Kuro5hin.org: technology and culture, from the trenches
create account | help/FAQ | contact | links | search | IRC | site news
[ Everything | Diaries | Technology | Science | Culture | Politics | Media | News | Internet | Op-Ed | Fiction | Meta | MLP ]
We need your support: buy an ad | premium membership

[P]
SearchKing sues Google

By kpaul in Op-Ed
Sat Oct 19, 2002 at 08:28:21 PM EST
Tags: Internet (all tags)
Internet

According to a press release on the SearchKing site, Bob Massa is suing Google because they "arbitrarily and purposefully devalued his companies' and his customers' web sites, causing his business to suffer financially."


From what I've been able to gather from information around the web, SearchKing was set up specifically as a site to drive traffic, dangerously close, imho, to the definition of a link farm.

To be fair, Massa explains his side in a press release:

SearchKing began business as an Internet search engine and web hosting company in 1997, approximately a year before Google's inception. In August 2002, PR Ad Network began placing text ads for businesses on web sites with a high PageRank from Google, thereby becoming one of very few competitors to Google's advertising service. According to the lawsuit, once Google became aware of this, it lowered SearchKing's PageRank and the ranking of the web sites it hosts.

PageRank (PR) is a Google-developed system of determining the value of a particular website. The PR of a site, which ranges from one to 10 (10 being the highest), is displayed publicly on each site visited through the use of the Google tool bar, which can be down loaded to any computer for free, PR value is determined several ways, including calculating the number of web pages (links) pointing to a particular page and how relevant they are to the topic at hand.


Now, this from Google's information for webmasters:


2. Why does my page's rank keep changing?

We update our index every four weeks. Each time we update our database of web pages, our index invariably shifts: We find new sites, we lose some sites, and sites ranking may change. Your rank naturally will be affected by changes in the ranking of other sites. You can be assured that no one at Google has hand adjusted the results to decrease the ranking of one site or increase the ranking of another. Google's order of results is automatically determined by several factors, including our PageRank algorithm. Please check out our Why Use Google page for more information on how this works.

You may want to check and see if the number of other sites linking to your URL has changed. This is the single biggest factor in determining what sites are indexed by Google, as we find most pages when our robots crawl the web and jump from page to page via hyperlinks. To find out who links to your site, use Google's link: tool.


Is Massa simply upset that Google's algorithm caught him breaking one of their rules, or is he insinuating that Google hand adjusted his site's ranking?

Google makes it very clear what kind of activity they don't like to see. Among the actions they frown on are when webmasters, "Participate in link exchanges for the sole purpose of increasing your ranking in search engines." This sounds a lot like what Massa was trying to do when he attempted to sell advertising based on Google's PR value of his site.

If Massa's PR has dropped, and Google says they don't hand modify the listings, could it simply be a matter of him losing PR (like many others) for trying to openly do something Google frowns upon? I'm sure a penalty for being in a link farm is coded into the algorithm in some manner. (Not to mention the fact that the September update was big enough to get a mention on Slashdot and Wired.)

What is the definition of a link farm? Internet.com describes it as, "The process of exchanging reciprocal links with Web sites in order to increase search engine optimization. The idea behind link farming is to increase the number of sites that link to yours because search engines such as Google rank sites according to, among other things, the quality and quantity of sites that link to yours."

Was Massa merely attempting to cash in on the hard work that has gone into making Google what it is today? Here's a look at Searchking's submission page as of posting: "Link Popularity - Increase your site's link popularity NOW! From just $29.00!"

Personally, I try to follow Google's advice - build a content rich site for human visitors and let Googlebot do what it does best without trying to 'manipulate' the algorithm. I'm a staunch believer that if you have good content, the visitors will come. As your site gets richer and increases in depth, people across the web will willingly link to it, which Googlebot will notice and give you your props for all your hard work.

Further reading:

Webmaster Dos and Don'ts - (From Google)

First appearance on SearchKing? - (From WebmasterWorld - 7/12/00)

PageRank For Sale - (PANDIA - 8/24/02)

Overture Sues Google - (InternetNews.com - 4/5/02)

Sponsors

Voxel dot net
o Managed Hosting
o VoxCAST Content Delivery
o Raw Infrastructure

Login

Poll
What is your favorite search engine?
o Google 97%
o Altavista 0%
o AllTheWeb 0%
o Ask Jeeves 1%

Votes: 142
Results | Other Polls

Related Links
o Slashdot
o Google
o press release on the SearchKing site
o link farm
o Google's information for webmasters
o Why Use Google
o Google's link: tool
o Google's algorithm
o Google's PR value
o September update
o mention on Slashdot
o Wired
o Searchking 's submission page
o Webmaster Dos and Don'ts
o First appearance on SearchKing?
o PageRank For Sale
o Overture Sues Google
o Also by kpaul


Display: Sort:
SearchKing sues Google | 38 comments (29 topical, 9 editorial, 0 hidden)
-1 My objections go beyond the (now fixed) link (4.27 / 11) (#6)
by HidingMyName on Sat Oct 19, 2002 at 05:11:21 PM EST

I'm not keen on this, it looks like a nuisance suit. Perhaps the best thing is for us not to dignify it with our attention. It looks like SearhKing as a typical bad dot.com business model .

communication (3.25 / 4) (#9)
by kpaul on Sat Oct 19, 2002 at 05:18:46 PM EST

i feel anything that sparks a conversation about bad practices on the web is a good thing.

who knows, seeing others talking about it might cause someone who would otherwise blanket his site with blink tags to just not do it.

;)


2014 Halloween Costumes
[ Parent ]

agreed (3.50 / 4) (#15)
by jcolter on Sun Oct 20, 2002 at 03:51:24 AM EST

What alternative do we have?  This does not sound to me like an attention grabbing kid that if we ignore will go away.  How does the fact that you find the suit wrongheaded make the story not worthwhile?

[ Parent ]
What the fuck?!?!? (4.35 / 14) (#12)
by Talez on Sat Oct 19, 2002 at 08:28:25 PM EST

I'm sorry but you can't use a competitor to promote YOUR wares and then complain when they notice you using THEIR resources to promote YOUR product.

I'm sorry. The world doesn't work like that. I hope Google wins this and countersues these twits for wasting their time and money that could have been spent on Google's excellent technical innovations like Google Labs or Google Compute.

Si in Googlis non est, ergo non est

Searchking believes google acted against them... (3.00 / 5) (#13)
by IronWolve on Sat Oct 19, 2002 at 10:16:19 PM EST

Searchking really believes that Google marked down Searchking sponsered sites. I doubt google does that sort of thing, Searchking better provide proof. It will be interesting to see what "proof" they have.

impossible to prove (4.50 / 4) (#14)
by martingale on Sat Oct 19, 2002 at 10:56:26 PM EST

It's certainly possible that Google could introduce a penalizing weight for certain web pages. In fact, they do so for pr0n sites. However, to have a case (irrespective of its merits) SK would have to show that the change in PR they suffered fell outside of the normal range of statistical variation. Since actual PR is not published, they can only rely on site rankings. The trouble with that is that site rankings are not a continuous function of PR. Let me explain:

If sites A, B, C have PR 1, 2, 3 say, then the order obtained is C, B, A. Easy, and safe under small perturbations - ie if the calculation gave 1.001, 1.98, 3.12, the order would still be C, B, A.

Now if the PR is 1, 1, 3, then the order can be C, B, A or C, A, B. A small perturbabion now has a large effect. Suppose that the true PR values are 1.001, 1.01, 3.07, but the calulation gave 1.07, 1.002, 3.10, then the ordering would simply be wrong.

In other words, it is mathematically impossible even in principle to deduce PR ranking from the site orderings, and site orderings are the only observable quantities. Since Google claims to order based on PR, it is effectively technically immune from the SK claims.

[ Parent ]

Okay. (4.00 / 4) (#18)
by mindstrm on Sun Oct 20, 2002 at 05:52:10 PM EST

What if google DID lower their ranking on purpose? So what?

I mean, that would say something about google... but...
Is it illegal? Should they be sued for it?
Was there some implied agreement between them and SearchKing? Does google not have the right to tailor their listings however they want? They don't owe anybody anything.


[ Parent ]

ianal (5.00 / 3) (#19)
by martingale on Sun Oct 20, 2002 at 08:04:32 PM EST

Certainly, Google can tailor the listings any which way they choose, however I'd expect two things where they have to be careful:

1) As a search company, they benefit from a lot of trust in their service. Historical examples however show that when the perception appears that their listings aren't objective anymore, they will lose mindshare and eventually marketshare. The perception of fairness is crucial. So they can't arbitrarily penalize some sites and not others.

2) As a common carrier(or whatever the exact legal term) in the US, Google benefits from immunity from prosecution under the Millenium Copyright Act. For example, the Google cache violates copyright, but Google is immune under the law, so long as they do not edit or otherwise modify the information. If they can be shown to tamper with their listings in a sinister way, I'd expect some lawyer to argue they don't deserve the immunity, and the next day they'd be dead and buried under a ton of lawsuits.

So it's not like they don't have to be careful what they do.

[ Parent ]

Common Carrier (4.00 / 1) (#34)
by mindstrm on Tue Oct 22, 2002 at 12:31:33 AM EST

I believe common carrier status is a specific legal status assigned generally to telephone companies.

It is my understanding that you just can't claim "common carrier" status because you want to. You don't get immunity from the law just because you claim to be treating everyone equally... immunity like that has to be carefully handed out.

[ Parent ]

Google is an OSP (4.00 / 1) (#35)
by martingale on Tue Oct 22, 2002 at 12:55:25 AM EST

You're right, my terminology is wrong. The correct term is Online Service Provider. See here for example, or just take a look at this earlier k5 discussion.

[ Parent ]
Okay..cool.. (none / 0) (#38)
by mindstrm on Sat Nov 02, 2002 at 08:40:54 AM EST

And that only applies to the DMCA.

I suppose, say, some author could decide that becaues google cached an infringing copy of his work, that he's going to sue google outright for contributory infringement (because he doens't have to notify them to take it down, because he believes they don't qualify as an OSP becuase they muck with their own content) then he's free to do so.. but give me a break.

Keeping how they rank sites a secret hardly disqualifies them from being an online service provider.


[ Parent ]

I believe the PR is visible in the Google tool bar (4.50 / 2) (#22)
by KWillets on Mon Oct 21, 2002 at 03:45:22 AM EST

Or at least the article above says so.  

The original pagerank algorithm just passed PR to linked pages on an even basis - if a page had a PR value of R, and k links, each linked page got a boost of R/k, or more like .85R/k.  So finding the PR's of a page's back links and summing accordingly should produce an approximate PR sufficient for checking.  Since all the terms in the summation are positive, it's sufficient to get a few major terms and sum to get a reasonable lower bound on the true PR.

If Google's PR is less than this approximated figure, it would indicate that the page was downgraded for some reason.  It's probable that the original flat R/k distribution across outbound links has been replaced by some less evenly weighted distribution; this change is workable as long as the weights sum to 1.

[ Parent ]

dunno (4.50 / 2) (#25)
by martingale on Mon Oct 21, 2002 at 05:04:17 AM EST

I don't know for sure, but I think I remember reading somewhere that the Googlebar's pagerank display is a number between 1 and 10. That's what I meant when I said it's not published. Someone with IE care to confirm?

For your method to be applicable, you'll need floating point PR values, though. In particular, you'll want the log of PR to be accurate, since the vast majority of web pages will have vanishingly small PR (the web's been verified to follow a power law for outlinks, and I've seen a paper claiming that PageRank has therefore a power law in this case, too). You've also not included PR due to external sources, but as you say, you're aiming for a lower bound.

I'm not sure that even Google calculate PR with constant relative error, though. The matrix iteration procedure everyone mentions usually leads to L^p error estimates, and that can mean large relative errors on the web scale.

So I'm not convinced you'll be able to test it quite like that. What do you think?

[ Parent ]

True, 1-10 sounds suspicious (3.00 / 1) (#26)
by KWillets on Mon Oct 21, 2002 at 01:23:38 PM EST

It might well be concealed.  It's probably still possible to guess one's decile from linking pages.  For instance, if you get ten links from ten pages whose pseudo-rank is X, having ten out-links each, you should rank X or better.  

It would be much harder to find such a situation, but it's technically possible.

I'd be interested in the paper on the power law.  Is it online?

[ Parent ]

no, it really is 1-10 (4.00 / 1) (#28)
by jayfoo2 on Mon Oct 21, 2002 at 05:52:11 PM EST

I've got the google toolbar on one of my boxen and it really does display a 1-10 slider bar for each page.

[ Parent ]
I meant... (4.00 / 1) (#29)
by KWillets on Mon Oct 21, 2002 at 06:25:02 PM EST

that the 1-10 range in the toolbar doesn't sound like what the real pagerank value should be.  I suspect the pagerank is remapped and rescaled to form a more even distribution for the toolbar.

[ Parent ]
power law, ranges (4.50 / 2) (#32)
by martingale on Mon Oct 21, 2002 at 07:48:15 PM EST

I was just lazy about quoting the paper, but you can download it here.

On the question of the real range of PR, it's obvious that 1-10 won't cut it. Google has 2-3 bn distinct urls in its index, and each has a pagerank value. Assuming for a moment that each url has a distinct PR value, that's 2-3 bn values which, if we normalize them to sum to 1, means a typical value on the order of 2e-9. My guess is that the PR score in IE is logarithmic. But even so, cramming 2 bn points in the interval 1-10 is going to get messy.

[ Parent ]

more linkage on toolbar/PR (5.00 / 2) (#33)
by kpaul on Mon Oct 21, 2002 at 08:33:52 PM EST

The Implementation of PageRank in the Google Search Engine

not my site, but an interesting read...

Although, this is greek to me ;)

Toolbar PageRank Real PageRank
0/10 0.15 - 0.9
1/10 0.9 - 5.4
2/10 5.4 - 32.4
3/10 32.4 - 194.4
4/10 194.4 - 1,166.4
5/10 1,166.4 - 6,998.4
6/10 6,998.4 - 41,990.4
7/10 41,990.4 - 251,942.4
8/10 251,942.4 - 1,511,654.4
9/10 1,511,654.4 - 9,069,926.4
10/10 9,069,926.4 - 0.85 N + 0.15

2014 Halloween Costumes
[ Parent ]
The recent update works in favour of Google (4.45 / 11) (#16)
by Jim Dabell on Sun Oct 20, 2002 at 02:33:49 PM EST

Think about it - they can't claim that Google hand-modified his particular website's ranking, since there was massive upheaval across the board. They might have adjusted the algorithm because of people like him, but he certainly doesn't have the right to force Google to use an algorithm that suits him just so he can make a quick buck.



Yup. (4.54 / 11) (#17)
by mindstrm on Sun Oct 20, 2002 at 05:50:38 PM EST

I've seen this kind of thing before.. though not in court....

I've met a few people who's businesses depend heavily on their google ranking, and when that ranking changes, they complain like google owes them something.

Searchking's business was selling advertising on sites with a high pagerank value. Now those values have all changed, so he feels he was ripped off. Who guaranteed him they would stay? Is there an agreement between him and google (even an implied one?) No, there isn't.
Google does not publish the specifics of how they rank sites for this very reason. They want to provide search results that best serve the people searching, not ones that drive up profits of some site.

Seriously.. I can think of many businesses who would pay HUGE amounts of money for the #1 spot on google for a bunch of search terms.

 

BTW Googlebot is the crawler, not the ranker (3.40 / 5) (#20)
by KWillets on Sun Oct 20, 2002 at 10:03:39 PM EST

I'm a staunch believer that if you have good content, the visitors will come. As your site gets richer and increases in depth, people across the web will willingly link to it, which Googlebot will notice and give you your props for all your hard work.

Now, if the googlebot would leave nice presents when it visits, it might be better than SearchKing.  If a webmaster goes to all the trouble you outline above, he has a right to sell ads to whomever he pleases.  The original PageRank algorithm distributed rank evenly to all the outgoing links on a page, but it looks like now Google is picking and choosing which links are more relevant.  

I don't really care about SearchKing, but it looks like Google is trying to preempt SK's service with paid links of its own, which isn't necessarily nice, and they're going to the "security by obscurity" model for their ranking algorithm, which is at least a partial admission of failure.  

Google has no obligation (4.60 / 5) (#21)
by Mysidia on Sun Oct 20, 2002 at 11:14:51 PM EST

To consider SearchKing's links more relevant to what its users are searching for.

Google's concern is to internet users using the service trying to find relevant information: not sites whose rankings on certain keywords have been pushed through the roof by advertisers who have determined to take advantage of google's particular rating system for their own profit.

If running searches on google turns up irrelevent searchking advertisements, then users get frustrated and look for a search engine that doesn't do that.

In terms of google's paid ads, google is the company running the **** index, ofcourse they are the ones who have the foremost privilege for putting any sort of advertising in results displays.

They also have the privilege of adjusting their ranking algorithms so that searches are more relevant to users' entries. Whatever obligation searchking might think that search engines have to it, it's not greater than the obligation of search engines to provide relevant search results and make their users happy.



-Mysidia the insane @k5
[ Parent ]
And Windows is all about usability (4.00 / 2) (#24)
by KWillets on Mon Oct 21, 2002 at 04:15:37 AM EST

It's fine if google wants to deliver a good search experience, but I sense the beginning of another range war if they feel they have to hide their ranking algorithm.  The original pagerank algorithm was published for everyone to read, and webmasters could take their lumps with some reassurance that they weren't being treated unfairly.  Likewise consumers knew what they were getting.

[ Parent ]
So.. (4.00 / 2) (#31)
by mindstrm on Mon Oct 21, 2002 at 06:59:24 PM EST

Why do you use google if you don't approve of their methods?

Google explains roughly how things work, just not precisely. They are under no obligation to treat anyone "fairly". THey are providing a way of indexing things to provide relevant results.

Can you think of another way to prevent link stuffing than to hide the method?

[ Parent ]

They claim to be objective (5.00 / 1) (#37)
by Mysidia on Sun Oct 27, 2002 at 02:02:58 AM EST

So they are under a sort of obligation to treat everyone fairly.

They are, however, entitled to come up with mechanisms for preventing techniques like link farms, link exchanges, google bombs, etc from heavily effecting their search results, if they like.

The obvious reason that they wouldn't want to release the details is that: if you know the various ways that they detect such things, you can craft a system that still causes the problem but singularly evades each detection mechanism that they have implemented.

In other words: if you know the method, you can be assured of getting around their filters and maximizing your score through technical measures (rather than increased popularity) without having to repeatedly wait for google to reindex, to see how particular things effected your PR.



-Mysidia the insane @k5
[ Parent ]
Security by obscurity? (4.00 / 2) (#30)
by mindstrm on Mon Oct 21, 2002 at 06:56:30 PM EST

It's amazing to me how often this phrase is misused.

Security by obscrutiy only means that if your only method of security is hiding the location of the gaping hole, then you really aren't secure at all.

IF you honestly think secrecy is not part of security, please explain why the Secret Service will not discuss security measures regarding the President. Please explain why the bank doesn't publish it's entire security model for analysis.

Google doesn't publish their ranking information because they don't want people abusing it. If someone DOES figure it out, and uses it, it's not a disaster. No security has been 'breached'.

Oh... what about money lending agencies who keep their selection criteria secret? IS that security through obscurity? No, it's common business sense.

Google runs a search engine; they are concerned with providing relevant results to searchers; that's why they are so popular.
It's up to them how they want to run their site.


[ Parent ]

how is it illegal? (3.00 / 3) (#23)
by dirvish on Mon Oct 21, 2002 at 03:54:42 AM EST

Even if they purposely devalued search king's sites is that illegal. Seems like Google should be able to do whatever it wants with its software. I would certainly prefer they didn't tamper with their software to squash competition, but I don't understand how it is illegal.

Technical Certification Blog, Anti Spam Blog
Google is anti-religion, too! (2.50 / 4) (#27)
by egg troll on Mon Oct 21, 2002 at 01:30:35 PM EST

I know this isn't the first time Google has worked to ruin someone. They've worked very hard to promote criticsm of this inoffensive religion.

He's a bondage fan, a gastronome, a sensualist
Unparalleled for sinister lasciviousness.

HOWTO: Be the biggest whiner (4.00 / 2) (#36)
by erlando on Tue Oct 22, 2002 at 10:47:15 AM EST

Find a commercial company to lean against
This one's easy. Let's choose Google. They have their own algorithm for ranking webpages.

Use said company's non-standard in-house-developed algorithm to base your business on
Make deals with sites with high rankings to carry ads totally out-of-context, e.g gambling ads

Whine and sqeal like a pig when said company changes their algorithm (legally btw) of which you have no control
Do not even once consider if your dot.bomb business-model is sound. Believe in free lunches. The Internet really can make you money overnight.

SUE SUE SUE SUE SUE
Easy. You do live in the US, right? This is the final step towards ultimate whinerhood.

It really does work
Look at SearchKing. Ultimate whiners galore...

SearchKing sues Google | 38 comments (29 topical, 9 editorial, 0 hidden)
Display: Sort:

kuro5hin.org

[XML]
All trademarks and copyrights on this page are owned by their respective companies. The Rest 2000 - Present Kuro5hin.org Inc.
See our legalese page for copyright policies. Please also read our Privacy Policy.
Kuro5hin.org is powered by Free Software, including Apache, Perl, and Linux, The Scoop Engine that runs this site is freely available, under the terms of the GPL.
Need some help? Email help@kuro5hin.org.
My heart's the long stairs.

Powered by Scoop create account | help/FAQ | mission | links | search | IRC | YOU choose the stories!