingly getting paid to advertise illegally distributed prescription drugs is fine and dandy.

Clearly Google can do no evil, because everything they do is good in your eyes. Next thing you'll tell me that collecting advertising for child porn, hitmen, or crack is fine too. They don't make moral judgments. Except in China when they're having problems -- then they're suddenly all about morality.

I love all the jerk-wad Google apologists downvoting this. Few companies I can stand less than this self-righteous company and its blind followers.

Laws, being constructed by humans are not infalliable, or otherwise new laws won't replace existing ones. I don't know how willing the google's decision to permit advertising of prescription drugs was, but I am assuming they went with it because there was no precedent for such a thing in the past (the article mentions this as this is the first time a search engine being held responsible). But I hope and am almost certain that google won't do such a thing again. We will see.

Your example of child porn and crack is taking it too far and I am sure no such thing will happen. But if you want google to be enforcing all the laws for you, then maybe it should also detect your location and disallow all merchants catering to gay marriages in the search results, in states where it is prohibited.

Also I don't agree with your generalist approach to corporations that they are all equally evil. I for one would prefer a corporation which makes a bold statement about its policies being non evil and one which is ready to be called upon when it goes astray its path, than one which makes no such claims. To me this is not self righteousness but self discipline.

What you might be seeing is the response of intelligent people who make up their own minds, and who consider Google overall to be much less evil than, oh, I dunno, MicrosoftOracleAppleFacebooketc? At least they have a bold statement that people like us, the media and their own employees will measure them up against, every hour of every day. Whether they always unfailingly measure up to it isn't as important as having it there. Especially for a profit-driven company that has unending opportunities to get just a bit more evil.

Even worse is the Orwellian Google love that's shitting all over this thread. Bill Gates started a charity because charities suck so bad and he wanted his billions to accomplish something. Steve Jobs doesn't even take money for his work (he gets stock options that he often doesn't even bother to extercise), hell, even Larry Ellison signed a thing to donate most of his fortune to charity when he dies.

But, no, it's Google who are the benevolent ones! Why? Well because their motto says so, duh?

Then maybe Steve Jobs should use his salary to make his products affordable to third world countries to do some real good than serving the riches of the world, and maybe Larry Ellison should have honored Sun Microsystem's policies of not suing anybody who uses Java APIs.

The point wasn't that Jobs/Ellison/Gates are great role models. I was showing that everyone that many would consider evil has also shown benevolence. There is nothing that makes the google founders stand out. In fact, I can't really think of anything they've done to make the world a better place besides say they want to. At least these other guys have done something.

I think you're taking the "evil" part a little to seriously. Is Microsoft really "evil" by pressuring you into a tie-in to their products? They're not killing anybody. They're not defrauding anybody. They're creating an annoyance for their customers. That's a far cry from evil, but in half-serious Internet jargon it's "evil".

Whaaaat? Don't you know thew world is a better place if Google gets that $30 per credit card instead of organic search? Or now travel money, G offers, local etc.

the naievety is amusing. what site isn't using some sort of tracker to better serve you ads or market to you better using your information to sell you something.

They specify what they mean by "evil": http://www.aaronsw.com/weblog/googevil

No they want us to use their services because they make more money that way. And sure you have the choice of using competitors but let's be honest, Google has the option of always showing their own services in the #1 position in search results(i.e. search for any mortgage term).

And they also dominate any local keywords. Instead of going to the actual business's website, the first 10 results tend to be Google places.

Google wants to control all information. By being the defacto middle man between all businesses and their customers they are able to carve a large chunk of advertising dollars.

If the results Google gave me were not relevant to what I needed (and that includes linking to third parties over their own excellent properties) then I would use bing, or find something else which accommodates to my needs better. However, this is currently not the case. I can quickly search the local weather without going to any random site, do quick maths or conversions without effort, search journals and books without second thought etc...

It looks like it was built quickly.

Why is there a giant search bar at the top of the screen? I came here to look for flights, not to search the web.

Why is there a giant map at the top of the screen? I already know where I want to go. I don't need to look at a map to decide which destination I want. This part seems like more of a tech demo than anything.

The whole UI looks like it was designed by an engineer. For example when I want to book the flight, I get a tiny left/right arrow box-thing to increase the passenger count. Very poor choice of UI.

> Why is there a giant search bar at the top of the screen?

This will be (if it isn't already) integrated with their search product. In other words, it'll be a part of search. You'll search for "flight from ny to sf" and that's what you'll see. Perfectly natural for them to have a consistent UI in their search product.

> Why is there a giant map at the top of the screen?

It's interactive. Try pulling the marker from one location to another. Very cool. I can also see the exact route that the flight will take. More intuitive.

I may have a general area of where I want to end up, but this lets you choose exactly where through a visual map. Again, very cool.

> I can also see the exact route that the flight will take.

That's not the exact route the flight will take; it's the quickest (great circle) route. Airliners fly along airways.

Let me paraphrase that.

I can also see the approximate route the flight will take, with layovers and other planned stops.

> Try pulling the marker from one location to another.

For me it just drags the map (Chrome 13.0.782.220 / OS X).

> I may have a general area of where I want to end up, but this lets you choose exactly where through a visual map. Again, very cool.

...is a map really helpful for this? How often do you want to go somewhere but not know the name? I'd rather they take all of that map space and use it for something else. I wouldn't mind something that says "hey, you could save $x on this flight if you went to $nearby_airport instead of the one you're on."

> I wouldn't mind something that says "hey, you could save $x on this flight if you went to $nearby_airport instead of the one you're on."

Definitely. They could also just overlay that on the map like they do the popular destinations.

> For me it just drags the map (Chrome 13.0.782.220 / OS X).

I'm on Chrome 15 (dev), which probably explains it.

> is a map really helpful for this?

I like maps. You never know what you'll notice on one.

There's a giant search bar because it's part of Google Search. Like Images, Shopping or any of the other verticals. It's likely under testing, but I would wager at some point soon it will show up automatically in the sidebar as an option when searching for flight related topics.

To see how the map is useful, put in an origin and then play with the price limit.

It's a cool in a geeky sense that destinations drop off the map when I click back and forth. How does this solve my problem as a customer though? I know I need to get from A to B. How does the map help me choose which flight from which carrier on what day?

If you leave the destination open and limit by price, you can answer questions like "Where can I go for $500"?

This is all I've wanted from a flight search product since forever. It will be awesome when it's more international!

http://kayak.com/buzz has had this forever.

Try http://www.kayak.com/explore/ which gives the same max budget parameter but on a map from an origin city.

True. The UI however is something else. Really gets out of the way and gives you what you are looking for. Feels effortless to use compared to the usual suspects in this domain, IMHO.

It's helpful in places like San Francisco or New York, where you have options about which airport you can use.

For example, I looked at a ticket from SEA -> SFO. When I zoomed in on SFO, I can easily see how much it is to San Jose or Oakland as well.

It's also just fun to enter your departure airport and see "where can I fly and for how much?"

The map is misleading when showing result for flights with stops. It should show/highlight intermediary cities in this case

Next to what other people have pointed out, you can also hide the map if you want to by clicking on the marker.

"Just how quickly this was built (really, it was quick)"

Wouldn't this have to be thanks to their relatively recent ITA acquisition? Not too surprised since QPX runs a few popular websites such as Orbitz and Kayak.

(I'm not saying it's not a feat to do it in such little time as any project does take time, but I would like to see credit where due.)

Its' tough for a company to keep impressing customers with everything that comes out of it. That is what google has been doing for a long time now. It really is impressive.

As a developer, I just HOPE that one day google takes a service like this and makes a long blog post out of it describing and bisecting every aspect of the app and how it was built. I'm tired of clicking 'view source' and finding nothing but minimized JS

A tip I found the other day is that in Chrome you can actually de-obfuscate minimized Javascript and throw break points in there if you'd like.

In Chrome 12 you just right click the scripts pane in the developer tools and click "De-obfuscate Source".

In Chrome 13 you navigate to the scripts pane and click the "{}" icon. I honestly have no idea why they got rid of the right click functionality, but it's incredibly useful for me.

Thanks for the tip. I did not know about this.

I think that would create a sort of turtles all the way down situation. Google has so much internal proprietary technology infrastructure that it would be pretty difficult to provide a compelling explanation that doesn't have you wondering about how 10 other google services work (that you just learned about).

Disclaimer: I work for Google and this is how I feel every day :) it's a good feeling though!

Many of the tools Google uses to build their products are open source. Flights source for example mentions GWT.


is there a API or a way to affiliate model to integrate this on our own sites? i could totally use this on my travel related sites/apps :)

I doubt there will be. Flight data is typically well protected by the industry. The only reason Google has access is they acquired ITA who is the main data warehouse.

ITA Software, acquired by Google in the last year, has always had this functionality available at [1]. In fact it is more full featured with ability to search international flight and a number of other options. There are even apps for iphone and android. This functionality pales in comparison to what is available there. I do not see the point of building this rather than have people use what is better and currently available.

[1] http://matrix.itasoftware.com/

How fast is Matrix? How does it scale?

"Just how quickly this was built (really, it was quick); and"

Sorry but not impressed, unless it was done overnight. Nothing special there, just copying from already done sites.

"2. Google wants you to use our services because they're compelling not because we don't give you any other choice (ie "Don't be evil")."

Yeah, that's why they keep filling the first page with in house junk that would NEVER make it without the Google advantage. I am already seeing adsense ads for this all over the web, how much does Google pay the webmasters for them?

It says "Ads" on the booking options. What's the business model?

Just curious about the backend. Python too or Lisp?

Why "Python too"?

Google is well known for using Python.

I'm willing to bet this is using ITA's software, which is written entirely in Common Lisp. See my comment above.

I said nothing about this specific product. I was simply simply answering the parents question about the use of Python at Google.

Using python yes. But for building apps with heavy traffic? Most of the SWE jobs on google.com/jobs have a Java/C/C++ requirement, not Python.

Just curious... how quick it was? (2-days for coding and 2-months for ideation??)

That really is good. Hipmunk is still better if you know what days you want to fly, but this is great if you haven't yet decided precisely when, or precisely where, you want to go.

A pity it only works within the US so far, but I'm sure they'll add international destinations eventually.

Also a shame that Southwest still won't cooperate with any of these guys. I guess it's a rational decision on their part: be cheapest most of the time and hope that people won't bother to compare your fares to others.

Yes, the first site that has JetBlue and SouthWest comparisons with the major airlines will have my traffic. I usually prefer to travel on price when it's not business so I love how Google has made this clear - that's my core problem with Hipmunk and I'm not sure why they don't just tell you the cheapest days to fly, like Kayak at least tries to. I also like how http://skyscanner.com does it.

We (Hipmunk) have Jetblue, and so does Kayak and Google. We also let you figure out the cheapest days to fly: just click "I'm flexible."

Good to hear.

You also have one killer competitive advantage: an adorable logo.

I love your site. It easily has the best UI. I knew in the back of my mind that you have the "flexible" feature, as I've seen it before, but never thought about or even bothered to use it, because it just isn't visible enough. It may help to make it a link on the results page, and perhaps mention it on the loading screen as well.

Thanks for that tip! i've been looking for a service like that, but I never saw that link on hipmunk.

Thanks for a great service.

Maybe you should have "I'm flexible" on the results page, too? Skyscanner does. I would not find it without your hint.

Thanks, I just saw this (new?) feature. Works beautifully.

Google Flights only supports the US at the moment. I see this as a good thing for Skyscanner which covers Europe well, they can adopt and implement new ideas from Google Flight search before Google gets around to other countries.

except that Hipmunk's interface is a clean clone of ITA's matrix search's gantt view (http://matrix.itasoftware.com/) - I'm sure it'll be a matter of time before that makes it into google's interface.

> except that Hipmunk's interface is a clean clone of ITA's matrix search's gantt view

Why clean clone? Gantt charts are (technically) 115 years old, and it's not like there are 5 billion ways to display a gantt chart, especially one with no paths. Other flight search startup have independently stumbled upon the idea of gantt charts for flight, why could that not be yet another independent reinvention?

> I'm sure it'll be a matter of time before that makes it into google's interface.

Or not, ITA does not seem to have met much success with their matrix so far.

They are the same as in yes, it's the same idea - a gantt chart but Hipmunk's way better looking and more interactive.

When I say clone, I don't mean to imply they simply copied it with no originality - I'm just pointing out this fact to all the people who claim Hipmunk's advantage is their gantt view..which ITA already has a versin of and can be improved to match or be better than Hipmunk's.

> When I say clone, I don't mean to imply they simply copied it with no originality

That's what "cloning" implies though: making the exact same thing.

> I'm just pointing out this fact to all the people who claim Hipmunk's advantage is their gantt view..which ITA already has a versin of and can be improved to match or be better than Hipmunk's.

But have not so far, so Hipmunk's remains an advantage. You "could always" copy to match, in theory. Does not mean it's technically feasible, let alone easy. Or that you think it's a good idea.

And yet so far no one has.

Not bad, but hipmunk's visualization seems far far better than Google's tables. The graphical representation of the flight path seems nice, as does the highlighting of cities involved, but that doesn't actually give me information I need to make a decision; Hipmunk's time-oriented chart of departures, arrivals, and layovers tells me exactly what I need to know to book a flight.

Exactly. This would intimidate my aunt; Hipmunk would amaze her.

I hope she'd like the chipmunk, too!

If your startup is involved with search aimed at the consumer market, watch out...Google is coming.

In a few weeks/months Google will be featuring this search result when you type "nyc to sfo" and take a big bite out of orbitz, kayak, and whoever's lunch that's in this space.

EDIT: The counterargument to this, is that orbitz, kayak, and friends are some of the heaviest purchasers of PPC ads. So Google be at risk of cannibalizing if they push this too hard.

The counterargument to your counterargument is that google could start taking over what they cannibalize and make more money directly.

Exactly, it's simple economics: If Orbitz et al. get $5 per Airline ticket sold, they can only afford to pay Google up to $4.99 for the click.

Google can take the $5 from the Airlines and laugh all the way to the bank.

The issue is that in order to do that google would have to only list flights from airlines that are willing to pay them $5.

If the airline refuses to pay and google doesn't display their flights then their results are incomplete.

The airlines will probably understand that google wants to be able to list their flights and will negotiate harder because of it.

This is a dangerous game however; the flip-side is that the airline has more potential profit to lose by not having google to sell their flights for them.

> If your startup is involved with search aimed at the consumer market, watch out...Google is coming.

Possibly... However sites like hipmunk will be used as a comparison in so many press articles and blog posts when google starts pushing this.

Many start ups have won against google. Google Notepad & Evernote, Google Wave & Twitter, Google Ride Finder & Uber etc...

.. but I still wouldn't want to be in this space right now.

Not entirely sure how you're equating Google Wave and Twitter.

I think Jaiku is the more apt comparison, for what it's worth.

Those are all ancillary to google's core product: search.

The problem with Google is that they don't seem interested in doing the hard work to go from 90% of the way there to a full product. It's like that with nearly everything they do. Often, it shows up as a US-only policy (flight search currently fits that bill). Other times, it's just random things. There are no Apps accounts for Google+. There's no "Tablet Optimized" section of the Android Market outside the US (and of course no books or movies either, but that's not their fault necessarily).

Most people aren't going to be interested in replacing something that works with something that sometimes works. Google might have a great flight search engine, but if I have to use Orbitz 30% of the time, I'll just use Orbitz. I already have an account set up there, and I can actually buy the ticket instead of having to find the flight, go to the airline web site, and search the flight again to buy it.

From playing around a bit, it seems like Kayak still has Google Flights beat on pricing by as much as 150 dollars. A quick comparison of a bunch of flights and I found that Kayak almost always had the lower price. I'm gonna have to agree with you on the 90% thing, I don't really see why Google would release something like this and not take pricing into consideration. I guess that will be fixed soon though.

Well, it was a matter of time, wasn't it? For now it's a bit too fiddly when compared to the likes of Hipmunk, save for the booking process which is quite streamlined. And the search itself, which is blazing fast. On second thought, this is really not a bad effort. I can see it gaining huge traction in no time.

The (well-hidden) "Limits" widget is a very cool bit of visual / visceral UI design. Click the "scatter graph" button next to the Duration field and you get a scatter plot of duration against price for all the flights - which is useful data in itself - then you can drag a boundary around to set the maximum duration or price.

You know what would my make flights better (I travel a lot for work - but work out of my house and book a lot of my own travel)

I want to be able to setup standard trips/itineraries and be able to single clik re-book them with simply a leave and return date.

Further, I want the system to auto arrange for a cab/shuttle/uber to pick me up and take me to my destination.

For example, I fly to LAX several times a month - Nome Alaska once a month, Soon it will be Dallas once a month - and various places in the bay area.

If I can setup my "Visit LAX trip" with all my details and know that a car was waiting for me when I got to LAX to take me to my office/hotel and I didnt have to do anything other than click "Re-book LAX" that would save me so much time and hassle.

I would setup my preferred airlines, times and seats (Virgin America etc...)

May I ask what you do, and how I might get into doing it? I am craving a job with > 0 travel...

I am a healthcare consultant. I consult on new hospital builds and ensuring the organizations are prepared to get licensed and then move in.

To me, that map seems like a very confusing bit of UI. The cities themselves are really tiny interaction points, I'm a relatively experienced FPS gamer and hitting those points accurately involves more fumbling than it should. You cannot drag departure/arrival pins. Streetview kind of taught me that this should be a mechanism for manipulating this sort of location reference.

The big one, though, is that layovers are not reflected on the map. Granted, it may push me slightly towards more expensive flights, but I would appreciate the ability to see the grid of ugliness and waiting I'm buying into to get that super cheap ticket.

I don't think that map as it is now is worth the screen real estate.

The whole thing has "designed by an engineer" all over it. From hidden UI that reveals itself by clicking/hovering on various elements, to the entire process being limited to one screen which various pop-out elements and transformations. There is little consistency in the design, and poor or no usage of contrast to separate information visually.

When clicking the large blue button to the right of a flight, several rows of return flights drop down below that. The only visual distinction between these specific sub-rows and the other departing flights below is a thin blue/gray border on the sides.

Going from A to B is called an "Outbound flight", while going from B to A is called a "Return flight". Copy is very important in UI design. It should be outbound/inbound or depart/return, not some mix of different terminology.

Durations are always "-- Xh -->". What is the point of the arrow if it is always pointing to the right? Why is the duration a lighter shade than the rest of the text?

The left column is a Gawker/Gizmodo style static column that locks in place when scrolling down. The Google search results page doesn't do that.

These are basics, and a company the size of Google with as many employees as it has should be able to get these things right.

Outbound/return is standard terminology all over the travel industry. Including in UK and US, and for trains and for planes: I'm surprised you're not familiar with it.

"Depart" means something different: it refers to the beginning of each leg of the journey. Its opposite is "arrive". So you might have an outbound leg which departs from LAX and arrives in SFO, and then a return leg which departs from SFO and arrives in LAX.

Hipmunk.com says depart/return and Iflyswa.com says departing/returning. Your claim is 0/2 so far.

I went to go and prove you wrong (I personally use outbound/return).

But webjet (Departing/Returning) and skyscanner (Depart/Return) seem to back you up.

"Sorry, locations outside the U.S. are currently not supported." Are you kidding me?

I also don't see any way to search for non-economy-class flights...

as usual

First iteration.

I have never even heard of a flight search not supporting international flights.

Release early. Release often. Minimum viable product. Any of these ring a bell?

Come on, that would be like releasing an MP3 player in 2001 that had no wireless and less space than a Nomad.

Bad memories are now flooding through my head of silly complaints to various products over the years (a few of which I'm quite guilty of making).


Well, now you have! It's a much simpler problem to solve first.

And now you've actually used one.

For some values of "used." I used it to search for "toronto" then Ctrl-W'd.

You should have tried ITA Software's Matrix site in 2000. It was already awesome, but it was US only.

It's an obvious next step then. Their hotel search tool is US only as well: http://www.google.com/hotelfinder

I guess when they are combined into /travel they will be internationalized.

Just for some context, here's the previous HN discussion when Google announced their acquisition of ITA in July 2010.


It's quite scary that a company like Google can come along and shit in a startup's cereal practically overnight.

I know we were all rooting for The 'Munk, but it seems that Google's use of Price x Duration matrix effectively steals the thunder from Hipmunk's agony filter.

I mean, it's not as if Google needs the affiliate revenue, and I bet the Hipmunk guys would have preferred it had Google decided to "organise the world's information" through an acquisition. I think I'd need a good cry if this happened to me :(

It wasn't overnight. ITA has been doing flight search since forever. In fact, they even had the matrix UI before Hipmunk (though it wasn't featured as prominently):


No tears over here! We've known this was coming for almost a year—Google announced their ITA acquisition before we even launched. We've also been preparing some sweet responses :)

Attaboy! Still rooting for you guys. You should add gigaballs of steel to the mascot.

Nice. You have a lot of people on your side.

General question here. Almost all flight search sites default the search dates to about 3 weeks out. Now, I find this annoying but I'm wondering if its done because there is statistical evidence that most customers search in that range of time, or rather if it's just what someone thought would be a good idea by "following their intuition".

Well this is some interesting research. Out of my off-hand and roughly scientific research, my statement of "almost all" was quite off.

I found that all of the major carrier websites I could remember off the top of my head did not employ the date-ahead practice.

Delta - no US Airways - no Southwest - no Virgin America - no Continental - no

I followed that by a Google search for "flight search" and checked the sites that came up in roughly the 1st and 2nd pages. Where I put "yes" and the date range, this date range was pre-populated.

Based on Google for "flight search" Priceline - no Travelocity - no Kayak - yes 9/27/11-10/4/11 Expedia - no Cheapflights - no Flightsearch.com - yes 9/27/11-10/4/11 Tripadvisor - yes 9/27/11-10/4/11 Farebuzz.com - yes 9/27/11-10/4/11 http://www.skyscanner.com/ - yes 9/27/11-10/4/11

Finally, I checked YC's own Hipmunk, and they do not do it either.

Hipmunk - no

So my original statement was at least partly false. It doesn't seem like "almost all" after all. And the ones that I could find doing the date-ahead practice all had identical dates. I suspect all of them use the same backend search product.

I don't care enough at this point anymore to follow up on this anymore, so I'll leave it at that.

I cannot vouch for the other sites, but at Skyscanner we looked at the aggregated search history (12 million users a month) to make a best guess at the default dates to set.Interestingly the defaults we use are different for the mobile platforms. (We use our own backend technology)

Can you provide a little more detail on this? What percentage of users actually use the defaults (for both start and end date), and how many change it anyways?

Have you thought about doing some A/B testing on providing a default and not?

The rule of thumb that I've heard was that anything about a month out represented the 'normal' price.

If you're buying a ticket for a flight less than a month away, the prices start to rise dramatically as the flight date draws nearer. But if you're buying a ticket for a flight more than a month away, it's more-or-less the same price as a flight that's a month away. [1]

So absent any other information, the 'normal' price can help people ballpark costs more quickly, something I'd expect to be statistically more common than actually searching to book a specific flight on specific dates.

[1] With exceptions for destinations with very defined/popular seasons, holidays and things like that.

It seems like the default search is just for people who want to see "about how much is a ticket between X and Y?" Anyone doing a real search has real dates. By placing the time frame a few weeks out, you avoid including very expensive last minute tickets, and do a better job getting general flight prices.

It depends on the kind of traveler (business travelers have shorter lead-time, leisure travelers have longer), but 3 weeks is a reasonable guess—probably on the shorter end—for what Google customers will be looking for.

I guess Google will be able to fill this automatically by checking your Google calendar for recently added events in other cities. :)

Bad news for hipmunk? Especially since google acquired ITA.

Hipmunk's key differentiator is design. It's a pain in the ass to book flights and hotels. Hipmunk removed that pain by making it fun.

Google's flight search doesn't look fun.

The really bad news for hipmunk will be when google integrates it into their search results.

Well, Adam has said in the past that only a small amount of Hipmunk's traffic comes from Google search results, so they don't really have that far to fall there.

On the other hand, users are fickle and traffic can be a bit butterfly-effecty, so it's hard to predict the impact of something like this. That said, I don't expect Hipmunk to panic or take this lying down.

So who are you rooting for? In the hypothetical world where there is a winner and a loser (which travel search pretty much is, given its maturity), are you rooting for the home team or for the guys you worked with for so long?

Well, I work at Google and some of my best friends are at Hipmunk, so the ideal outcome would be for Google to buy Hipmunk. Or vice versa.

>Or vice versa

I like you.

What do you mean by "integrates it into their search results"?

Hipmunk's killer feature is the ability to easily visualize the flight in terms of the time you leave, the time you arrive and the time spent at various intermediate locations. It becomes more useful when there are no nonstop flights. Google is taking a different approach with their UI, which is fine.

In terms of whether I'd use google or hipmunk to search for flights in the future, I'd say that it depends on what my priorities are for that particular flight.

by "integrate into their search results" i mean.. if you search google for "sfo lax" and it brins up this new Google Flight page with their organic results.

Ah, I see. Sorry, the ambiguous "it" made me think "it" was hipmunk. Somehow.

The bad news for hipmunk is it wasn't that much of an improvement over actual travel sites. I never had that much trouble using Expedia and I already know how to purchase my ticket right on the site. So there wasn't much reason to use Hipmunk in the first place.

This Google search interface is pretty nice, and makes a lot more sense than Hipmunk did.

Really? Google gives you a sort-able grid which doesn't seem that helpful to me.

Hipmunk gives me a "heat map" style presentation which I can understand multiple factors at once without scanning text:

Color = carrier / layovers

Line length = total travel duration

Timeline position = flight start/end time

Ordered by Agony (or whatever)

I can take one quick look and grasp all the flights without a second thought. In my opinion, Hipmunk's UI is genius where Google's looks like a school project.

Game over.

Disclaimer: This is a throwaway account. I'm a person who has had a travel startup. I've decided to halt, once Google has finally acquired ITA Software.

Travel is one of the world's largest industries (~5th), and online sales are its significant part (~10-30%, depending on the market), growing strongly. Online travel agencies (OTA) are among the few companies on web that get real money (~$150 per sale) from customers (it is relatively easy to make serious revenue).

Online travel sales consist mostly of flight bookings, and hotel reservations.

On these markets, there are roughly three categories of players. Airlines and hotels _provide_ the inventory, that is flights and hotel rooms. Computer reservations systems (GDS) _manage_ the inventory. Online travel agencies _sell_ the inventory.

Specifically, there are thousands of airlines and hotels (e.g. Hyatt, Lufthansa). However, there are only three major GDS operators in the US and Europe (i.e. Amadeus, Sabre and Travelport), as well as only few big OTAs (i.e. Expedia, Travelocity, Orbitz, Priceline, Hotels.com, and Opodo).

Few decades ago, before Internet, airlines and hotels were unable to sell inventory on their own. It definitely takes more time to set up an office than a connection between two airports. Thus, it made sense to use travel agencies for this purpose.

Over time, airlines and hotels became also unable to manage their inventory. Synchronization of reservations between thousands of third parties is a non-trivial task, not a core competency of involved companies. Thus, it made sense to use a middle man.

That's what GDS systems do. They manage inventory, what includes reservations (PNR), its availability, prices, and exchanging data with others. As far as I know, right now, airlines and hotels pay them for the service, and for each reservation made (~few bucks).

Internet has complicated things a bit. Travel agencies are no longer so vital, as both airlines and hotels are able (at least, they think so) to sell the inventory on their own. Unfortunately, they were too slow again, and OTAs has emerged meanwhile (Expedia), giving the second youth to GDS companies.

The ecosystem is like an old marriage couple, although a threesome. Each party hat hates each other, but there's no other way around. To oversimplify a bit, OTAs have _customers_ (traffic), GDS systems _manage_ the inventory (with airlines, hotels, and OTAs) and both airlines and hotels _provide_ the inventory, after all.

From time to time there's an affair. Low-cost airlines try to distrupt the market by selling tickets so cheaply partially because they sell them directly. Major OTAs, like Expedia, partially grow to a GDS category. Some airlines or hotel chains withdraw from GDS systems, and return eventually with negiotiated better fees. However, it's mostly business as usual.

Today, if you want to start an online travel agency you have to speak with a GDS company. Depending on your market, it might be Amadeus, Sabre or Travelport. After a long selling process, you get access to the system, and you can start selling the reservations.

What's important, though, nearly all systems used today were created a decade, or two, ago. As core competency of GDS companies is in selling, then, as far as I know, they outsource the software evelopment to third parties, and it's not that easy to innovate on a critical part of the world's infrastructure.

What you end up with, then, is an access to an undocumented API that lets you to search, and manage your reservations. Insiders are used to the quirks, like waiting few seconds until you get the response, random issues, or hinting the system so you get a better response than others. Importantly, you're actively discoured to cache the data, as the prices change rather frequently.

The critical part here is search. It's a mathematically non-trivial problem to very quickly find rates within thousands of connections, definitely beyond technical know-how of GDS operators, airlines and hotels. ITA Software has managed to get access to the inventory and while, as far as I know, they do not sell resevations, they've created a much better (faster) search.

Meanwhile, few years ago, metasearches (e.g. Kayak, Hipmunk) emerged. Smart folks have realized that the competition is on price, customers look for a single place to compare prices, and operate under assumption that what really counts is traffic. From both customers and metasearch perspective, it does not make that big difference where do they buy the reservation from, an OTA or directly from an airline.

So, here we are today.

As a beggining travel agency, you likely have to pay annually for access, and for each request made, especially if you exceed the quota negotiated with the GDS company. Few years ago you were able to make profit by incurring a transaction fee to each ticket sold, but now transaction fees are nearly non-existent, and it's more frequent to rely on provisions from GDS companies, and, sometimes, airlines.

What's your competetive advantage? Basically, you cannot provide much better product than your competitors, as everyone relies on the similar legacy GDS system that returns the flight details rather slowly. Most of the time, only choices are either to show the results a bit differently, or bet on more trustful brand.

The focus is on efficiency. Profit per ticket is so slim, so cost of customers acquisitions is what really matters. OTAs, metasearches and, increasingly, airlines, together with hotels, master SEM, SEO, and other forms of advertisement (newsletters, banners). They live and die by the numbers. If you've figured out how to scalably make $1 more profit on each reservation made, you're covered for some time. The novel methods are, obviously, eventually realized by others, too.

In this race to the bottom there's one clear winner. Google's AdWords is a major source of traffic for all parties, and I bet that they already make the biggest profit off each reservation made. Once Google has acquired ITA Software, they now have both traffic and the inventory.

This is one of the reasons I hate Google. I love to see normal people like me go make a startup and become successful. Escaping their corporate bonds. With companies like Google around who can use ad income to supplement dabbling in all sorts of things they can really weaken that option. And what's Google going to do with this? Add a couple percent on their revenue sheets? Great.

And make my and million's of other live's much better. It takes a mammoth to disrupt such a consolidated (and crappy) industry.

It does not take a mammoth anymore. A small group with a big idea can change the world these days. We don't need Google.

Actually yes and no.

True you dont need Google. We would all love their traffic but having played with the tool for the past 24 hours. I have noticed the changes in inventory state and they are doing exactly what we could have expected them to do.

Sadly that is also the downside - they did exactly what we expected them to do. A mediocre job that is fast and of course "good enough".

A small team can (and has) replicated what Google has done. In fact there are many of them. But Google does not provide Trust and in my reasonably jaundiced view - this is another rearranging of the deck chairs.

The issues are now out in the open. The core infrastructure of search is REALLY hard if you work on the same premise as Google. If you change the basics then there is a different answer.

Will the basics change? I believe the answer has to be yes because otherwise the original premise of the first poster must logically be correct.

I believe not. I am working specifically to change that - if you are interested - then ping me.

My First Impression:

No international flights yet: Fail but not surprising

That's okay I want to fly from NY to Vegas next year. No flights for April 2012: Fail that was not expected.

I don't really understand why Google launches stuff like this when it has a lovely user interface but is a bit halfbaked in terms of the data that they've put in.

I guess I'll go back in a years time when they remind me about it.

Many airlines don't allow you to book tickets more than 6 months in advance at their own websites. I don't think it's surprising that Google didn't prioritize 6+ month advance ticket purchases.

We have detected that you are using an unsupported web browser. We support Firefox 3.5 or later, Chrome, Safari 4 or later, or Internet Explorer 8 or later.

?!? I'm using the Android browser!

If I were expedia, orbitz, priceline, travelocity, kayak, hipmunk. etc. I'd be very very afraid. Sure, google won't actually do the sales, but they are linking direct to airline websites for now. Also this doesn't exist for hotels. Yet. The odds that in 1 year that any of those sites have a better UI or superior search capabilities than Google is low. So, what will they bring to the table?

Also, what's up with all these people saying that it's so limited because it doesn't do international and such. It seems very obvious that this thing will get better and better and like others have said, eventually end up on top of the search page. This is assuming it gains traction quickly and doesn't get nuked.

Honestly, the lack of a good API by any provider is one of the biggest barriers to entry in this space. I wonder if the Hipmunk guys might release one? I'd be thrilled to send affiliate traffic to the first company that makes good on this.

Unless I'm missing something fairly obvious, they might have missed something pretty big - is there any way to book a one way flight?

Agreed -- couldn't find it in the first 30 seconds, which is sort of a red flag...

what will be a game changing tool, and something befitting the resources of google, would be to pair flights together that are cheaper than what is offered by the airline. for instance: flying from NY to LA. Airlines sell this route non stop, or through one of their hubs. But imagine if a flight search can figure out that a oneway flight from ny to new orleans on Delta and than a flight from new orleans on AA, had a low layover and was significantly cheaper than the published routes.

The problem is that as this expands to more cities and takes more stops, you end up hitting an NP problem.

It's unlikely that the Delta fare from NYC to MSY and the American fare from MSY to LAX permit end-on-end combination with each other, so it would have to be issued as two separate tickets. What happens if your Delta flight to New Orleans is delayed and you miss your connecting flight to Los Angeles on American? Neither carrier is going to offer you any compensation, and your travel insurance probably has a clause which excludes misconnections caused by an airline's late arrival.

There are occasions when this strategy produces good results, however: last year I needed a last-minute one way flight from a small southeastern city (TYS / Knoxville) to San Francisco, and the fares were extremely high ($800+). But flights to Orlando were very cheap due to competition from low-cost leisure carriers, and the last-minute fares from Orlando (MCO) to San Francisco were quite reasonable. I booked, on a single ticket and flying on a single airline the entire way, TYS-IAD-MCO-IAD-DEN-SFO. This gave me a TYS-MCO fare plus an MCO-SFO fare, which was hundreds of dollars cheaper than the more direct TYS-DEN-SFO. At the time, saving money was a lot more important than getting to San Francisco a couple of hours earlier. A cunning plan that cannot fail!

Except that I missed my first flight of the morning due to an overnight road closure that blocked my only way of leaving the house. By the time that I got to the airport, my originally booked flight had already left and I was faced with a very long wait for the next flight to IAD, which virtually guaranteed I wouldn't even make it to San Francisco that day. It was only through a combination of charisma and blind luck that I managed to convince the airline to reroute me direct through Denver, which is completely against company policy. They only charged me the $50 same-day standby fee.

Needless to say, I would never sell that sort of a ticket to the general public. Far too many different ways in which things could go horribly wrong. It's also likely that the airlines would promptly revoke your ticketing privileges with them were they to discover you were routinely issuing these sorts of tickets.

perfectly fair, i admit i've done some flight "hacking" myself. The reason i bring it up is because it is well know that airlines prices their flights between city pairs, not based on the legs. For instance a flight, between jfk and lax through chicago could cost more than the sum of those two same exact flights. This would be an opportunity to build your own ticket. I admit there are significant downsides with delays and reroutings, but if google presented the information to the public it would add transparency to a very opaque system and it may lead to a change throughout the industry in the way tickets are priced. After all google's mission statement is to organize the worlds information

The market for building complicated routings is very small -- typically only mile/segment runners. ITA's backend is already used by that community for this, but not sure why the public would be interested. Each segment adds additional landing fees and taxes, making it more expensive, unless you're exploiting a rare hub/partner mistake fare.

Looks nice and it has great potential, especially since it isn't cluttered with links to Expedia/Priceline like the others are.

That said, I can't really use it until it supports multiple destinations/open jaw in a single search.

This is just beautiful. Data matrix is innovative, UI is out of the way, slick and to the point interface, and it even takes you to the correct 3rd party booking pages.

I can't wait till this is implemented in Canada.

Kayak beware!

> Kayak beware!

This definitely looks better than Kayak. ^_^

I love the scatter plot graph (with duration/price), it is so geeky :)

As is evident from some of the comments here, the map and the search bar on top are totally extraneous to the basic task that the user wishes to perform.

The absolute barebones should be shown: Starting location, Destination, Dates. And then maybe a less noisy version of the price-points plot from which one can just drag and drop certain options into a bucket for comparison.

The fact that so much scrolling needs to be done to even get a basic understanding of the results means this needs to be reworked.

I try to get a flight from Frankfurt - the 9th busiest airport in the world. - and I get "Unavailable".

Post this again when I can actually use it.

When you search for a destination airport outside of the US, you get a very clear message:

"Sorry, locations outside the U.S. are currently not supported."

They're taking the release-early-iterate-often approach to this new service, which means that they're not going to make everyone happy immediately.

Rather than dismiss this new service just because it doesn't fit your specific use case right now, think about this from a different (i.e. hacker) perspective -- this is a cool new service that will obviously become even cooler as it gets more advanced.

I agree totally...just pointing out that the site is useless to us living outside of the USA. Hard to get excited about something that is, from a European resident's perspective, useless.

Decent initial release, but I wish the search was more flexible.

My common use case is "I will spend $X to go anywhere for under Y days any time in the future" and I haven't found a service that makes this easy.

Kayak's "explore" page is the closest I've found, but they rarely have the cheapest flights listed and have no way to set duration. Has anyone found a better solution?

These guys almost fulfil that requirement - I use their service pretty regularly. http://adioso.com/

Is there anyway to do one-way flights? I really like it so far but cannot for the life of me figure out if one-way is possible.

I can't find it either. I would have actually used it today since I found it right as I was about to search for a flight on Kayak... but w/o one-way I can't use it.

I don't understand why anyone would use a federated flight search when looking for a domestic flight in the US. None of the major airlines come close to Southwest in terms of price or hassle free travel and none of the flight search engines include price information for Southwest.

Perhaps for the still-common case of flying to an airport where Southwest doesn't go? Southwest gets to be cheap in significant part by only flying routes with enough demand to fill a 737.

And on rare occasions another airline does beat Southwest's price. Sometimes I see flights in and out of JFK turn up cheaper on Delta.

When are people going to stop solving this problem and start building super sonic airplanes like the concord? The airline industry moves backwards in technology and trying to build a flight search engine to wrap around it is really a band-aid to the real situation. If I could get to Vegas in half the time, or even a quarter of the time (currently takes 4 hours, could take 1 in supersonic jet) I would pay at least double price for a ticket, and I wouldnt even need a seat.

My point is if you fly coach especially with Delta they treat you like a slave and stuff you with almost zero space to move (I am american, but not obese (5'9, 200lb). Forget flight search engines I can find a flight, make a me a faster more tolerable flight, you are friggin GOOGLE!

Drag (and hence power required) is roughly proportional to the square of the velocity. You want to go twice as fast? You better be willing to pay (at least) 4x the ticket price. And since most people want cheaper tickets over flight time, you'll have to pay even more to compensate for the lack of economies of scale.

Computer efficiency still is rapidly moving up. Airplane efficiency is near its limits.

This had to happen sooner than later after the ITA aquisition. I wonder why it took them so long.

Is it me, or that app is super fast?

Yes, you're looking at ITA's InstaSearch product, which delivers near real-time results, compared to the traditional QPX product.

There were many articles about Google's acquisition of ITA, and various stipulations around InstaSearch:

"As a condition of the deal, Google must continue to license ITA's QPX software for airfare websites on "fair, reasonable, and non-discriminatory terms" through 2016. And it must commit to continued investment in ITA's QPX software and in ITA's unreleased InstaSearch product, which aims to provide near instant airline fare quotes."


Disclosure: I worked for ITA until rumors of the acquisition.

It probably is caching data from other people's searches, similar to what Kayak does. This is cool right up until you try to actually book a flight and it sells out or goes up in price dramatically.

Wait, is this actually the case with the Google product?

I was under the impression that Amtrak from Newark Airport to 30th St Station in Philadelphia was available as an airline code share (and that train segment is fast enough that there's no good reason to get on a plane for Newark<->Philadelphia, even if you are connecting to another flight out of Newark), but I haven't figured out how to find such a code share with Google's flight search tools.

(The governor of IL and the Chicago mayor also asked Amtrak to study how to extend some of the Amtrak routes to O'Hare, presumably to replace some ``commuter'' flights, and they wanted the study done by the end of this summer, but I haven't seen any evidence of the study being done yet.)

It's only a problem for hipmunk as long as google doesn't 'retire' the project.

Give it 6 months or so.

We built this at Yahoo! in summer 2007 after the FareChase acquisition. Took about a week or so to prototype, and a couple months to offer various filter/search criteria and prepare for launch. With access the good flight data APIs it's a pretty simple app. Of course for us it was never allowed to launch.

Glad to see Google was finally able to push something like this. Main feature missing that Y! Faremaps had is the ability to specify a span of time in which you wanted to travel and the cheapest trips in that timeframe were shown. Also, you could search "weekends only" in that timeframe.

In true Google style, they are focusing on speed, with a minimal UI.

Unexpected was the minimal data part... They have a short look-ahead and don't work with any international destinations.

The folks who are most interested in this app are the ones who are most likely interested in the above two points.

When I want to do a flight search, I don't usually think of a map.

I wonder if they allowed ITA to keep writing in Common Lisp...

"Knowledge of LISP" is still showing up on job postings:


I work for Dohop (www.dohop.com), a direct competitor of Google's new flight thing. We have been worried about Google's entrance into the field for a while, but after today we are breathing easier.

No international? No one-way? I know Google will change this, but why put out such a wildly underwhelming product?

And finally, since they are basing the whole thing on ITA anyway, we don't expect them to do anything Kayak isn't already doing.

Wow ... but then I am not really sure it meets my needs. For one thing, it does not recognize may local airport and keeps wanting me to fly out of another airport a good 4 hour drive away! And while it is good at displaying cost differentials for different destinations it does not seem to do as well for the same destination at different times with different carriers. Still, it is interesting.

My eyes dilated when using the auto price feature. Also the speed and ease of adjusting the dates by a day is impressive.

Please build this in Europe, searching for flights is so painful. I hope you're looking at this web site Expedia!

This makes a ton of sense to me. Users already trust Google with search and connecting people with airlines has baked-in revenue. This also has tremendous advertising potential. Imagine if local restaurants could target ads at people who will be traveling to the area in the next three months.

They must be accessing prior flight search history in some way or they made a really lucky guess with their default selections.

I've recently searched for flights on Expedia and Southwest (I don't think I've visited anywhere else). Google already had those same dates and cities selected by default.

I think they will integrate this with google maps and give around the world directions. Pretty cool.


Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact