Matt Cutts Gives Insight To Bigdaddy

I dropped my Google sitemaps months ago. I think they are worthless!! I have heard similar things from others.

All I know is MSN is sending me a lot more traffic these days than Google. With the new Windows and IE browser (which will default to MSN) and MSN's ad program coming, Microsoft is looking to be an great shape.

Google better get its shit together fast or it is going to lose a huge share of the market.

 

The timing is perfect for Microsoft, Google is ripe for the picking.

 

I don't remember anyone being this upset with Google since the Florida update. But even that worked itself out in a relatively short amount of time. (couple months vs almost half a year) My stats went up last couple weeks, even though I dropped another 20 places for my main keyword.

 

Loooks good

 

What makes you think they can't tell who is selling links? Remember greed and statistics are on their side. They don't care if you sell one link or buy one link, because that's not going to affect the ranks. You don't think they can detect this?


Matt Cutts says they can:

He goes on to say:

He pretty much confirms here that different links have different weights.

In the more recent post, he corroborates that they are weighting individual links by some trust factor:

And, he goes on to mention exactly what links are given low weights:

And he says they are getting better at weighting links:

Now, a couple words on the TrustRank idea. Just because the original Stanford paper starts with a bunch of "trustworthy" seed sites, that doesn't mean thats what google does. They could also start with with the "Untrustworthy" seed sites and go from there.

Since they've said that other people linking to you can't hurt you. An "untrustworthy" site that links to you doesn't help or hurt. The links are just weighted low enough that they don't count.

However, when you link to an untrustworthy site, then you become less trustworthy yourself. You get their cooties. Linking to a bad neighborhood, etc.

I'm sure Google has a huge database of manually reviewed "bad seeds" and algorithmically detected spam sites.

This Idea of Inverse TrustRank has other SEO implications. If this is the case, instead of pursuing inbound links from high trust sites, you should be avoiding outbound links to low trust sites.

I also think that trust factor is domain based. A clue from here:

It maybe that the neighborhood in "bad neighborhood" is domain based.

How does google detect link buying?

In Two clicks to lesbian porn, Matt talks about how link buyers turn around and also sell links. I think detecting something like this is very consistent with the idea of inverse TrustRank, or untrustworthiness that propagates via backlinks.

I'm not sure I'd get caught up with the idea that off-topic links are bad. As Matt says on the topic:

So off topic outbound links are ok, unless they are to untrustworthy sites.

I should wrap this up by saying that it doesn't look like buying links is any big deal, except that in some cases it may be a bad investment. But it looks like one should be very careful who one sells a link to.

 

Selkirk, I appreciate the time you took to put this reply together, however I am just not sure I trust Mr Cutts.

The last post I read at that blog was detailing some of the "why am I not indexed" questions. His casual responses didn't sit so well with me. The "Oh look at the footer - these links have nothing to do with the site!!!1" make perfect sense from the human point of view, but I don't see that kind of logic being coded into the algorithm, multiple PHD dudes not-withstanding.

Further I think it would be more interesting to see what sites do rank well with "off topic" links in the footer (do they really parse css fully to determine what is exactly the footer anyway? I doubt it.)

"However, when you link to an untrustworthy site, then you become less trustworthy yourself. You get their cooties. Linking to a bad neighborhood, etc."

Alright what defines these terms? As a human it is not hard to say "I know it when I see it", but that is not what we are talking about - we are talking about a machine that is calculating it. If we don't know what a "bad neighborhood" is, it seems kind of pointless to talk about it.

I am frustrated with Google. Mr Cutts "explanations" do nothing but irritate me.

 

While there may be a case to try and detect bad links, that is not the reason behind the current indexing problems.

Of my 6 sites, 4 have more or less disappeared except the home page, and none have this type of 'dubious' linking. One is a site for a holiday rental property I own - the site is more or less unchanged in three years, and has never engaged in even the slightest degree of odd link exchanges, yet most pages have disappeared from the index. Hence I will get less actual bookings for the property if it isn't sorted. Less real, actual, income.

These aren't technical SEO manipulation problems, they are a google error, clear and simple.

 

That is exactly my situation. The only thing I have ever done beyond natural/organic linking and internal linking is a few months of the DO Coop and that ended last fall.

NONE of the "causes" given by Matt Cutts apply to my sites.

The problem is NOT the site. The problem, as Rasputin says, is Google. Period.

 

I just don't what google would gain by disinformation. There are only 10 slots in each SERP. It only takes a handful of nonbelievers to game google. I think preventing search engine spam via disinformation would be a completely unreliable and ineffectual technique.

I think the quality guidelines describe pretty well what google would consider a bad neighborhood. Hidden text, keyword stuffing, javascript redirects, unnatural linking. These are all algorithmically detectable things.

What I'm suggesting is that if you link to a site that does one of these things, then the trustworthiness of your site goes down. If you do it enough, then you will get penalized. In effect, moving into the bad neighborhood.

What I meant by the Inverse TrustRank idea, was that its not just linking to a bad site that lowers your trust, but linking to a site that links to a bad site does too, just not as much.

 

Whilst I appreciate the time you have spent putting together the reply, I am not sure why you have addressed this at me?

Paint it whatever way you like and quote what and whom you like. NO ONE can tell who is selling links, they can ONLY identify patterns that are similar to those of sold/bought links. Which is exactly what I was getting at, and always have said. So you really are trying to preach to the converted .

But to answer your question 'what makes me think google can't tell who are selling links'? The answer is SIMPLE. Because they can not see if money has changed hands, easy as that.

I do not buy nor sell ads, never have, but the fact is that Google might decide that footer links in small text = sold/bought links, and penalise me if I have such links on my site. They are wrong of course, but like I have always said. If you're a dolphin, and you don't want to get caught in the tuna net, do NOT swim with tuna.

With regard trust rank, again we are in agreement, and the article I wrote a couple of months ago finished saying this very thing. I suggested that if there are trusted sites, then there MUST be untrusted sites. This concept is not new though is it. Google have rated bad neighbourhoods for a long time now, and dragging yourself into there by bad linking choices has always been the case.

 

Sure, I agree. I am not saying that he is *trying* to mislead us, but I think that he is misleading us anyway. The explanations he offers do not add up to me and do not seem to fit the empirical evidence.

I do understand what you are saying.

"Hidden text, keyword stuffing, javascript redirects, unnatural linking. These are all algorithmically detectable things."

I do not agree with the last factor. "unnatural linking" is not defined. What is it? If it is buying/selling links then that is not "algorithmically detectable" as mentioned above only the pattern that "may indicate" a purchased / sold link can be detected. And again for a human it is not as hard to make a reasonable guess, but for an algorithm not easy at all and prone to serious mistakes.

If Google is in fact attempting to do this, I think they are in for serious problems. There are either going to be far too many false positives, or it is going to be ineffective. It will simply lead to better spamming techniques - at best.

 

This inverse approach might make a lot more sense, but it's a pity that you've pitched in with your ideas here because emotions are high right now.

I have spammy link exchange sites that are doing well, and other squeaky clean sites that have gone home page only.

The whole thing's completely arbitrary and unfair. It's my view that Matt Cutts has tried to dismiss our problems as being due to "spammy link exchanges" when it's clear to many experienced webmasters that Google have screwed up bigtime.

Cheers,
Paz.

 

Alext, it is easy for an algorithm to detect potential paid links. But as you rightly point out, there is massive room for error. As I have said previously in this thread, Google really do not mind throwning a few babies out with the bathwater. They see collateral damage as par for the course. They TOTALLY do not care, who gets hurt, as long as it doesn't hurt them financially, and as long as the SERP is improved. Simple as that.

 

The only thing Matt Cutt is doing is to take a stick and shove it into the ants nest. I don't take anything he say for real since it just seems he's sent out to minimize and disguise Googles latest screw up.
Here's my suggestion to Google.
Create a web template that all can use. Then we can hace a script where we can post content and if Google doesn't like it they can have a filter to either sort it out or tell us we need to change it before saving it.
We can not put in any meta tags, because that is done by Google. Then we can select from a list to whom we want to link to and if they agree then a link from our site will be visible.
That'll make Google happy and all they will have in their index will be Google approved sites following the Google guidelines!!

 

I had a few points that I wanted to make after reading the whole thread and yours was the last post that I found interesting. Perhaps I could have been better with my context-fu.

 

Thank you for saying what I was fumbling with. You captured just what I think.

With Yahoo and MSN having caught up in the quality search area (or close enough) missing those few good sites that they threw out with their overly agressive (not saying this is what is currently happening btw) techniques will hurt them though. As far as the average surfer is concerned, if Google disapeared over night, everyone would be fine switching to Yahoo or MSN.

 

As crazy as that sounds I think Google is actually trying to do this, f**kin internet Nazis

Unless they develop a super efficient AI that can process trillions of links each week and be atleast 50% accurate at determing bad practices from good what they are doing is completely idiotic, they are fighting a fight they can't win, and in the mean time hurting themselves and others, like someone said the best they can hope for is newer/trickier linking techniques.

They are control crazy, they don't understand that it is only natural to have people who break the rules, but I don't think a cop would ever shoot through the hostage to get to the bank robber

I think the best they can do is revert back to the way serps were before and add a more user friendly reporting system for spammy sites, which would need to be actively promoted much like voting in politics, a stupid bot/algo can never replace a human no matter how good it is.

Emil

 

don't fear the hype

 

An algo is far better than a human IMo.

The BIG problem is that what we want and what Google want are different. WE want our sites to be top, and consider the definition of spam to = 'any site above mine ' <joke>

Google ONLY care about relevance, they ONLY care about the quality of their serps, and they monitor this by a simple Maths calc. If you look at the code of a google SERP, you will see this command after every url

This is effectively telling google what position people are going to to find what they want.

Put MEGA simply this is how google look at it.

Pre algo change the average serch cctr was sat 3.567345 (this being the value of all the clicks based on the info above divided by the volume of searches) They introduce the new algo, and the average position the searcher has to go to drops to 3.564444 Google are happy, because it means
their SERP's are delivering better results. They MUST be more accurate, as people are finding what they want closer to the top of the SERPS.

 

not that it is better search result, just that sites (spam sites?) have managed to get to the top and not that someone found a better site, just that they clicked on a link in the serp's that they thought was the right one.
Maybe Google should check how long people was at the site before they came back and did another search and so on. Then they might be close to what people actually was looking for....