DescentBB

A friend has a biz and is being told by AT&T yellow pages that for $170 for 3 months he will get min of 60 hits on his yellow pages ad.

I've heard that no one can get you to the top of Google/Yahoo unless you get a bunch of hits per chance.

Any info or ideas?

or you spend money on google adwords.

Boy, this seems like a crap shoot.

Page Rank is based on who links to you, not on how many people visit your page.

There's a bigger perspective here...

By putting ads out there, you don't directly influence your google ranking. You do, however, create more exposure for your brand, which can result in links to your brand, which, as Jeff250 noted, is what boosts your rank. I should strongly note that not all links are created equal either.

Read Beginners Guide To SEO on SEOmoz for a good understanding of how it works.

It's not only who links you, because that alone is quite easy to abuse. Popularity plays a part - that's why Wikipedia entries figure near the top of almost any search. There are probably a lot of other things as well. I recall that Bing uses several hundred factors (maybe thousand but I think hundred), and I wouldn't be at all surprised if Google uses more.

Although there are various heuristics, I *think* that Page Rank is still their primary algorithm. It's all very hush hush though, so I don't think you can know for certain.

The way I remember it explained to me a year ago (corrections welcome) is to first take the adjacency matrix of the Internet M where M(i,j) is the *fraction* of site i's links that point to site j. Normalize M so that it's stochastic, and then apply some fancy but otherwise negligible noise to M to ensure it has a stationary distribution. Then your rank is your value in the stationary distribution (the matrix's eigenvector with eigenvalue 1).

It's harder to abuse than it might seem. You can't just create a lot of dummy sites that no one links to to link to your site, because they won't have any contribution in the stationary distribution. Popular sites have limited power to abuse since only the *fraction* of their links to another site, not the absolute number, is considered. You have to control a lot of nontrivial nodes before you can abuse Page Rank, but as you pointed out, people are trying to do this. A purely topological ranking algorithm does leave something to be desired... nevertheless, this algorithm is what distinguished Google from other search engines back in the day.

Sirius wrote:It's not only who links you, because that alone is quite easy to abuse. Popularity plays a part - that's why Wikipedia entries figure near the top of almost any search. There are probably a lot of other things as well. I recall that Bing uses several hundred factors (maybe thousand but I think hundred), and I wouldn't be at all surprised if Google uses more.

Links = popularity That means, the more links there are to a page (depending on the anchor text) the more popular google can know a page is.

Like Jeff said, it's not that easy to abuse. I've never heard it put so technically (what the heck is a elgen???), but in the SEO world, they say everyone has link juice, and you can only pass a certain amount of juice around. Not only that, but not all links are created equal. Lets say Wikipedia and MyEncyclopedia.com had a link using the same anchor text to another website, Wikipedia is going to have a much stronger "vote" value in Google's eyes.

Wikipedia is so popular simply because they have multiple pages ranking high with many keywords (PageRank), and they have that because so many people "vote" for them by linking to them. It's all pretty simple once you break it down.

Thenior wrote:what the heck is a elgen???

eigan vectors and eigan values are properties of matricies. It's linear algebra, yo.... pretty cool stuff.

popularity is a factor, but more significant is the length of time until some returns to google after clicking a result.

The "popularity" I referred to was the amount of traffic from Google to that site in general. Flip would be right that they'd also look at how long people stay there (Wikipedia wins pretty big on both counts, come to think of it ). Jeff would be right that we can't know for sure exactly how the algorithm works - it's top-secret for good reason.

Incoming links can indeed be abused if you have a sufficiently large scheme to do so - and I recall hearing that people are actually doing it as well. If a sleazy SEO company has enough customers, they might merely ask them to all link to each other.

Are you sure? It seems like length of time spent at a site would be the most sketchy of all. If I open four results in new tabs, how long am I spending at any of them? Could I not easily lower my competitor's rank by spending no time at his site (and conversely raise my own)? I can't find any reference to this.

Jeff250 wrote:Are you sure? It seems like length of time spent at a site would be the most sketchy of all. If I open four results in new tabs, how long am I spending at any of them? Could I not easily lower my competitor's rank by spending no time at his site (and conversely raise my own)? I can't find any reference to this.

2x I don't mean to sound prideful, but I have to study and keep up with this stuff a lot for my job. I am 99.99% positive length of time isn't a metric.

On top of that, I am pretty confident traffic from google doesn't matter. Think about it - people do a search for a kw, a website appears on top, users click the website, google keeps the website on top - it'd be an endless cycle. Google merely presents what it views as the highest/most relevant based on keyword information and links.

It's also not a given indicator by itself. Nothing is. But what I mean is the following:
Scenario 1: Someone clicks a link from Google's page, clicks "Back" 5 seconds later, and then clicks another link.
Scenario 2: Someone clicks a link from the page and you don't hear from them for a few minutes.

Scenario 2 could mean many things, but 1 looks distinctly like the user didn't find what he/she was looking for. This information could be used to adjust the page's ranking for that search.

Is there any hard evidence that Google actually does this?

Nope.

(At least, not in my possession... can't speak for fliptw.)