Slashdot is powered by your submissions, so send in your scoop

 



Forgot your password?
typodupeerror
×
Google The Internet Businesses The Almighty Buck

False Ad Clicks Cost Google 1 Billion Dollars A Year 233

Meshach writes "There is an interesting story at CBC which claims that Google loses one billion dollars per year to fraudulent ad clicks. The article contains an interesting description how how the company determines if a click is false. 'The company explained that it determines which clicks are invalid through a three-stage system. Most of the illegitimate clicks are automatically detected analyzed and filtered out in the first stage ... The second part uses automatic and manual analysis of the AdSense network to weed out false clicks before they are logged to an advertiser's account.'"
This discussion has been archived. No new comments can be posted.

False Ad Clicks Cost Google 1 Billion Dollars A Year

Comments Filter:
  • Upside down logic (Score:5, Insightful)

    by Anonymous Coward on Tuesday September 18, 2007 @06:41PM (#20661215)
    It's more like $1B dollars in fraud is not passed on to the advertiser. Many billions more probably are. Google isn't losing a thing.
  • by JoelKatz ( 46478 ) on Tuesday September 18, 2007 @06:41PM (#20661221)
    The logic behind this story is bogus. The $1 billion in money that these fraudulent clicks cost Google doesn't exist. If not for the bogus clicks, these clicks wouldn't exist.

    It's like a software company claiming that false orders cost them $10 billion dollars last year because they received an bogus order for 100,000,000 copies of a $100 product. Had they not received the bogus order, they would not be $10 billion richer.

    Duh.
  • Re:Ledgerlines (Score:5, Insightful)

    by moderatorrater ( 1095745 ) on Tuesday September 18, 2007 @06:42PM (#20661231)
    Assuming that each account has $99.99, and that Google can get a 10% interest rate with that money (they may have to pay out on it, after all, so they've got to keep it close to liquid), it would require 100 million such accounts. Somehow, I'm doubting that they make that much off of it.
  • Ads (Score:5, Insightful)

    by Threni ( 635302 ) on Tuesday September 18, 2007 @06:43PM (#20661241)
    I've been on the net for about 11 years now, and I've not one single time ever deliberately clicked on an ad because it was interesting. I've clicked on accident; I've clicked to allow a download to proceed, or to get a limited time pass to an otherwise charged-for service/site, and I've clicked just for a laugh to fool people into thinking I give a shit, but the day I start to get interested in and buy products based on commercials (online or elsewhere) instead of reading reviews, comparing alternatives and talking to friends/family who've bought something is the day you can take my brain out and give it to someone else.

  • by MacDork ( 560499 ) on Tuesday September 18, 2007 @06:48PM (#20661313) Journal
    Click fraud is only inflating the number of clicks made by $1B. Roughly 10% according to the article. Sounds suspiciously low to me. Spam certainly comprises greater than 10% of all email sent, why are click fraud rates so low?
  • Re:Ledgerlines (Score:5, Insightful)

    by ajs ( 35943 ) <{ajs} {at} {ajs.com}> on Tuesday September 18, 2007 @06:55PM (#20661403) Homepage Journal
    I'm often confused by why people think that there's some mystery cash floating around in Google's (or any other online advertiser's) pocket. There's an obvious need, when payments are tiny, to limit the frequency of transactions so that aggregation can happen. However, it's not like your Google AdSense account is a money market account with cash sitting in it, gathering interest. Google simply has a line-item in their budget for payables that cannot be issued yet because the transaction fees on cutting someone a check for $0.03 cost them more than the payment itself.
  • by ILuvRamen ( 1026668 ) on Tuesday September 18, 2007 @07:01PM (#20661465)
    My web host has a nearly perfect system. If someone clicks on the "hey, buy a package from my hosting company" link a cookie is generated. If he buys a web package from that computer before that cookie expires in 60 days or whatever, I get $100. That's right $100 in cash, not hosting credit, an actual check for $100. Yeah some people have cookies disabled but there's other methods that would work better. So basically that purchase based system removes 100% of false clicks and still leaves in like 99% of valid customers. Seems fair to me and the people who let the ads on their website stand to make a way bigger amount of money if they bring valid customers cuz it's basically a large commission on sales. The only problem is if the ad company keeps serving up crappy ads for products nobody would ever buy you don't make any money but you wouldn't make very much on that type of ad through clicks either.
  • Re:Ads (Score:5, Insightful)

    by sampson7 ( 536545 ) on Tuesday September 18, 2007 @07:01PM (#20661469)
    This is cruel.

    My family runs a small business, http://www.beadstore.com./ [www.beadstore.com] We are not Apple or Microsoft. We do not gross anywhere near a million dollars a year. Each time you click on one of our ads you take somewhere between 5 cents to $1.00 directly out of our pocket.

    Now we try to target our ads only to those who care about beads and jewelry and such -- but our ads sometimes display for completely random searches.

    What on earth is possibly wrong with buying something off the internet? A Google search for "African King Beads" (including the quotes) and my store is the first hit.

    I also happen to know many of the merchants listed on the right advertising for those key words; after all, the high end collectible bead family is relatively small. I would *never* click on one of their ads, because I know it costs them money every time I do. If I wouldn't do this to my competitors, why would you do it to a random stranger?
  • by p0tat03 ( 985078 ) on Tuesday September 18, 2007 @07:02PM (#20661477)

    A cost-effective system might not be able to detect them, but there are ways to get a good idea about where this number may lie. 100% human-powered filtering of random clicksets? That would catch almost all fraudulent clicks, but be horrifyingly expensive to implement across the board, but by comparing results from a random human test to the machine system you can get some idea of how many you miss in the automated method.

    This is also how we figure out QA numbers in manufacturing - not every bad device that goes out will be reported for warranty or otherwise exposed, so there are alternate ways to derive a good quality index out of mfg'ed goods.

  • Re:Ads (Score:3, Insightful)

    by swv3752 ( 187722 ) <[moc.liamtoh] [ta] [2573vws]> on Tuesday September 18, 2007 @07:05PM (#20661521) Homepage Journal
    I have clicked on adds to buy a product, but it was for products I was looking to buy anyways. I figure i can help support a site without costing me anything extra.
  • by Anonymous Coward on Tuesday September 18, 2007 @07:10PM (#20661575)
    But fake clicks that pass through the system are paid out by the advertiser, not Google.

    Google does not charge its advertisers for clicks it determines to be invalid. For example, if 10 out of 100 clicks were excluded Google would not charge its advertisers for the invalid clicks, cutting into the company's revenue.

    So what they are saying is that if it is an invalid click, they don't charge the AdWord owner. Right, that makes sense. Because of this they somehow lose money. Why, because they still pay the person who is displaying the advertisement? Nope, that's not the case. Infact, the persons AdSense account where the fraudulent clicks are coming from is banned.

    So is it the bandwidth and processing power that incurs this hefty bill? I doubt. I can understand that the manual analysis they talk about in TFA could add to the bill, but 1 billion dollars.

    To me, acting much like Dr. Evil, they decided to say 1 billion dollars because it sounded good.
  • by karmatic ( 776420 ) on Tuesday September 18, 2007 @07:10PM (#20661577)
    Actually, paid out, or not, makes no difference. Either way, it doesn't cost Google anything.

    Let's take a look - clicks are either legit, or aren't.

    Legit Click - Money comes from publisher (not Google), and Google gets a cut.
    Bogus Click, Caught - No money changes hands. Without said bogus click, Google makes exactly the same amount of money.
    Bogus Click, Not Caught - Money comes from publisher (not Google), and Google gets a cut (profit).

    If you look the scenarios, the only for Google to "lose" money is to mis-detect a legit click as fraudulant, as the publisher gets a legit click for free. Google, of course, minimizes this likelihood, and makes sure it's more likely to have false negatives than false positives (I spend enough on Google, and do my own metrics to know this is the case).

    Fraud costs _publishers_ money; it _makes_ Google money (up until the point when advertisers start jumping ship).
  • by vrtladept ( 674792 ) on Tuesday September 18, 2007 @07:37PM (#20661801) Journal
    Exactly.

    Honestly I don't understand how they "loose" 1 Billion in ad revenue due to fraudulent clicks. In fact they don't loose anything. If the clicks were a fraud, then they weren't earned. This is the same kind of funny logic that is used to say that piracy costs billions. There some weird assumption that if this fraud click didn't happen that a legitimate click would happen. In a world with no scarcity ("selling" a click does not prevent selling another click). There is no loss due to clicks that shouldn't count.
  • by catbutt ( 469582 ) on Tuesday September 18, 2007 @07:39PM (#20661815)
    I think you are failing to factor in that the amount google makes on ads is determined by the market rate. And the market rate takes into account whether or not the clicks are real or not, since if it doesn't produce sales equal to the amount they pay for the advertising, people won't pay for it, and the price therefore adjusts downward. Google doesn't necessarily make more or less money because of false clicks, they just are paid less per click, but there are more clicks.

    You say "up to the point where advertisers start jumping ship". They don't "jump ship" per se, they just pay less.
  • by moosehooey ( 953907 ) on Tuesday September 18, 2007 @07:40PM (#20661833)
    This is actually even worse. It could be argued that someone might buy a CD or a movie if they were unable to pirate it. But, you can't say that google would get more legitimate clicks if they could eliminate the fake ones.
  • by DragonWriter ( 970822 ) on Tuesday September 18, 2007 @07:45PM (#20661899)

    But, you can't say that google would get more legitimate clicks if they could eliminate the fake ones.


    You could say Google would charge advertisers more in total, given its rates, if it didn't detect as many of the fake ones, though.

  • by timeOday ( 582209 ) on Tuesday September 18, 2007 @08:42PM (#20662417)
    For google, click fraud is a quality control issue. It dissuades potential advertisers from paying google for clicks because they don't know how many of those clicks are worthless.

    Saying click fraud costs google nothing is like saying bad transmissions cost Ford nothing because the customer eats it. People aren't stupid, so pretty soon things that decrease the utility of the product also hurt the market for the product.

  • Re:Ledgerlines (Score:4, Insightful)

    by DrSkwid ( 118965 ) on Tuesday September 18, 2007 @08:51PM (#20662491) Journal
    People probably have that idea because every $0.01 of every click in your unpaid Adsense account was paid for, up front, by the advertisers, with real money.

  • by tompaulco ( 629533 ) on Tuesday September 18, 2007 @08:52PM (#20662493) Homepage Journal
    We just completely stopped our Google Adwords advertising. We have looked at print, mail, radio and other media advertising, but every time we run the estimated numbers, we find that spending money on advertising would bring us less money in increased sales than we spent on the campaign. I have long suspected, and now believe strongly that the marketers are best not at selling your products but at selling the idea that you need their services.
  • by moderatorrater ( 1095745 ) on Tuesday September 18, 2007 @09:04PM (#20662615)
    To keep its lead in text ads and to avoid law suits, Google puts a lot of time and effort into catching fraud. That costs them money in development and the projects the developers could have done instead. Then there's the percentage they would have taken had that been a legitimate click which they no longer take because they have to nullify it. Then there's the lost bandwidth from having to serve out and record the click, plus the extra servers because of fraud. Let's not forget that they have to pay their lawyers to defend them from the lawsuits that would come whether google's at fault or not.

    It's true that publishers lose more money from fraud than google does, but that doesn't mean it doesn't cost Google either.
  • Re:Ledgerlines (Score:3, Insightful)

    by micheas ( 231635 ) on Tuesday September 18, 2007 @09:39PM (#20662859) Homepage Journal

    Yes, it's called Google's profit. Why that should be of interest to anyone else, including their advertising customers, is not obvious.


    Well, if you are depending on your advertising distributor for your business, you would want your advertising distributor to be profitable, so that they don't disappear into bankruptcy.

    A profitable business is one that you might be able to sue and get money out of. If they are not profitable, and they steal your money, you may have no recourse as the money may not be there to recover.

    Hope that clears up why someone might care if the people they are doing business with are profitable or not.

    Personally I am not concerned if the bike shop I patronize is profitable, however if I made my living with bikes, (racing, courier service, etc.) I would care about the profitability of the bike shop I did business with.

    Hope that makes sense.

Do you suffer painful elimination? -- Don Knuth, "Structured Programming with Gotos"

Working...