Forgot your password?
typodupeerror
Google Technology

Google Algorithm Discriminates Against Bad Reviews 175

Posted by CmdrTaco
from the this-algorithm-sucks dept.
j_col writes "According to the official Google blog, Google has altered their PageRank algorithm to not give back linking points to bad reviews of websites belonging to online retailers, following the publication of a recent article in the New York Times describing one woman's experiences in being harassed by an online retailer she found via Google. The specific changes to the algorithm are of course a guarded secret. So considering that these changes are already live, how do we know how the algorithm determines a bad review from a good one, and whether or not innocent online retailers will be wrongly punished by having their rankings downgraded?"
This discussion has been archived. No new comments can be posted.

Google Algorithm Discriminates Against Bad Reviews

Comments Filter:
  • ...of google search results for 'discount designer sunglasses'

    So is it really 'buried?'
  • Poor summary... (Score:5, Informative)

    by msauve (701917) on Thursday December 02, 2010 @01:30PM (#34420614)
    "Google has altered their PageRank algorithm to not give back linking points to bad reviews of websites belonging to online retailers"

    Uh, no. Google changed it so that websites of poorly reviewed retailers lose points, not the reviews themselves.
    • by Jugalator (259273)

      Thanks, I was pretty confused there for a while after re-reading the sentence three times. :-S

      • by msauve (701917)
        Yeah. The chosen link anchor didn't help, either. It might make you think that kdawson was logged into CmdrTaco's account.
      • I had the same problem. I didn't understand WTF the summary was saying, so I had to RTFA.

        Hey, wait a minute. googleblog.blogspot.com is gaming the system to get more pageviews, by getting people to post bad summaries to Slashdot!

        Slashdot's editorbot should use the coherency engine on summaries, and if the summary doesn't make sense or has too much non-humorous ambiguity, it should penalize the linked articles.

  • by rsborg (111459) on Thursday December 02, 2010 @01:31PM (#34420624) Homepage

    ... not "bad reviews", which would be very anti-consumer.

    Instead, the poorly reviewed products and services are going to lose index.

    This kind of selective pressure will reward those companies whose services and products garner better reviews.

    I just wonder if this will lead to more astroturfed reviews and payola for review-sites like Yelp.

    • by Desler (1608317)

      This kind of selective pressure will reward those companies who can afford to pay people to destroy the page ranking of their competitors.

      FTFY.

      • Your point would have been better accepted (to me anyways) if you hadn't used the "Fixed that For ya" meme -

        But yes, I fear this won't encourage more positive reviews but only more negative reviews between competitors.

        Perhaps it will eventually reach a point where competitors will push each other into the dirt with bad reviews and a completely unreviewed product will be the one with the highest rank.

        • Except as the google blog post stated, negative reviews don't harm the company's site, they just don't get positive boosts and credit from it.

          Essentially before, both positive and negative reviews helped and counted, now only positive reviews help and count. This doesn't allow other companies to destroy other's page rank.

        • by xero314 (722674)
          The whole point of the Google change is to not include bad reviews in the page ranking. I actually means that you can't destroy a competitor with bad reviews because those bad reviews will not affect page ranking. Previously both bad and good reviews increased the page ranking. This caused products that where poorly reviewed to be at the top of the page rankings. Now it will take more positive reviews to move to the top of the page ranking, while negative reviews will have no affect on the page ranking
      • by rsborg (111459) on Thursday December 02, 2010 @01:49PM (#34420900) Homepage

        This kind of selective pressure will reward those companies who can afford to pay people to destroy the page ranking of their competitors.

        FTFY.

        I thought about that but the article states that

        an algorithmic solution which detects the merchant from the Times article along with hundreds of other merchants that, in our opinion, provide an extremely poor user experience

        .. I presume this means that the weighting would not be linear, but more like an exponential dropoff when reviews are numerous, time-disjoint, and all negative. I'm sure Google has done at least a sample analysis using their mountain of data. I think the biggest point made here is that (as a vendor) services to monitor your product/service will become increasingly important so you can reply to negative reviews and actively manage any trolls... whether this leads to more engagement or simply more astroturfing is yet to be seen.

      • by xero314 (722674)
        How can a competitor utilize poor reviews to destroy their competitor when Google is ignoring links from poor reviews. The positive reviews will still boost the page ranking while poor reviews, rightful or astroturf, will be ignored. So yes their competitor could create bad reviews that Google will ignore, but that seems like a big waste when they could just write no reviews and Google can ignore those as well. At worst some of those bad reviews would get included in the page ranking and actually improve
    • by gstoddart (321705)

      I just wonder if this will lead to more astroturfed reviews and payola for review-sites like Yelp.

      Safe bet ... as they say in the article, people are trying to game Google rankings constantly ... if there's money to be had, someone will keep trying.

    • by RandCraw (1047302)

      If Google's change does what's intended, downrank URLs of merchants who invite furious web opinion as a marketing ploy to game search engines, only the losers, like bile-thriving DecorMyEyes' Vitaly Borker will seek alternative means of self promotion. Frankly, I suspect that's a pretty small contingent of potential astroturfers -- 'hundreds' according to the Google blog.

      I suspect that Google has indeed applied Sentiment Analysis, but done so narrowly, targeting only (1)merchants described with (2)domain-s

    • by bonch (38532)

      I just wonder if this will lead to more astroturfed reviews and payola for review-sites like Yelp.

      Well, of course it will.

      By the way, double-spacing every sentence is annoying to read.

      Don't you think?

  • by gblues (90260) on Thursday December 02, 2010 @01:32PM (#34420650)

    The blog does not say what the contributor says it does. The closest it comes is noting that the links from the negative reviews never counted in the first place because the sites hosting the reviews used the "rel=nofollow" attribute on the links. What it does say is that they have altered the algorithm to punish bad businesses more effectively in response to the NYT article that suggested that being bad could be good for business.

    Move along, nothing to see here!

  • Simple (Score:5, Funny)

    by CheshireCatCO (185193) on Thursday December 02, 2010 @01:33PM (#34420664) Homepage

    They look for phrases like

    • ...burst into flames...
    • ...still sobbing for her pet rabbit...
    • ...sucked into the trans-dimensional vortex...
    • ...shouldn't even have been any radioactive material IN a children's book...
    • ...and that's how little Tiffany learned about death and accidental dismemberment...
    • ...came to my home and set it on fire and then kicked my dog...
    • ...never knew I was capable of that sort of pain...
    • ...ordered the complete Beethoven Symphonies and the discs had nothing buy Justin Bieber on them...
    • ..contained a live bobcat... (obligatory)
    • ... would not buy again...
    • Re:Simple (Score:5, Funny)

      by Anonymous Coward on Thursday December 02, 2010 @02:03PM (#34421070)

      Great, Slashdot's PageRank just dropped like a rock thanks to you...

    • by Chapter80 (926879)

      I am no algorithm expert, but if I were in charge, I would start with that last phrase that you listed ("would not buy again") along with other similar phrases ("is a scam", "feel completely ripped off", etc). Then I would scan the massive Google database to see if I could find other phrases that are frequently located "near" to those key phrases, and see what I come up with.

      I bet I'd come up with a pretty good list of phrases that have negative sentiment. I'd eyeball those, and get rid of the obvious bad

  • by nog_lorp (896553) on Thursday December 02, 2010 @01:33PM (#34420674)
    ... who feels like Google results have gotten really, really bad? I know it can come in waves as the SEO arms race progresses, but srsly. I feel like Google's user base has shifted from technical people to the average populace, and so have the results.
    • by inputdev (1252080) on Thursday December 02, 2010 @01:39PM (#34420770)
      have you tried Bing? ;)
    • The "technical people" population hasn't driven Google's user base since about 6 months after their inception.

      I still normally find what I need on the first page of results in any case, so without knowing what you're looking for... I couldn't say you're doing anything more than typical nerdbitching because they're popular.

      • +1. Seriously, I've never even heard ^anyone^ (in person, this excludes places like /.) complain about the quality of results returned by Google (and I live in Silicon Valley, the conversation in the bar turns to tech all the time).

        I'm a huge fan, but only because it is the best I've seen. Hell, I used to be a huge fan of hotmail - at one time (way, way long time ago) it was the best free email available...

        • by vux984 (928602)

          I'll complain about them. Finding useful review of something on google is a pain... search for a game...

          IGN ... your "ULTIMATE SOURCE FOR news and media"... half the time its a place holder page; I just roll my eyes every time I see that in the search results.

          or you get Nextag or Dealtime or any of a dozen other price / review aggregators with a place holder page populated with generated content that was scraped from somewhere else, and then barfed on.

          And its the same content on all of them, because they'r

          • Well, like I said, places like /. are the only times I hear complaints about them... I don't search for games, so I've never had that particular result. Every time I test google against the competition, I'm happier with google... and that is just the results, the appearance of some of them makes me want to puke. Take Yahoo.com for example - bunch of flashy crap and ads that make me feel like I'm surfing in 1995.

            As always YMMV, and I really don't care if the whole world does not see things the same way I do

      • I'll complain too, try searching for something that has a non-alphanumeric character in it. Or put something specific in "quotes" only to have google completely ignore the quotes, strip out numbers or characters you are trying to find and return that result instead.

        Or all the times it ignores what I typed completely, "Hey we think you are an idiot, so we searched for FOO instead of BAR, if you really want to search for BAR click here. But we think you want FOO.
    • by drinkypoo (153816)

      I have definitely observed greater difficulty in googling for pages that I am familiar with containing technical info but whose URL I cannot find at the moment, even when putting the site name into the search terms (With or without site:). On the other hand, I can pretty much ALWAYS find ANYTHING on my OWN site by putting its name (just the name, not even the .org) and a term or two into the google query, which I find interesting. Much of what I can't find is forum or mailing list archives, but of course, s

      • by leenks (906881)

        PageRank is well documented, and scores for pages are available (or were) through the Google Toolbar, but you are correct in that to be effective the actual algorithms Google use to rank results need to be fairly closed (to avoid exploitation).

        Nowhere in the linked blog post does Google mention PageRank - this is a poor summary.

        I've been at technical events where Google engineers have given talks and explained that PageRank is now used for less that 0.1% of all queries. Clearly graph algorithms are still us

    • I think it's trying to be a bit too smart lately, constantly being "helpful" by searching for what it thinks I want to search for instead of what I actually wanted to search for.

      Many times it actually corrects a real spelling mistake of course, the difference might be that when I make a spelling mistake I can only blame myself, but when google corrects something it shouldn't I can blame google..
      • by 0123456 (636235)

        I think it's trying to be a bit too smart lately, constantly being "helpful" by searching for what it thinks I want to search for instead of what I actually wanted to search for.

        Bingo. 'Smart' search engines suck, particularly for technical information, because you can never really tell what it's going to search for... the 'smarter' Google searches get, the less useful they become.

        Of course it doesn't help when people pick names for their projects or products which are the same as or very close to some other word that's in common usage.

      • Append "&nfpr=1" to the search URL. If you use keyworded bookmarks in Mozilla: http://www.google.com/search?q=%s&nfpr=1

  • by Todd Knarr (15451) on Thursday December 02, 2010 @01:34PM (#34420686) Homepage

    One hopes the guys at Google took into account the business that sets up a fake review site for the purpose of posting negative reviews of competitors to get Google to falsely downgrade them. My bet's on a manual filter to remove such sites, probably based on a discrepancy between those sites and every other review site.

    • Wouldn't that be similar to (and perhaps offset by) people already doing them same with favorable reviews for themselves? It doesn't seem like a new problem, anyway.

    • My understanding is that this would not affect results at all. What they are doing is no longer giving points to a site because a negative review linked to that page in the review. Before the change if you wrote a review that said "You should not buy a book [amazon.com] from Amazon because their books in particular are absolutely terrible an no one should go there!" Then Amazon's page rank for the keyword "book" would get a bump from the back-link to it. If you think about it this doesn't really make sense as you'd b

  • by kwabbles (259554) on Thursday December 02, 2010 @01:36PM (#34420736)

    Google is the worst company ever. They sold out and went evil. I give their company a poor review and personal rating.

    .
    .
    .

    good great wonderful cheese love flowers butterflies excited appealing chocolate yay amazing cool googlicious

  • by acomj (20611) on Thursday December 02, 2010 @01:39PM (#34420780) Homepage

    Its a long interesting read.
    Quite the character mr. Borker is.

    http://www.nytimes.com/2010/11/28/business/28borker.html [nytimes.com]

  • You have to wonder if Mr. Borker is familiar with that phrase.

  • bad reviews return to make decisions about company..namely not to use them.

  • How do we know that all of the "Barack Obama sucks" websites out there won't make it harder to search for the White House? Just one example of where semantic inclusion may not work.

    • by blueg3 (192743)

      That would be sentiment analysis, which TFA specifically mentions they don't use because of exactly the problem you describe.

  • Article submitter sounds like a SEO moron suffering from a case of sour grapes.

  • by gurps_npc (621217) on Thursday December 02, 2010 @02:06PM (#34421108) Homepage
    I tried to post this earlier (guess I was too slow). But, considering that Google specifically said they would be looking into this later, that means: In the (paraphrased) words of Coots and Gillespie: They are making a list, And checking it twice; gonna find out who's naughty and nice"
  • I wonder how well that algorithm works with sarcasm.

  • TFA specifically indicates that they don't do something as straightforward as is described (which would be sentiment analysis). Instead, they implemented some algorithm that lowers the ranking of some merchant websites that, according to them, provide a poor user experience. No further details on how their algorithm behaves are given. It doesn't even indicate that people giving poor reviews of the merchant or website factor in to the ranking change at all. (The only part where this comes in to play is their

  • by fuzzyfuzzyfungus (1223518) on Thursday December 02, 2010 @02:14PM (#34421240) Journal
    I wrote my favorite escort a glowing review, all about how much and how hard she sucks, and now google is going to downrank her and it will be all my fault! This is terrible...

    On a more serious note, correctly assigning "positive" or "negative" to a given adjective or phrase, across a wide range of subject areas, must actually be something that would give the computational linguists a bit of trouble(or 10,000 interns a very boring time of it)... Simply parsing star ratings or "out of 10" is easy enough; but is a vacuum cleaner that sucks good or bad?
  • Why is everyone acting like page ranking should be anything but whatever Google wants them to be? They are free to use whatever they wish to determine the search results. If they decide to never show Hotmail when you search for email, there is absolutely nothing wrong, ethically or legally, with this. They are free to shun a competitor. They are free to put you last on every search if they just simply don't like you. They can put whatever they want into the search results, in whatever order they wish,

  • by Animats (122034) on Thursday December 02, 2010 @02:50PM (#34421692) Homepage

    This is the fundamental problem with "crowdsourcing" reviews. Where the number of reviewers is large compared to the number of items being reviewed, as with movies, it works fine. Where the ratio is small, it doesn't. It's far too easy to game the system. There are automated tools for that. [wikipedia.org]

    This problem has become worse since the October 27th change to Google, when Google Places/Maps results were merged into web search. This made "local" results much more prominent. Look at the first screen of Google search results for a local product or service. Most of what you see are Google Places results, maps, or ads. The organic results are so far down they don't matter.

    As a result, the "black hat" SEO companies are now aggressively targeting Google's places and maps system. "Convert Offline" is quite open about this, with their article Dominating Google Maps- The Most Effective Spam Ever And What You Can Learn From It" [convertoffline.com] In some ways, Google Places is more vulnerable to attack than organic search. The number of web mentions of a local business tends to be small, so the amount of phony material that has to be generated to make a business look good is also small. Each mention carries a lot of weight.

    Google might lose this battle. Craigslist did. Back in 2008, Cory Doctorow wrote about "Spammers discuss breaking Craigslist verification system" [boingboing.net]. It's become much worse [techdirt.com] since then. Personals were the first to go, and are now over 90% spam. Then Computer Services and Self Employment fell to the spammers. Jobs and Real Estate are under attack. Along the way, Gmail became a spam haven [google.com], especially after Jiffy Gmail Email Creator [cnet.com] became widely used.

    The fundamental design assumption of Google is that important stuff has lots of links to it. That's not a valid assumption in local search.

  • The specific changes to the algorithm are of course a guarded secret. So [...] how do we know how the algorithm determines a bad review from a good one

    I think you pretty much answered that one yourself.

Little known fact about Middle Earth: The Hobbits had a very sophisticated computer network! It was a Tolkien Ring...

Working...