No Press Is Bad Press Even Online

otter42 writes "The NYTimes has an 8-page exposé on how an online business is thriving because of giant amounts of negative reviews. It seems that if you directly google the company you have no problem discerning the true nature; but if you instead only google the brand names it sells, the company is at the top of the rankings. Turns out that all the negative advertisement he generates from reputable sites gives him countless links that inflate his pagerank."

Comments

Hyperlinks and Pagerank 101

[Haedrian]

This shows the failure of how hyperlinks works and how the page rank algorithm works.

The Page rank algorithm determines how useful a site is based on the amount of hyperlinks TO the website. Each count is multiplied by how reputable a website is - so if its a huge website which brings in millions of users - then its more likely to be reputable than a website on a free host which gets 10 hits a year.

Now the problem with hyperlinks is that there is no semantic information attached to them - if you place a link on a page - there is no way to mark it as "This is a dangerous page" for example, or "This guy is an idiot, someone shut him up" or "This is an adverstiment, they have nothing to do with us". So the crawler notices a reputable website is linking on another site - and gives points accordingly.

The best solution is to add semantic information to hyperlinks - but that's not supported yet...

Re:Hyperlinks and Pagerank 101

[Anonymous Coward]

Sounds like a "great" idea; in a "We'll only set the good-link attribute, if you pay us!" kind of way.

Re:Hyperlinks and Pagerank 101

[haystor]

So you've reduced the problem to parsing English. That should eliminate the 200 programming languages too.

Nice job.

Re:Hyperlinks and Pagerank 101

[Anonymous Coward]

Don't be ridiculous. Google is international, this magical search engine will have to understand all languages.

Re:Hyperlinks and Pagerank 101

[Haedrian]

Natural Language processing is one of the hardest problems there are in ICT.

If only Search Engines could actually understand the Natural Language we type our websites in - then we wouldn't even need to research anything, just ask the Search Engine and get the correct answer - already pre-filtered and understood.

Won't be happening very soon - because there are tons of ways of saying the same thing.

"The following website gives irreputable medical advice - which could be dangerous for health. [Link]"
"[Link], [Link2], and [Link3] - one of the three scam sites which were detected by [Reputable site]"

The second one is a very interesting problems, how do you mark it? Without marking reputable site as well?

Re:Hyperlinks and Pagerank 101

[underqualified]

The best solution is to add semantic information to hyperlinks - but that's not supported yet...

or you could read the entire post that contains the link

Re:Hyperlinks and Pagerank 101

[datapharmer]

sure there is, use rel="nofollow" if you don't want to share link love.

Re:Hyperlinks and Pagerank 101

[devbox]

This whole conversation is stupid, because Google and other search engines understand rel="nofollow" attribute in links. It's already there. Besides, PageRank is just one system Google uses - they have countless amount of other algorithms too.

Re:Hyperlinks and Pagerank 101

[Alwin Henseler]

This shows the failure of how hyperlinks works and how the page rank algorithm works.

Hardly: a search engine just makes it easy to find popular sites. With popular meaning 'much visited, linked from many other sites', not 'containing useful info from trustworthy source'. A user's mistake starts when he/she loses track of that distinction.

Of course you'd want the 'useful info from trustworthy source', but that's not a search engine's job (to decide what's 'good' or 'bad' info), is it? Google finds it, you decide what it means.

Re:Hyperlinks and Pagerank 101

[blahplusplus]

"Natural Language processing is one of the hardest problems there are in ICT"

That's because they are looking at strings not intents or emotions. Language is not merely symbols there are whole other layers of information embedded in it. Think about words as human beings use them especially in politics think about this word:

Pedophelia.

That word causes people to feel emotions or think immediately of other things on an unconscious level it activates EVERYTHING associated with the word in their minds.

What really needs to happen is that they need to stop looking at language and create models for what is the emotional content of each context they are measuring and what it is pointing at.

Re:Hyperlinks and Pagerank 101

[rolfwind]

How about people do some research first? Caveat Emptor.

I was going to suggest a star system. But any rating system can be gamed, by the web site itself or its competitors.

Also, instead of searching the lowest possible price, find a reasonable one from a known and trustworthy vendor. I have little sympathy for the low dollar hounds that then scream about bad customer service.

OTOH, that's not a criticism of the lady here, she obviously paid money for her glasses and just got stuck with an abusive lying ass. Which goes back to Caveat Emptor.

Re:Hyperlinks and Pagerank 101

[Anonymous Coward]

Pedophelia.

Hamlet prequel?

Re:Hyperlinks and Pagerank 101

[Moryath]

No, how about the regulatory agencies responsible for handling business licensing, fraud, and criminal prosecutions actually get off their ass and do something about this guy?

It should be easy enough to set up a sting on him. Buy a pair of glasses, check if they're fakes, attempt to return them, see what he does in return and record any messages he leaves along with any other interactions with a "honeypot" credit card account.

The problem is that law enforcement agencies in the US aren't interested in doing their jobs when it comes to criminals like Borker.

Re:Hyperlinks and Pagerank 101

[rtfa-troll]

This shows the failure of how hyperlinks works and how the page rank algorithm works.

In fact what this shows is the failure of the current idea that people don't need to be educated about computers. In this case, the question the person has asked is "which is the most 'interesting' link related to Lafont eyglasses". And "interesting" is defined as something like "most discussed". The person thinks they have asked "magically tell me where is the most interesting Lafont link for me".

Page rank is doing exactly what it's meant to do. Now people have to understand what that is. Once they do that, they have to work out how to get what they want (e.g. look for "Lafont shopping comparison" then track through different comparison sites).

Google can add tools which help people understand results. A direct link to shopping ratings for example, but people still need to be willing to do the work to understand what the computer is doing for them.

Re:Hyperlinks and Pagerank 101

[Anonymous Coward]

Google and other search engines understand rel="nofollow" attribute in links

Yeah - but how many authors understand it? How many use it?

Re:Hyperlinks and Pagerank 101

[Cid Highwind]

That's because they are looking at strings not intents or emotions. Language is not merely symbols there are whole other layers of information embedded in it.

...which you can only extract after parsing the strings and symbols!

Getting to the emotion and intent behind a sentence isn't a shortcut to having the computer to correctly interpret language, it's the next step that can only be taken after you solve the problem of getting the literal meaning.

What really needs to happen is that they need to stop looking at language and create models for what is the emotional content of each context they are measuring and what it is pointing at.

And how are you going to determine what a context is and what it "points at" without understanding the literal meaning of the sentence first? It's hard to figure out who to attach the negative emotional content of "pedophelia" to when you don't know whether the sentence containing it is about a molester or about the detective who busted him...

Re:Hyperlinks and Pagerank 101

[Anonymous Coward]

Setting aside the fact that natural language (including spelling) seems to be beyond you, you've now reduced the complexity from
[add a bit of metadata to links (like rel="nofollow")]
through
[create a program with semantic understanding of natural language]
to
[invent a computer with emotions and a strong affinity for the think-of-the-children voting block]

If this thread continues much farther, we'll have this simplified to
[build a Dyson sphere and harness Sol's energy to drag the entire system to the Large Magellanic Cloud, where hypernet search is a solved problem]

Re:Hyperlinks and Pagerank 101

[AvitarX]

Except when shopping they do provide extra information. I would simply recommend people using the product search. It conveniently has a seller rating.

Re:Hyperlinks and Pagerank 101

[ikkonoishi]

Er... this is not Redundant.
Its a real thing.
That does exactly what the GP was talking about.
http://en.wikipedia.org/wiki/Nofollow

Re:Hyperlinks and Pagerank 101

[chebucto]

The example in the article is that a lady searched for some brand of eyeglasses, and the scammer's site was the first result. Call me a populist, but I don't think people should be expected to learn how PageRank works in order to decipher the results of a simple search for some brandname. Google should have a mechanism to handle the bad-reviews-give-good-rank bug.

Educating people is still a good thing, but at first blush I would think teaching people to check the background of the sites they do buisness with would help more than teaching them how Google's sausage is made.

Re:Hyperlinks and Pagerank 101

[Anonymous Coward]

It also really sucks if your name is Peter File.

Re:Hyperlinks and Pagerank 101

[drinkypoo]

Now the problem with hyperlinks is that there is no semantic information attached to them - if you place a link on a page - there is no way to mark it as "This is a dangerous page" for example, or "This guy is an idiot, someone shut him up" or "This is an adverstiment, they have nothing to do with us".

It's called rel=nofollow. HTH, HAND.

Re:Hyperlinks and Pagerank 101

[Anonymous Coward]

Hamlet prequel?

What a piece of work, this A.C.? How funny in posting,
How comedic in faculties, in form and typing,
I would express admiration by action to this Anon Cow, sans apprehension,
But I'm not Mod!

Re:Hyperlinks and Pagerank 101

[Anonymous Coward]

Approximating the overall gist isn't necessarily as hard as parsing the exact meaning, especially if you are averaging a lot of results.
An earlier story featured similar data mining on Twitter messages.

Re:Hyperlinks and Pagerank 101

[acon1modm]

Of course you'd want the 'useful info from trustworthy source', but that's not a search engine's job (to decide what's 'good' or 'bad' info), is it? Google finds it, you decide what it means.

Successful businesses deliver something that people want. Google and others get this, you don't. You're looking at the scenario from the point of view of an employee who does only what he's obligated to do and nothing more.

Incidentally, I personally *DO* occasionally rely on Google to identify a site's legitimacy even now. In situations where a project/organization/company's domain name is not obvious ( e.g. mplayer), googling their name will generally show the top result as the actual, legit site, instead of perhaps a scammer selling a fake.

Re:Hyperlinks and Pagerank 101

[blahplusplus]

"...which you can only extract after parsing the strings and symbols!"

My whole point is they are starting with the wrong idea i.e. trying to parse strings. Strings are merely pointers.

What they should be doing is creating concept networks themselves and building databases that map to these concepts networks of meaning using human input by trying building most common variations of statements that equal the same thing. Then have the interpreter try to guess what the meaning is of all these different statements that mean the same thing to hone in on on the patterns that allow parsing of seemingly an infinite number of statements.

To put it another way they are trying to "Extract meaning" instead of building concept networks from human minds, humans are the arbiters of meaning, so humans necessarily will have to be involved in building these networks.

The string isn't what is important, because the string is only the first layer. It points to actual images/difference definitions in the mind - the true meaning is spacial/structural. What we see as "strings" are merely window dressing (packages), they aren't meaning they are pointers to meaning. Meaning is in our minds - images and differences our minds have recorded.

It's mapping the meanings in our minds that counts

Re:Hyperlinks and Pagerank 101

[Anonymous Coward]

or you could read the entire post that you are replying to and realize that he isn't talking about people, but web spiders.

Re:Hyperlinks and Pagerank 101

[twidarkling]

It was marked redundant because it was mentioned further up the page. It's now marked informative now, though, and probably rightfully so, since it's the kind of information that SHOULD be repeated.

Re:Hyperlinks and Pagerank 101

[Anonymous Coward]

What really needs to happen is that they need to stop looking at language and create models for what is the emotional content of each context they are measuring and what it is pointing at.

Spoken like a psychologist. Did I misunderstand "emotional context", or are you suggesting emotionless people (sociopaths) shouldn't be capable of thinking and communicating? Emotion is not a way to describe thought. Even if you meant emotion is the most abstract, primitive thought, I still have to disagree. Things measurable as emotional responses are present without a brain or central nervous system, in simpler lifeforms. Emotions are basically a physically decentralized reflex coordinating different cells of an organism in response to stimulus, because any appropriate coordination improved their survival and the trait got passed on. It makes sense human beings could survive with a lot less emotion than other organisms, because we have an additional mechanism for coordinating our actions. For example, no emotional incentive for procreation would be bad, but an emotion lacking human could choose to regardless but other creatures under those circumstances would probably avoid it.

In fact, I'm sure the reason people can "read" pedophile and react appropriately by another person's standards, is due to their lifetime of experience involving other people. Because if Mike read it, and responded with nonsense, then you know Mike is illiterate or your brain's NL processing isn't the same as his. If Mike wasn't illiterate, he could explain his response, and you could work out an understanding, retry or fail.

In the end, the only way anyone would accept NL processing as being accurate, is if it was an algorithm that modeled them specifically. And everybody has their own. Otherwise, (using the earlier example) pedophiles would be disappointed or non pedophiles would be appalled.

Re:Hyperlinks and Pagerank 101

[gdshaw]

The Page rank algorithm determines how useful a site is based on the amount of hyperlinks TO the website. Each count is multiplied by how reputable a website is - so if its a huge website which brings in millions of users - then its more likely to be reputable than a website on a free host which gets 10 hits a year.

I think it would be more accurate to say that pagerank measures how notable (as opposed to reputable) a site is. Therein, of course, lies the problem.

Re:Hyperlinks and Pagerank 101

[Anonymous Coward]

I don't know why nobody has thought of this before: http://lmgtfy.com/?q=how+do+i+write+a+search+engine+with+natural+language+recognition

Re:Hyperlinks and Pagerank 101

[Firehed]

Two problems.

One: many sites don't use nofollow where they should, or use it where it doesn't naturally make sense (in an effort to shape their link weighting for SEO reasons). GetSatisfaction is obviously guilty of this, as the weight of their outbound links is counting towards this scam site. Granted it's not quite as simple as "Store below 75% positive reviews? nofollow everything" since that's almost as easy to game, but that conceivably shifts the intelligent nofollow burden from Google to GS. Google (and Yahoo!, Bing, etc) have far more resources to throw at that problem, and them fixing it will be far more helpful.

Two: recently, Google has started giving a (very!) small amount of link juice even to nofollow links. I'm sure they have their reasons, some I'd agree with and some I wouldn't. But that makes it even more important to use the context of a link in weighting it.

Re:Hyperlinks and Pagerank 101

[Zoxed]

> The best solution is to add semantic information to hyperlinks - but that's not supported yet...

It would never work as it relies on the honesty of the web page writer (same as meta-data tags are not useful for determining the contents of a page.

Re:Hyperlinks and Pagerank 101

[icebraining]

The necessary semantic information to prevent this already exists: it's called rel="nofollow".

nofollow is a value that can be assigned to the rel attribute of an HTML a element to instruct some search engines that a hyperlink should not influence the link target's ranking in the search engine's index. It is intended to reduce the effectiveness of certain types of search engine spam, thereby improving the quality of search engine results and preventing spamdexing from occurring.

The problem is simply getting people to tag links at all.

Re:Hyperlinks and Pagerank 101

[Anonymous Coward]

He was saying that the hyperlink is working just fine. It's doing it's job and it's not trying to decide anything about the page except for the fact it is related to what you searched for. This is how hyperlinks should work. If you want more then you need to look at the semantic web, which is where the idea of a search engine that takes context alongside content belongs.

In the normal web, you should be reading the linked post. Let's not try and do the human brains job for it until we're a lot better at it.

Re:Hyperlinks and Pagerank 101

[blahplusplus]

Emotions can be modeled in different ways by rating sentiment of a given word "up or down", i.e positive or negative, you begin with simple emotions that are easily expressed or modeled, you don't try to bite off more then you can chew.

The truth is you don't have any idea and thats why you posted as AC.

Re:Hyperlinks and Pagerank 101

[f.ardelian]

The best solution is to add semantic information to hyperlinks - but that's not supported yet...

There are other solutions, but people just don't care about what and how they link to: https://secure.wikimedia.org/wikipedia/en/wiki/Spamdexing

Re:Hyperlinks and Pagerank 101

[Anonymous Coward]

I would say that there is a lot of information around your average hyperlink (This is a dangerous page) but the search engines are not yet taking advantage of it.

Fuck, I wised I didn't used up all my mod points. Fuck.

Re:Hyperlinks and Pagerank 101

[Calydor]

And how exactly do you suggest doing that on, say, a consumer forum using very limited BBcode? The moment you don't have control over the code generated on the page you may as well give up on being fancy.

Re:Hyperlinks and Pagerank 101

[Matrix14]

It's only *finitely* improbable...

Re:Hyperlinks and Pagerank 101

[Reziac]

And since according to TFA there's a good chance what he's selling are largely counterfeits, and we just discussed THAT issue ... why isn't he being taken down for selling counterfeit merchandise?

Re:Hyperlinks and Pagerank 101

[Haedrian]

See also: Semantic Web, Semantic Markup.

Getting semantic meaning out of a sentence is far harder for a computer than it is for us humans. Unfortunatly humans are lazy, and its pretty hard to markup everything.

But if only we had a perfect complete Onthology and format, then we could map the universe and have computers do our thinking for us. If only.

Re:Hyperlinks and Pagerank 101

[Haedrian]

Because there's no Eyeglasses Association of America (EGAA) yet.

Re:Hyperlinks and Pagerank 101

[Anonymous Coward]

I don't think that anyone would disagree with you that if we could find a way to somehow model knowledge and emotion, we could use it to do some pretty cool things with it including natural language processing. But they might disagree with your statement that researchers are "starting with the wrong idea" by focusing on "parsing strings".

Bottom line is that for the time being, NLP frameworks that apply probabilistic methods to process text (and pretty much ignore the underlying semantics) are orders of magnitude more effective at real-world tasks than any attempts to solve the problem from the approach you suggested. We've gotten useful new technologies (machine translation, spam filters, news aggregation) and we've learned a lot in the process. How is that "starting with the wrong idea"?

Posting anon because I already modded some comments in this discussion.

Re:Hyperlinks and Pagerank 101

[Reziac]

True :) But there is a division of the FBI that deals with counterfeit goods, is there not? I know they've busted dealers of fake-label clothing; eyeglasses should be in the same bucket, perhaps moreso considering how much the damned things cost.

Re:Hyperlinks and Pagerank 101

[Anonymous Coward]

Damn - I accidentally slipped an modded this "insightful".

Re:Hyperlinks and Pagerank 101

[cduffy]

That's the job of the BBcode rendering engine. If they're not adding rel=nofollow universally (or, at minimum, by default), the folks maintaining the relevant libraries need a smack upside the head.

Re:Hyperlinks and Pagerank 101

[cduffy]

Are you saying that a person writing a review saying "Company X screwed me!" would intentionally add semantic information to the hyperlink indicating that Company X is trustworthy? Or are you saying that the folks who own review clearinghouse sites (specifically, ones with high Pagerank values) would intentionally provide metadata contrary to the reviews that they host?

The topic for this article is presumptively honest negative reviews on trusted review sites (as other sites -- such as those run directly by the scammers -- wouldn't have the necessary PageRank "weight") pulling up the rank of the companies they pan. I don't see how adding the opportunity to provide metadata with hyperlinks (with the nofollow tag as a trivial example) would in any way make the problem worse.

Re:Hyperlinks and Pagerank 101

[Anonymous Coward]

Then don't post the link, or don't use such a forum. The solution exists, it's there.

Re:Hyperlinks and Pagerank 101

[leenks]

Supposedly Google only use PageRank in less than about 0.1% of queries now (from a Google Engineer that presented at a previous employer of mine)

Re:Hyperlinks and Pagerank 101

[Anonymous Coward]

"This guy is an idiot, someone shut him up"

Off topic (thus the AC post), why shut up an idiot? Let him speak and prove himself an idiot rather than shut him up and allow the doubt to continue.

Re:Hyperlinks and Pagerank 101

[lwsimon]

Exactly. If I found a page of bad reviews that was PR5 and none of them had nofollows - I'd add every website I could fit :)

Re:Hyperlinks and Pagerank 101

[Anonymous Coward]

Google should have a mechanism to handle the bad-reviews-give-good-rank bug.

Not sure that it's fair to call this a "bug". It's more of a "trade-off" or "consequence" that resulted from certain design decisions that Google made implementing their search engine (i.e. deciding to implement the PageRank algorithm). Any changes to their search algorithm targeted at addressing this issue are likely to incur other trade-offs and other undesirable consequences.

Overall I agree with both the parent's and the grandparent's assessments (after all, they are not mutually exclusive):

• People should familiarize themselves with the limitations of the technologies that they use.
• Engineers should continue to make improvements to technology where possible.
• Consumers should be wary of who they buy from online.
• The government should enact laws to protect consumers from scammers, and enforce those laws to the best of their ability (I added this last point myself).

Re:Hyperlinks and Pagerank 101

[davester666]

or you could post the link as plain text instead of a hyperlink, if it's such a crappy site...like what Google does for links to sites with malware.

Re:Hyperlinks and Pagerank 101

[HTH NE1]

So it's like The Three Amigos: unable to tell the famous from the infamous.

No need to RTFA

[Anonymous Coward]

Thanks to the NYT's valiant efforts, you can be spared from reading TFA: just check out the comic instead.

Re:No need to RTFA

[sootman]

No, read the article. It's much better (worse.) There's a special place in Hell reserved for this guy.

Re:No need to RTFA

[Anonymous Coward]

They should do this for every article!

Nofollow?

[WPIDalamar]

Perhaps review sites should add nofollow attributes to their external links

http://en.wikipedia.org/wiki/Nofollow

Maybe just for negative reviews?

Re:Nofollow?

[janek78]

Or maybe not use hyperlinks? I mean, if I want to warn people about www.crazyscammer.com, why would I need to make it into a clickable link? I don't. Now he gets linked from NYTimes. Well done.

Re:Nofollow?

[Haedrian]

I had never heard of this attribute before. And I'm pretty sure that "Blogger using automatic content system to input the article" wouldn't have heard of it either.

Sure you can do it with all links, but that'll ruin how Page Rank works.

Re:Nofollow?

[plover]

How about combining the worst with the worst? Add "nofollow" to any links in negative reviews, and copy negative reviews to a shadow site that's filled with links to sites known for porn, pedophilia, viruses and malware.

Re:Nofollow?

[mounthood]

rel="nofollow" isn't new and it's useful for large websites with user submitted content. Slashdot, for example, adds it to outbound links to remove any incentive to increase page rank by spamming links. All the major search engines respect it AFAIK.

Re:Nofollow?

[ais523]

Well, if a link's flagged as "we think there's a decent chance this link is irrelevant or spam and will contaminate your search results", what sensible search engine would pay attention to it? This is one of the cases where nofollow helps both the website that uses it, and the search engine that looks at it.

Re:Nofollow?

[TheTurtlesMoves]

Since the story broke, things have changed a bit. The ISP has or was going to take down the site. Master card and visa are blocking his service and ebay has blacklisted him. It seems others don't want to test the "negative" publicity angle.

Re:Nofollow?

[modecx]

One problem is that various forums don't have a way to add nofollow or other attributes to links... Where do people go to complain? Besides the various merchant review/complaint sites, they often go to forums. The second problem is that many forum scrips see something like a URL, they automatically parse that into a link as default behavior.

Re:Nofollow?

[cduffy]

Sure -- but if those forum scripts don't tag such user-submitted links nofollow, they're doing it wrong.

Re:Nofollow?

[modecx]

Agreed, that would be ideal, if only to keep link spam from fudging your own sites' rating. AFAIK most forum packages don't nofollow external links, and either require the admin *cough* to modify some code or install an add-on (and indeed know to look for it in the first place)

Only users in the mod, admin groups should be able to share link love, by default. Dunno why they haven't figured that out yet.

Re:Nofollow?

[gregzeng]

How about combining the worst with the worst? Add "nofollow" to any links in negative reviews, and copy negative reviews to a shadow site that's filled with links to sites known for porn, pedophilia, viruses and malware.

Good idea. Scamware appears, but no unsubscribe link. If you do unsubscibe, it validates for more scamware. Could you please be more explicit though, with an example? So I forward the unwanted scam onnto where, and how? All my emails are handled in real time by Gmail & Yahoo.

ORLY?

[Gr33nJ3ll0]

Seriously tried a couple of the searches in the article.... nothing came up. Maybe they fixed it for this guy, who knows. Anyway, I always check for feedback on new (possibly sketchy) sites. If it sounds too good to be true, it probably is.

Re:ORLY?

[Rick Richardson]

Google: Christian Audigier glasses

Re:ORLY?

[Gr33nJ3ll0]

Good point, 5th from the top.

Re:ORLY?

[oDDmON oUT]

Google: Christian Audigier glasses

DME is #6 on that results page, what's your point?

A smart shopper would click here, and quickly find that DME is *not* a low price leader for any style of frame from that designer, pictured on the results page.

Not to mention that lower priced vendors on that page are associated with e-tailers that have standards and pull...or is Amazon not "all that" anymore?

Do your homework on the vendor you intend to buy from, don't take links blindly, and you'll be fine.

Re:ORLY?

[StikyPad]

In TFA, he mentions that he also sells through Amazon using a business name he won't disclose. To be fair, he says he offers much better service through Amazon because they have a low threshold for dissatisfied customers, but personally I don't want to risk giving him my money, which now rules out Amazon.

New Twist

[the eric conspiracy]

A new twist on the old aphorism that there is no such thing as bad publicity.

Re:New Twist

[Anonymous Coward]

Bad Publicity 2.0

(I'm sorry; someone had to.)

The kneejerk reactions

[baegucb]

I read the NY Times article last night at work, and did a bit of looking around at metafilter, reditt, and other sites. It was going to show up on slashdot eventually, and thought about the comments to follow. I'd group them as:
1. Google is teh evil for it's algorithms
2. Let's harass this asshat
3. The BBB/police/FBI/NY AG is inept
How about we go in another direction in the comments? (I will admit some amusement at what he will go through now that he's come to wide attention).
What should we do?

Re:The kneejerk reactions

[Blue Stone]

Let's not forget CitiBank who apparently didn't give a shit that someone fraudulently closed the woman's disputed charge.

I have an account run by CitiBank, and this has made me decide to close it - before I get subjected to their don't care approach.

Re:The kneejerk reactions

[Antique Geekmeister]

Perhaps some education of web users is in order. Looking up the actual comments on Ebay, rather than merely whether there is a negative one, would help reveal the sort of vindictive comment this bully threatens other vendors with. And educating consumers on how to get their complaints actually dealt with by blocking the accounts of abusive vendors, rather than merely refunding a particular purchase, would be invaluable.

I've also read the article. Keep in mind that if Google starts manipulating page ranks based on the "semantic content" of the linking pages in order to lower the ranks of such awful businesses, they open themselves up for a _world_ of pain refusing requests from governments or litigious companies to "lower" the scores of links that criticize them, effectively censoring critics without having to actually block them from posting their material. And that kind of linguistic analysis is computationally expensive. It would burden a system that's already maintaining incredible amounts of data: it might cost quite a bit more money to implement, money out of Google's pockets, that they don'w want to spend.

Re:The kneejerk reactions

[slackbheep]

Yeah, we all knew the BBB and co are useless, but it was a little shocking to hear about such terrible service from the costumers bank.

Re:The kneejerk reactions

[Sponge Bath]

...terrible service from the costumers bank.

The funny masks should have tipped her off.

Re:The kneejerk reactions

[dangitman]

1. Google is teh evil for it's algorithms

2. Let's harass this asshat

3. The BBB/police/FBI/NY AG is inept

So far, not a single comment along those lines. Nor would I expect them. Perhaps your slashdot crystal ball is broken?

Re:The kneejerk reactions

[larry bagina]

What part of "CitiBank" did you not understand?

Re:The kneejerk reactions

[baegucb]

You didn't give it enough time:
http://tech.slashdot.org/comments.pl?sid=1886536&cid=34364802
http://tech.slashdot.org/comments.pl?sid=1886536&cid=34364792
http://tech.slashdot.org/comments.pl?sid=1886536&cid=34365434
http://tech.slashdot.org/comments.pl?sid=1886536&cid=34364720
http://tech.slashdot.org/comments.pl?sid=1886536&cid=34364794
http://tech.slashdot.org/comments.pl?sid=1886536&cid=34366990
http://tech.slashdot.org/comments.pl?sid=1886536&cid=34365822
Of course, Alan Ralsky was the classic case for slashdot.

Re:The kneejerk reactions

[jpatters]

Seriously, IMO, everyone should just fire their bank and join a credit union. This shit would never happen with a credit union.

so?

[underqualified]

big deal. it's simply functioning as designed.

Re:so?

[Anonymous Coward]

not as desired though.

Something has changed

[microcars]

for this guy and not for the better as he expected.
A quick search on LAFONT CIBA VISIONS (they contacts first mentioned in the story) now show 11 results on the first page, of those SIX are either this story or re-posts of the story on other websites.
The other 5 results appear to be for things like BizRate and for legit CIBAVISIONS outlets.

Clicking through page 2, 3 etc there is at least ONE site showing a discussion about the NY Time article but DecorEyes does not show up.

maybe I am not using the correct keywords or this guy has some other websites he controls?

Re:Something has changed

[Bigjeff5]

You're doing it wrong, Lafont and Ciba Vision are different brands. She found Lafont frames on DecorMyEyes first, then decided to order the contacts at the same time.

DecorMyEyes is halfway down the list for both "Lafont Eyeglasses" and "Discount Designer Frames".

No longer #1, but still top 10 for both.

Re:Something has changed

[microcars]

ah thank you, I figured I had done something wrong.

Blaming Google?

[MDMurphy]

Reading the article I had to shake my head at the complaints that "Google should do something". Do they think the phone book people should boot him out or circle his name in red too?

I've seen people complain in online forums about ads or intrusive product placement in TV shows and movies. And what do they do? They mention the product or company 3-4 times in their rant. On threads where I've commented I've tried to encourage people not to name the specific product or company in their rant, but too many clueless people out there. You hate the placement but remember it well enough to bitch about it online. Advertiser searches for online comments and finds many, so the campaign was successful.

Heck, the Times article just boosted the guy's profile even more! There wasn't just one mention of the company, but many, increasing the rank. Of course the company name is still likely to just show up in a search for the company not the brand names.

So the guy is an ass, but all the clueless people who want to blame someone else ( Google ) and not do research on a company but just buy whatever is claimed to be the cheapest. They may not be getting what they deserve, but they did contribute to their problems by their lack of due diligence. "Too good to be true" is still a true statement. If you find something online where everyone has it for about the same price but someone magically has it much lower you're asking for trouble. That's when you really need to check on the reputation of the seller.

Re:Blaming Google?

[TheLink]

On threads where I've commented I've tried to encourage people not to name the specific product or company in their rant, but too many clueless people out there.

Heck, the Times article just boosted the guy's profile even more! There wasn't just one mention of the company, but many, increasing the rank.

So the guy is an ass, but all the clueless people who want to blame someone else ( Google ) and not do research on a company but just buy whatever is claimed to be the cheapest

1) How are you going to check the reputation of the seller/product if everyone does what you suggest and not name names?
2) Search engines like Google and Bing don't simply rank by "number of mentions".

You don't seem that clueful yourself.

There will always be clueless people (and he wants to only do business with the weak and/or clueless). The real problem is cops and regulators aren't doing their jobs and stopping him- assuming he really made threatening and fraudulent phone calls.

Re:Blaming Google?

[Software Geek]

Reading the article I had to shake my head at the complaints that "Google should do something". Do they think the phone book people should boot him out or circle his name in red too?

Google should do something. Not because of any moral or ethical obligation, but because it is good business. There are numerous examples of internet businesses that have followed the pattern:
1) Attract lots of traffic, become immensely profitable.
2) Attract enough traffic to become appealing to parasites.
3) Become so infested with parasites that your site becomes aggravating to use.
4) Lose lots of money when all your traffic starts going elsewhere.
MySpace is a good example of this pattern. EBay and Facebook are well on their way.
Google has a long and mostly successful history of combating page rank parasites. Google understands that in order for them to succeed in their business, they must be trusted. Not trusted in the sense "I got screwed, but it wasn't google's fault." Trusted in the sense "I use google because they seem to know the site I was trying to find even when that's not exactly what I asked for."
The fact that one parasite has successfully exploited google search using this technique means that many more parasites will be following soon. Google needs figure out how to defeat this kind of parasite, or risk losing their hard earned trust.

Re:Blaming Google?

[drinkypoo]

Google SHOULD do something. That's what pagerank is all about.

Re:Blaming Google?

[leenks]

Except PageRank barely features in Google search results these days. But hey ho!

Re:Blaming Google?

[Anonymous Coward]

Reading the article I had to shake my head at the complaints that "Google should do something". Do they think the phone book people should boot him out or circle his name in red too?

Well, the phone book doesn't try to rank numbers on importance. They allow highlighting or buying bigger entries, just like google allow advertising around the search results.

Google provide that service, and their popularity is based around the reliability of those results. Both ethically ("do no wrong") and for their own business reputation, yes they should do something. And no doubt they are!

Should they be chastized for how it is now? Nah. Any system will be gamed somehow, it's still an incredibly useful system.

Re:Blaming Google?

[MDMurphy]

Giving you a clue for free:

You can mention the product or site and still obfuscate the name. You can keep from posting links to the seller's web site if you don't want to add to their business. Also I was referring to product placement comments not seller rating sites. ( You omitted that when quoting me to attempt to make some sort of argument )

"simply rank by number of mentions" is your clueless comment. Adding "simply" to what I wrote changes it to nonsense. More mentions in a page makes that page more relevant to a search.

If the seller is violating laws or regulations as it appears this one is, then the appropriate authorities need to do their job. It's not Google's job, which it what I also wrote.

Google will block links to sites with malware, but if the site is perfectly fine, only the guy who owns it is an asshole, how is that Google's problem? There's no harm in viewing the site, only if you have the misfortune to do business with them. Should they also remove all linkage to fake diet pills, magnetic bracelets and good luck charms? If they start doing that, and a site does show up in a Google search should the dimwitted web surfer assume it's passed the Google seal of approval? If Google claims to filter out crooks does the dimwitted surfer have a claim against Google if they get ripped off from a vendor they found via a Google search?

Re:Blaming Google?

[TheLink]

You can mention the product or site and still obfuscate the name.

There are so many ways to obfuscate the names. So how are you going to research it? If everyone uses obfuscated names it's going to be much harder when researching.

The same clueless people would have even more difficulty researching obfuscated names than the clueful, so they'll just be as screwed as before, while those with a clue would have to work harder to figure out obfuscated names. So what's the net benefit of your suggestion?

It should be pretty obvious that obfuscation is a stupid idea.

Re:Blaming Google?

[drinkypoo]

Please post a link to the currently-running search engine code.

Re:Blaming Google?

[greap]

When did the phone book start ranking businesses and suggesting the method of ranking should provide a degree of confidence in results that appear up top?

Google already harvest review data from a significant number of sites (have a look at local sometime) and integrating this in to pagerank should be a no brainer, not only does it improve the quality of the results but it builds consumer confidence in the results themselves.

This is the fundamental problem with the democratic ranking mechanism google use, it’s extremely open to gaming and lacks the semantics to express the intention behind a link.

Re:Blaming Google?

[MDMurphy]

I was making a comparison to discussions about intrusive advertising both times I suggested obfuscating company product names. The whole purpose behind that suggestion was to limit the ability to search for them, to limit the effectiveness of the annoying advertising or product placement. So you got that part right, you just missed the point.

4Chan assemble

[frap]

I wouldn't wish 4Chan's wrath on my worst enemy but it sounds like this guy needs a taste of his own medicine to me.

Re:4Chan assemble

[Anonymous Coward]

Why should 4chan assemble against him? He's running his business just like one would expect of a /b/tard.

Re:4Chan assemble

[Anonymous Coward]

4chan is a semi-organized sack of limp dicks. grow up dipshits.

Re:4Chan assemble

[Anonymous Coward]

Hell, given his apparent sociopathy, I wouldn't be surprised if he was a /b/tard.

Re:4Chan assemble

[Anonymous Coward]

Oh you've done it now. Consequences will NEVER be the same.

Re:4Chan assemble

[M4DP4RROT]

Unfortunately, TFA makes this guy sound like someone 4chan would worship, actually: a man who's made internet trolling into a successful business model! If anything, 4chan users are more likely to emulate than emasculate.

Re:4Chan assemble

[Anonymous Coward]

You just don't understand 4chan at all.

Re:4Chan assemble

[Anonymous Coward]

And yet they accomplish more with their limp dicks than yours ever will.

Re:4Chan assemble

[trytoguess]

Then please explain. I know 4chan is at it's core a random collection of people, but it does look like it's active members only care about laughing at "idiots" and on occasion taking up crusades against universally hated things like killing kittens and Scientology. Perhaps 4chan would take actions against this man. But history suggests otherwise.

Douche much?

[spynode]

If this guy is all about the traffic, maybe he would appreciate a DDoS attack? Not that I'm implying anything.

Mass versus count nouns

[noidentity]

The NYTimes has an 8-page exposé on how an online business is thriving because of giant amounts of negative reviews

No. Choose one:

giant amounts of negative review
giant numbers of negative reviews

Re:Mass versus count nouns

[Tacvek]

What about "a giant amount of negative reviews"?

Re:Mass versus count nouns

[noidentity]

What about "a giant amount of negative reviews"?

The problem is still reviews. You don't have an amount of something countable; you have a number of them.

Short-sighted?

[jfengel]

It may be that the guy is raking in cash today, but he's not just being a jackass: he's committing crimes. It's fiendishly difficult to prosecute some kinds of online crimes, especially when routed through overseas sites, but this guy does not seem to be protecting himself.

It's always wise to be suspicious of "trend" stories, since newspapers love to spot a single instance, call it a "trend", and get everybody yapping. But even if there is a "trend" here, it'll get cut right short if this guy gets arrested.

Which may be the real purpose behind the piece: take an injustice that is too small for authorities to take notice, raise its profile, and take some satisfaction when the police step in.

There may well be a marketing tactic to be had in providing rotten customer service and benefiting from the links provided by sites too dumb to use "nofollow". But there's a line between "rotten service" and "outright fraud", and this guy is well over it.

Re:Short-sighted?

[dcollins]

"But even if there is a 'trend' here, it'll get cut right short if this guy gets arrested."

RTFA: Guy's been arrested twice already.

Re:Short-sighted?

[sampas]

The guy in the story is taking advantage of the fact that the authorities (who we're paying for via tax dollars) will do absolutely nothing for months and months until more and more complaints pile up. Most police reports are now online so that crime victims don't waste officers' time reporting crimes. The victim in the story reported the crime to numerous authorities, who responded by doing nothing for a long time. Likewise, the bank did nothing. It would be cool if Google did something about it, but it's not their problem. Our tax dollars pay the salaries of those who are supposed to protect us. It's a fairly simple case of fraud, harassment, and threat of violence. If you ever thought the your bank would protect you in case of fraud or identity theft, they won't. The only thing the banks protect is their own money. Maybe you've noticed the news about mortgage fraud? Not many convictions there, either. On the upside, though, they did manage to catch Bernie Madoff eventually....

Re:Short-sighted?

[jfengel]

Thanks for pointing that out. I did RTFA, but I musta missed that.

I guess he figures the money must be worth the time, but it seems unlikely.

Microformats

[mounthood]

I'd really like to see a microformat for reviews take off. It could start in parallel to existing link-counting schemes so we just need a critical mass implementing it. Counting links is easy for search engines but we could get much better information by saying what we mean instead of just a link, possibly with nofollow.

There is hReview but it's in really bad shape. There isn't any agreement on nomenclature for the reviews -- no scales or weighting schema, or any way to communicate a rating schema. Anyone know of better or more common microformats for reviews?

Re:Microformats

[Anonymous Coward]

http://www.google.com/support/webmasters/bin/answer.py?answer=99170

Don't blindly use search engines for shopping.

[bcrowell]

The solution to this seems pretty obvious to me. Don't blindly use search engines for shopping. Why in the world would anyone sane google for a product, go to that site, and enter a credit card number?

I buy my running shoes online, but I don't just google '"New Balance" 883 9-2E' and then go to the top hit and buy. I buy through Amazon, which shows me retailers that are willing to sell me those shoes and gives me users' ratings for the various retailers.

Granted, it's unfortunate that this has the effect of siphoning off revenue from mom-and-pop businesses to giant corporations like Amazon and E-Bay. But it reflects the fact that online reputation is hard to evaluate, and that Amazon and E-Bay are doing something useful by building up a database of online reputations.

In theory, I could do a couple of other things that would be better for small retailers. I could go to a local bricks and mortar store. Well, that sort of works, but unfortunately the local bricks and mortar stores don't stock shoes in the wide size I want to buy. I could also google small internet retailers, then try to evaluate their reputation by googling for their name, or by going to some online site that rates online reputation. The problem is that those methods of evaluating reputation aren't very reliable.

Re:Don't blindly use search engines for shopping.

[dangitman]

Why in the world would anyone sane google for a product, go to that site, and enter a credit card number?

Because it's quick and convenient?

Re:Don't blindly use search engines for shopping.

[leenks]

Having been burnt many times by Amazon "stores" I'll quite happily take my chances with stores I find on a search engine that appear reputable to me. And over all those I'll try and source things locally first, even if it costs a bit more.

Case in point - I just bought a new thermostatic mixer shower (and fitted it at yesterday) from a local supplier rather than Amazon or eBay because the local supplier has been here for donkeys years and it is easy to deal with them for returns / warrantee issues. Too many Amazon stores disappear or are a nightmare WRT warrantee or returns.

Or at least be more thorough...

[Junta]

If I'm about to spend enough that I'll care if it goes south and find an apparent bargain, I may give weight to early results, but I'll also search on the companies names, look for reviews, and see if the BBB has any complaints.

Of course, the downside is that reviews are going to have a negative tendency if at all possible. When things go ok, people aren't motivated to say anything. When things go bad, they will step up. This is one reason I look hard for comments associated with bad reviews to see if it is something unreasonable or legitimate being complained about.

In this case, the text would point to completely unacceptable behavior that would kill the sell. Other times, I'll see people complain due to their own mistake (e.g. rating a video card down for not having DB-15, which is self-evident from product pictures).

Re:Or at least be more thorough...

[bcrowell]

Of course, the downside is that reviews are going to have a negative tendency if at all possible.

Yeah, and although I hate putting money into the pockets of Amazon and E-Bay, this is the reason that they're really useful. They both try pretty hard to get buyers to rate sellers, and therefore you get a relatively unbiased set of ratings compared to what you'd get from the BBB, etc.

No Website is an Island

[Tink2000]

I find that using MyWOT coupled with Chrome keeps me from stuff like this. Both of those reporting tools are crowd-sourced and goes back to the old word-of-mouth method of advertising. Otherwise, you just stick with the places you know.

"Press" is streaming media...

[crovira]

After the initial hubub dies down in a few hours of days, who'll remember why?

The idiot consumer'll just remember that he'd heard the name somewhere.

That's the fault of the "non retentive nature" of media.

Unless there are "survival reasons" to remember, we forget.

You're countrymen ate children from a certain economic/religious/ethnic group in the previous war?

Get into a media relegated memory economy and you'll be able to generate positive press in no time.

You'd have to really f*ck up badly to become a pariah. (To this day, Gypsies aka Romanche are reviled and relegated to impoverished lives all across Europe because of something that happened back during the crusades. Nobody remembers exactly what or why, but something did.)

I do not doubt the ranking

[epseps]

But I do doubt that it is good for his business.

He claims that using google maps to stalk angry customers combined with harassing phone calls threatening rape helps his business with Search Engine Optimization from generating negative reviews.

His techniques match those of Brooklyn based discount camera and electronics retailers of a few years ago, with the object of the personal phone calls being to get a few $100 more out of suckers who think they are getting the best price on something. These camera scammers used to have multiple websites, cheap cameras listed and they would call and threaten customers using "Italian" names. Then complaints would mount and they would leave and move on to the next website, but I do not think they earned much money by doing this.

One reason I doubt Borker (hilarious name) is making loads of cash on this is because he handles the phone calls himself, that means there are not many calls (he answers "Eyewear" in the same way the camera guys used to answer the phone "Photography" or "Cameras") but I do believe he is making money off of something. Perhaps it is his other company that is referenced in older whois lookups of his websites called AOSI, which appeared to be a search engine optimization company. I am not sure the company has the same name now, but that might explain why he was happy to be interviewed about his crappy businesses.

Oh, and I really doubt his Wall Street story too. He used to have an office at 305 Madison Avenue which is a temporary office / mobile office rental address. He formed "OpticGenius.com" and ran it from there. I do not know too many "Wall Street" people who give up their jobs and devote themselves to running scams from home or temporary offices as a better source of income.

Not that simple...

[denzacar]

The best solution is to add semantic information to hyperlinks - but that's not supported yet...

From TFA (which CAN be linked without the loginwall):

“If you have a lot of people who hate Obama, for instance, and you decided to rank on love or hate, you might not be able to find the White House and that would be terrible,” he says.

Russian Mafia, pure and simple

[Anonymous Coward]

This guy's got Russian Mafia written all over him.

If you do decide to fuck with him, I suggest you make VERY sure whatever you do
cannot be traced back to you. I've known a few of these characters, and they
make anything you saw in "The Godfather" look like Sesame Street.

Semantic metadata is supported by HTML

[Anonymous Coward]

The words between the start and end of the a tag carry that meta information... remember Google bombing? Not enough? The title= and the rel= add yet more metadata. The problem isn't a lack of metadata about the link.

The problem is that people, especially Joe and Jane User posting on forums, don't fill in the metadata, don't know it exists, wouldn't know html if it bit them on the arse and barely know how to copy paste a URL! Google thus doesn't scan for more metadata than the contained words and rel="nofollow".

Are People Really Reading the Results?

[oDDmON oUT]

Sure, Google DecorM****s.com and the top five hits are all for that company, followed by a link to Resellerratings.com where the company has a stunning 1.39 lifetime rating out of 10.

Googling Lafont (with multiple suffixes), designer glasses or designer frames brings up a number of vendors, DecorMy... not being one of them.

Seems to me people need to hone their search skills and *always* search for ratings on vendors when ever they make a purchase, particularly when using them for the first time.

[Please note, I munged the company name, just to insure he gets no wuffie from this post]

More thorough...

[Mateorabi]

Yeah, if a deal even has a whiff of to-good-to-be-true or I've never heard or dealt with the company I will usually do a search for (company name fraud) or (company name scam) in the Google. You have to take the results with a grain of salt and use some judgment: there are some loud, unreasonable customers out there. But typically a company trend is fairly evident.

Popular != Reputable

[pcardno]

You seem to have confused popular with reputable, i.e popular = A Good Thing. Just because a thing has many fans does not mean it is reputable. I give you Justin Bieber.

If you don't have NY Times Account

[Anonymous Squonk]

Via Google News:

http://bit.ly/fEKab6

Re:If you don't have NY Times Account

[nashv]

Eh? I don't have an NYTimes account and I could read the article just fine...

Not in jail???

[Dthief]

How is this person not in jail. Threats of sexual assault, threats of regular assault (in email), harassment (in email or voicemail, easily passed to the police), fraud (pretending to be her when calling Citi).

The fact that the police id not follow through is appalling

Because that works SO well

[SheeEttin]

No Press Is Bad Press Even Online

an online business is thriving because of giant amounts of negative reviews.

...because that worked out SO well for Cook's Source Magazine.