Slashdot Mirror


A Closer Look at Google Adwords

zaphle writes "This article describes an interesting experiment with the Google Adwords service; in an effort to fine-tune the price per word, a mirror site was set up, paying a different price per word. I turns out the second site had to pay more in order to reach a similar click-through rate. My questions to the slashdot community: are organizations like Google redefining the law of demand and answer? To what extent does this imply a competitive advantage for larger companies? Do we need an ethical framework to direct companies to make such algorithms open source?"

24 of 224 comments (clear)

  1. Forgot some experiments... by luvirini · · Score: 3, Interesting

    Like making a site with the $ .05/word.. and few others.. the total cost and time to do that should be quit low as said in the artickle. anyone using adwords to generate sales should definitely try to find the best results by doing those types of doubleblind tests.

    1. Re:Forgot some experiments... by Anonymous Coward · · Score: 5, Interesting

      On the contrary, Cringley has utterly misinterpreted these results and they actually prove that you should not do these tests.

      If Google does anything with AdWords the obvious thing to do is examine the BIDS for AdWords that it receives and figure out a VALUE for each AdWord.

      By setting up a rival site and BIDDING AGAINST HIMSELF this guy drove up the VALUE of his AdWords.

      How is this not obvious? Google just coded the free market.

    2. Re:Forgot some experiments... by MemeRot · · Score: 2, Interesting

      That's a pretty good theory. Also a proven track record probably figures into their algorithm somewhere. So when one account had generated a lot of click thrus for a long time, and one was brand new (and overpaying) they probably looked at the new account with suspicion. If the new account is willing to throw money away with $1 per click, google will be glad to help them. But if they pick a more moderate amount, they probably stop getting the "I've got money to burn" special treatment. I think if the experiment continued for a long time the results would trend towards each other, minus whatever dupe penalty the second account gets. And of course page rank may also factor in somehow, and again with the proven track record of the first page, it's probably got many more links to it than the new site.

      By the way - wasn't this guy doing this to experiment with different ads and offers? There's always the chance with Cringely that he's just completely wrong in saying that the two sites were identical. The whole point of the experiment was to make them not identical.

  2. Google scares me, this I know! by mister_llah · · Score: 3, Interesting

    I've said it before, I'll say it again. Google is a company. They are out to make a profit.

    Companies grow from profit, Google has grown a lot. To maintain growth percentages (which as you become gi-normous like Google, becomes harder to do) ... you need to branch out and you may also do things that may be questionable.

    I think that paid search priority is somewhat ethically questionable, but I am not at all surprised.

    Given that Google has been taking efforts to make themselves appear even more friendly to the open source community (those huge contributions awhile back) ... I think with enough contact/interest from people in the community that Google may become interested in sharing these algorithms (good public relations never hurt) ...

    Who knows, though?

    ===

    However, Google doing things that are questionable and quite publically, in my opinion, spin a dangerous message for the future. They may be progressing into a more pervasive position than Microsoft in the years to come, with increased power comes increased corruption.

    Scary stuff!

    --
    MoM++ - A Classic Expanded - [Master of Magic 1.5]
    http://mompp.sourceforge.net/
  3. Just a thought by mikkom · · Score: 4, Interesting

    Just a speculative thought because no one really know about googles algorithms.

    The result may be due to the original pages higher pagerank. I wouldn't be surprised if google would give higher position to "better" sites even in ads. In Googles context, higher pagerank means "better" site.

  4. Hmm... by TheSHAD0W · · Score: 4, Interesting

    I have an alternate explanation. I believe what existed before the guy's experiment was a more-or-less stable "ecology" around those particular keywords. There were probably a number of people paying a premium for a limited number of clicks for those keywords, well above the 10-cent level he was originally paying for. Google probably sorted the higher-paying advertisers onto the best pages and left the dime-a-click ones for others and everyone was more-or-less happy.

    Then when the experiment began, it disrupted things. The advertisers who were initially offering a premium found themselves with fewer clicks as their ads were placed on less advantageous pages, or when their ads were displaced entirely. They then changed their own behavior, perhaps by choosing different keywords and/or paying higher rates. This would have cascaded, causing other advertisers to change their behaviors.

    The end result would've been a shift in AdWords' performance with those keywords, one that wouldn't easily be reversible, and which could account for the poor performance when the experimenter reduced his bids for clicks.

  5. Quality score by Anonymous Coward · · Score: 0, Interesting

    There is also this quality score factor that will seriously impact the bid. It appears that if your adwords is performing badly, youe min bid amount can increase many fold!

  6. Google Has become Advertising Platform by xoip · · Score: 3, Interesting

    Long gone are the days of getting a simple answer from Google without advertising. As a search engine, it still does a very good job but, there are far too many people out there thinking that Google is the Only online advertising platform. This may have begun skew the results in favor of sites that have no real content other than a bunch of google ads. Google wins either way, click on the paid ads based on the search or ads loaded on the landing page of the search result. At some point, Google will loose it's lustre.

  7. Google has ethics: make money by tod_miller · · Score: 2, Interesting

    It works for all, it is capitalism at its best:

    Anyone can bid prices, but the costs you need vary based on how well your ads do.

    Useful pages that are well typed to keywords (lemonhead ads for a search for lemon chicken would have to pay more for instance) are good for us, which means good for google, because google want to make more money.

    Advertisers are stupid, they want to be top of EVERY SEARCH no matter how useful it is, and they want it cheap.

    Google says, the less relevant you are to the search, the higher the click through, the higher the cost.

    If you happen to convert your audience, and you now become more relevant, you prices go down.

    So if I start selling neckties to skaters, I might have to stump up a bit in the long run, but if I hit a craze, they would go down, until some chump makes his own neck ties and starts bidding above me.

    I think it is dumb to make this public, and the guy behind this has an ulterior motive anyway.

    Misleading ads change the equation, but what can you do.

    please type the word in this image: revamps
    random letters - if you are visually impaired, please email us at pater@slashdot.org

    --
    #hostfile 0.0.0.0 primidi.com 0.0.0.0 www.primidi.com 0.0.0.0 radio.weblogs.com
  8. I wonder.... by squoozer · · Score: 3, Interesting

    .... if this guy didn't trip over a duplicate content filter. I would be very surprised if Google didn't check to make sure it wasn't being fed the same content from multiple sources. From Googles point of view checking for duplicates is a good thing. They don't want their natural listings (or ad listings I imagine) to be filled with hundreds of copies of exactly the same site.

    I would have been more interested to see the results of a test that modified the wording of ads and how that affects placement.

    Finally, I wish I was getting 15000 click throughs a day. Sigh.

    --
    I used to have a better sig but it broke.
    1. Re:I wonder.... by Anonymous Coward · · Score: 1, Interesting

      Can I ask a stupid question since I don't understand AdWords fully.

      If this guy is paying $0.10 per AdWord and is getting 15,000 click-throughs per day ... correct me if I'm wrong, but he's paying $1,500 per day to Google for advertising? $500,000 per year?

      What do you do if you get 15,000 click-throughs per day but only 15 sales because of it?

  9. Re:Sadly its all true: An insiders view of Google by SpecBear · · Score: 5, Interesting

    I skimmed this and thought, "Hmm, this looks like the kind of text that would be generated by a script." A couple of minutes of searching (via Google, ironically enough) turned up the Automatic Complaint Generator.

    Sigh. Remember when trolling was an art form, when people would put time, effort, and (dare I say it?) heart into inciting flame wars, even when posting as Anonymous Cowards. The kids these days are just phoning it in, and that saddens this oldtimer's heart.

  10. I welcome relevant ads by core+plexus · · Score: 4, Interesting
    I know there's a lot of google-haters, but I have to say...

    I was thrilled with the ad-block extensions of Firefox, and welcomed the relevant ads from google. I'll admit, I have actually clicked on, and even (shocked) bought a few things.

    I hate desperate ads, like those on TV and everywhere else. Advertisers realise that they are failing.

    When/if google starts flash, popups, then start to complain.

    Tired of online retailers charging extra to ship products to Alaska?

  11. Something that I only realised the other day.. by Idimmu+Xul · · Score: 3, Interesting

    Google not only knows what you like, thanks to the things you search for, but it also knows who your friends are, and what they like thanks to relationships formed through GMail invites and Orkut which could come in very handy when it comes to targetted advertising in the Christmas season (and any other gift buying season).

    And I'd be quite appreciative of that as I've no idea what to get my Dad this year, and a few casually placed Google Adwords undermining my own thought process wouldn't go a miss!

    --
    The problem with slashdot is that most of its users were bullied and stuffed into lockers as kids!
  12. Re:Why they can get away with it by KiloByte · · Score: 3, Interesting

    I have yet to see a Google ad that is relevant to what I would like to see. And, I'm afraid that they already wasted their chance.

    So... what ads
    http://adblock.mozdev.org/
    are
    zone "googlesyndication.com" {type master;allow-query {any;};file "/etc/bind/db.blackhole";};
    you
    apt-get install adzapper
    talking
    http://www.customizegoogle.com/
    about?

    I don't ever pay for random software -- I buy only things I need to (because @#$%^& customers won't switch to usable systems), and I sometimes help with Free Software projects (donating code, not money). For non-software related things, the banking system in Poland is so abysmal that purchasing material things online is simply out of the question; also, I have a strong negative response to ads -- I make conscious decisions to boycott products that are advertised in an annoying way.

    Losing the clicks from the rest of the company I happen to admin the servers for is just collateral damage.

    --
    The creatures outside looked from Alt-Right to Antifa; but already it was impossible to say which was which.
  13. Article is just wrong by Proto23 · · Score: 5, Interesting

    First of all if he gets 15000 @ 0.15 clicks per day he is paying $1500,- a day in advertising that is $547.500 per year in advertising alone. His friend must be an internet miljonair! I somehow doubt that such a large operation would execute the test as described in the article. I think the figures are a bit inflated to make it a better story. Furthermore I use Google Adwords for my company (http://www.tiouw.com/). I spend around $5000 a year on Google. As everyone who has ever worked with AdWords knows is that when you change your ad, it changes your click through rate. Changing URL's, text, anything has a direct result. It takes some time before the system gets used to the changes and then you are back on track. I expect that if the company in the article would have run for some time with the new settings it would generate more and more hits. Finally, tests have shown that people do not click on the no.1 position, but prefer no. 2 or 3.

  14. Vote with your dollars, not your brand of ethics by Heembo · · Score: 3, Interesting

    are organizations like Google redefining the law of demand and answer?

    You mean supply and demand, the cornerstone of capitalism? More like - Google is redefining the rules of advertising and IT for the entire world.

    To what extent does this imply a competitive advantage for larger companies?

    Well, just like the superbowl, only companies with big bucks can get prime time advertising real estate.

    Do we need an ethical framework to direct companies to make such algorithms open source?"

    Keep your ethics and morality out of my consumer choices. If I dont like how google does business, I will stop buying from them. I live on Kauai, and I turned my girlfriends dying massage business into a thriving business (www.kauaioutcallmassage.com) spending only 20$/month over the last 2 years. Google has been incredible for my family, please don't rain on or change my parade with Google!

    --
    Horns are really just a broken halo.
  15. Re:Huh? by assert(0) · · Score: 1, Interesting

    > Correct me if I'm mistaken, but I keep on hearing that "open source" is about freedom. What's google got to do with "Open Source"?

    --
    (founded 95,000,000 yrs ago, very space opera)
  16. Re:The Google Business Model by Motherfucking+Shit · · Score: 2, Interesting

    I wonder if it's any coincidence that if you search for a phrase in Google's "you had invalid clicks so we aren't going to pay you" letter (that phrase being "that invalid clicks have been generated on") in Google, you get 545 results, but if you perform the same search at MSN, you get 832 results.

    Things that make you go "hmm." MSN is rarely if ever better than Google at search results.

    --
    "BSD: Free as in speech. Linux: Free as in beer. Windows 10: Free as in herpes." --Man On Pink Corner in #52607549.
  17. Re:Huh? by Ortega-Starfire · · Score: 1, Interesting

    Someone here has obviously never read ANY pre-Christian history...

    --
    ---- Liquid was a patriot ----
  18. Deeply flawed experiment by G4from128k · · Score: 2, Interesting
    This experiment is deeply flawed because he compared Adword performance of a new site to that of an existing successful site. This flaw biases the results in two ways. First, it assumes that Google's ranking system isn't biased by click-through history -- that Google doesn't up-rank a site with 12 months 15,000 click-throughs per day versus one with only 1 month of 15,000 click-throughs per day. This seems very very unlikely. If I were Google I would always up-rank incumbents that had a good history of click-throughs (and payments). Second, it assumes that people don't remember site names (that the existing successful site has no brand value). I know that I often refind casually interesting sites (those not worth remembering or book marking) by rerunning a search that found that site the first time. How many of that existing site's 15,000 click-throughs are repeat customers who recognize the site's name?

    The better experiment would create two or more new sites and test adword on an even footing of history with both Google and searchers.

    --
    Two wrongs don't make a right, but three lefts do.
  19. Comment removed by account_deleted · · Score: 2, Interesting

    Comment removed based on user account deletion

  20. Re:The Google Business Model by Anonymous Coward · · Score: 1, Interesting

    You little clue.

    $120 billion is their stock worth. Not their generated revenue, and not their profits. Stock worth is not very correlated to revenue/profit really.

    Their quarterly profit estimate was $650 million a little while back off of $1 billion in revenue (so minus salaries, equipment costs, network traffic, etc. etc.). $14 million would be 2% of that, but that percentage is meaningless because there is no context to the $14 million (esp. wrt time) versus a quarter (1/4 of a year) revenue number.

    Further, you don't know how many people who claim to be scammed were scammers or actually scammed, or the timeframe for all the claims (spoken or not). You don't know if this number is high or not. Neither do I.

    But if *any other company* was screwing people out of $14 million over, say, the past 2 years since Adwords likely became prominent, their practices should be investigated. To date, I haven't heard much on that end. What concerns me is that I'm not sure why, if you suspect that a company is screwing people out of millions, you are giving them any sort of pass whatsoever. Seems a double-standard because you "like" Google or have been brainwashed by their "do no evil" ex-mantra.

  21. AdWords Experiment Flawed by Anonymous Coward · · Score: 2, Interesting

    One thing that wasn't considered in this experiment is the fact that Google frowns upon websites that have duplicate content. By doing so, the tester may have notced he was experiencing a lower page rank as well as fewer listings in organic listings. This could have caused the Google ad to be discounted and even viewed as a SPAM website (obviously a big no-no). While we cannot contest the power of Google, the system is much more intelligent than the everyman and will not reward people trying to receive more traffic by creating duplicate websites (even if a test).