Google Warns of Javascript Redirects Crack Down

28 comments
Source Title:
SEO Mistakes: sneaky JavaScript
Story Text:

Matt made a great post on his blog entitled "SEO Mistakes: Sneaky Javascript"

It is good because it goes into a greater depth of explanation than GoogleGuy would have and because of that it is easier to understand for the layman, non technical webmaster.

One thing that interested me though was the final small "weather report"

In fact, I’ll issue a small weather report: I would not recommend using sneaky JavaScript redirects. Your domains might get rained on in the near future.

Matt essentially said that JS redirects DO work at the moment! :D

Comments

If you’re considering

Quote:
If you’re considering scraping content, doing it in the SEO industries is one of the worst places to do it.

Apparently it's ok to scrape content, so long as it is only in certain industries. I take it this means that not all subject matter on the web has is ranked or looked at with the same algorithm. Some special categories, such as SEO, have a different algorithm that google uses to rank them.

Because, of course, Google doesn't give any site a hand job.

(DaveN, Did I just take the piss?)

Actually...does google have

Actually...does google have more than one algo? And depending on which subject matter your site falls into, different algos are used? I actually feel stupid for never considering that before.

More than one.

Of course, Google has more than one algo.

At times I think they have

At times I think they have the JasonD algo and one for everyone else. Normally that algo makes me deliriously happy but sometimes........

ok

back to old fashion ip detection cloacking
/BP

What?

> back to old fashion ip detection cloacking

... BACK to? Where have you been LOL

>>>does google have more than one algo?

Try this;

Block cookies from google.

Do a google search, any search.

Take note of the order of the results and the number of results returned.

Refresh the page (you may have to refresh up to 10 times - it seems to be set up on a random less than 10 basis).

everytime you refresh the page take a look at the number of results returned and the results order (it is also interesting to watch the adsense ads)

Then you can answer your own question about more than one algo.

But instead of algo, I would say index.

Multiple copies of the index in "shards"

Quote:
lots0 wrote:
But instead of algo, I would say index.

Every data center has a copy of the index. One can rightly suggest that each copy may vary slightly from the others. However, Google engineers have stated that the index is too large to be contained in a traditional structure, so they break it up into what they call "shards". Some data may occur in multiple shards, or in only one shard.

So you can virtually see different results on the basis of which shards are used to satisfy a query.

grnidone

You are learning faster than I can teach you ;) Yep that was a piss take :)

...

>>Apparently it's ok to scrape content, so long as it is only in certain industries. I take it this means that not all subject matter on the web has is ranked or looked at with the same algorithm

but then he says

Quote:
people in that industry are much more likely to notice someone using their content.

I'm not arguing the algos exactly the same in all industries but on a purely 'how stupid is this person' front some things are definately less sensible than others :)

But then why would the

But then why would the results change for the same query made from the same location?

And the simple fact there are only two sets of results that are being displayed.

If your theory were correct we should get a different set of results everytime the query was refreshed, not just the two sets of randomly alternating results.

Can you think of a better way to screw up the automated queries?

lots0, results should vary in a multi-part system

Quote:
But then why would the results change for the same query made from the same location?

And the simple fact there are only two sets of results that are being displayed.

Remember, no one location has a single complete index. So, each query is going to be satisfied by whatever collection of resources can be brought together at a moment's notice. It may or may not be the same collection of resources satisfying the previous query.

Quote:
If your theory were correct we should get a different set of results everytime the query was refreshed, not just the two sets of randomly alternating results.

I cannot imagine what you think is a theory [ON EDIT: I see Mikkel jumped in -- sorry, but this thread is too active]. I have only reiterated a fact stated by Google engineers: that their system is so large, they divided the database into multiple segments they call "shards", and that some of the data is duplicated across multiple shards.

That distribution of data addresses multiple issues, including maintaining integrity in the case of hardware failures (they claim that every data center loses one or more hard drives a day and no one ever notices), satisfying a large storage requirement, and allowing for concurrent processes to operate on a large array of machines without having to compete for the same resources all the time.

It's a truly flexible and resilient design scheme. I don't often compliment Google, but based on what they've released about their hardware setup, I think they have achieved some remarkable results.

What's good for the user

I'm gonna stick with doing what's best for the user, just like everyone says.

And I think a JS redirect to a nice landing pg is much better than a free blog or subdomain with ads all over it.

Is the correct answer

>'how stupid is this person' front some things are definately less sensible than others :)

The site he mentions was ranking top 10 for his name last week.

> If your theory were

> If your theory were correct we should get a different set of results everytime the query was refreshed, not just the two sets of randomly alternating results.

This is not theory but facts as publicly presented by Google engineers (live and at a video-stream once). There is really no magic to this.

yeah well

I asked your daughter for a hint NFFC ;)

You misunderstand me

You misunderstand me Mikkel, I am not disputing that google uses index 'shards'. I also understand load balancing.

This has every appearance of being designed to randomly jump between two slightly different sets of results.

Also we have not been able to get the jump between the different sets of results with browsers that do allow google to set cookies. We are still testing, so nothing definite yet but that does seem rather strange, at least to me.

[added] ah to hell with it, I am going to go get all screwed up... Its friday talk to ya all later

This has every appearance of

Quote:
This has every appearance of being designed to randomly jump between two slightly different sets of results.

They've been doing similar things for years. I remember a few years back when they jiggled the results (possibly for the first time?) - ciml demonstrated it to me with his "99 green bottles" pages - they were not as neat as they were before the implementation :)

"Apparently it's ok to

"Apparently it's ok to scrape content, so long as it is only in certain industries"

Either that or you need to brush up on your reading comprehension. The statement is a perfectly clear "stupid people do stupid things" one.

Ok, I feel a little better now...

I don't disagree with a thing you just wrote MM.

Your just missing my point.

Now, I feel better, I am going to go back and fin the job of getting wasted...

Oh ya, MM where did you get the info on googles HDs going out? Is there a link you can share?

Not missing your point, just seeing it differently

lots0, the dual results you report are perfectly fine for me, based on what (little) we know about Google's mode of operation. There may be more to it than can be shown, but there doesn't have to be.

As far as the reference to the hard drives goes, I don't have a specific reference available. I know it's mentioned in the video lecture that Mikkel mentioned, but I believe it has also been mentioned in several media articles (including one I read recently with Matt Cutts on the BBC Web site -- but I don't have it bookmarked so I cannot check).

Hot swapping faulty drives is a step up from hot-swapping faulty sectors and segments on drives, so I'm sure it didn't take Google much effort to figure out how to do it. Everything runs on Linux, so they can modify the operating system if they need to.

Don't matter now

The anomaly has been corrected, no more 'dual' results.

I should know better than to post 'stuff' till after we get done checking it out...

Just for the record I came

Just for the record I came across a network of just over 70,000 sites today hitting (estimated) 100,000 search phrases all deploying JS redirects and quite a few of those sites ranking in G.

Will I miss them once they've gone ? Nope!

MM - if as you say "results

MM - if as you say "results should vary in a multi-part system." Why arn't the results varying now? Could it be you were WRONG again...?

lots0, what am I supposedly wrong about?

As Mikkel confirmed for you, I was only reiterating what Google engineers have publicly stated.

And people do report varying results in the SEO forums. So, maybe instead of being hostile, you should just focus on the facts and not worry about who is reporting them.

jetboy, good find!

That's about the most concise and complete treatment of the Google architecture as I have seen. It reiterates all the points I have seen dropped as asides or sub-topics in other sources.

PhilC actually raises a good

PhilC actually raises a good point on Matt's blog about use of JS redirects in framed pages. Is the algo capable of telling the difference?

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.