Is the importance of backlinks highly overrated?

[dburk]

Originally Posted by Kurt

Hi dburk,

First, my use of "multiple algos" is simply a way to say that Google doesn't treat all pages or queries the same.

If Google did only use one algo, it would be way too easy to reverse engineer?

Now we can use your definition that it is one algo, with many variables applied in different ways. But the end result is the same.

The way you put it, is that a single algo has multiple varibles...But is applied to each and every query/page the same way. My choice of words should imply that not all pages and queries are treated the sane and the concept easier to visualize.

The point is, IMO it's very unlikely that Google treats each search or each page in its index the same.

And if I over-simplify, it's because my method of SEO, which I've used for 13 years, means I don't concentrate on one SEO method or belief. Instead, I try to cover every reasonable variation.

Plus, if you consider that Bing and Yahoo can also give traffic, we can assume there's at least three algos. I have just proven we have to deal with more than one algo. So it seems obvious to me that we should perform SEO with the thought that even if there's just one algo/one variation, then we should be prepared for it to change.

As far as page rank, we don't know the true PR, only the little green bar Google tells us is the PR. What about Google's internal PR?

Actually, many advanced SEOers believe, myself included, that the more popular the search query, the more likely google uses more off-page elements. However, the more obscure the query, the more likely Google uses more "on the page".

This is because there are many searches, estimated between 40-60% of all searches that are so obscure/one of a kind, the engines struggle to return any relevant results.

Compare the SERPS for "computers" and "cheap computer boxes for sale in jackson hole wyoming". Google's going to struggle to find anything close to being relevant to the later. And if it does, I'll bet it uses more "on the page" than off.

Back to reverse engineering...Let's be honest, this is a great security risk for Google. Not only from SEOers, but internally. Their entire fortune rests on protecting their algo.

How would you protect your SERPs from both of the above? I'd do it using multiple algos, because there are many that will return good results, and none are perfect.

I doubt Google does this just like my following example, but the theory is important...

Instead of a single algo, let's say Google has come up with a few, or to please you, different variations of the same algo.

There's no doubt they have tested numerous algo, or multiple variations. Some that rely on On Page. Others on PageRank. Others on Trust Factors. Others on human bahavior.

All these listed can achieve good results, in most cases. So why not use them to protect reverse engineering and from anyone internally from stealing the "secret sauce"?

Let's say you have these basics (plus others):
1. On page
2. PR
3. Human behavior
4. Linking relevance/anchor text

And, all the possible combos.

These aren't random algos/variations. They are the ones that have withstood Google's testing. And they could be variations of the same algo, each giveing more "weight" (priority) to different elements in the algo.

Let's say that Google now has 10 algos, just for conversation. Each is lettered A B C D E F G H I J

Again, each algo A - J has been tested. And while none is perfect, each is good and will bring up a different "flavor" of site.

Now, Google RANDOMLY assigns letters for different queries. For example:

Computers - Uses algos A G D F F E A H C B

Dogs - Uses algos J I D A C H D J B G

With each letter representing a different algo, with different priorities. And each algo would apply to a different results in the SERPs.

For "dogs", the #1 result would use "J", #2 "I", etc.

Assuming each algo was legit in the first place, each would return viable results for "dogs" and "computers", yet use totally different ranking "variations" in different orders. This would make it just about impossible to reverse engineer.

We know my example isn't accurate, as the same page will drop/raise a few positions. However, my point isn't to explain the exact process. It is to show that there are many ways of incorporating different algos, with the end result being to protect their money-maker, their "algo". They may even apply some random elements to their algo, simply to protect it.

Has it ever occured to anyone that if there were truly one method of ranking pages, that the logical outcome would be that the pages would be more similiar? Similar in size, similar in links, similar in content, etc. But they are not? Why?

Now back to why I simplified this...All this SEO doesn't matter. We know there are at least 3 major algos, one for Google, Yahoo and Bing. We know each has changed. We can assume each will change again in the future.

Instead of worrying about all this mess is to concentrate on what the engines must do: Return relvant results. And this will be based on:

Words on a page
Words on pages that link to the page (inlcuding anchor text)
Human actions

This is really it. It's just how they tweak the "knobs".

So instead of worrying about tweaking, knowing every Google update, etc, I find it far better to create as many pages, sites and links, using as many variables as possible.

Instead of using 2% keyword density and early prominence in the page title, I say make a bunch of content that uses 0-10% keyword density. I say use keywords early, late, repeated and not at all in page titles.

Of course, "weigh" your efforts using the most excepted principles of SEO more often, but the end goal is pure numbers:

How many pages use your keywords? How many of those are your's?
How many links to those pages are there? How many are your's?

In other words, set up a site, and try some of the spammy linking methods. Create another and aquire only "quality" links. Create another and use a little of both.

Create pages on sites with authority. Create low-medium quality pages that are high risk to test, (keeping them away from sites that have produced money). Create high content sites for linking and affiliate programs.

I've done SEO for 13 years and virtually every method I've heard about has had people saying it worked for them (at least for a while) and others at the same time saying "That doesn't work".

We can either study and plot according to the lastest SEO discovery, or we can simply make sites and pages that use different blends of what is considered "reasonable" SEO.

PS..Concerning this:


Sorry, but logic doesn't dictate that we use a multiplier, nor does it assume anything is a "zero".

Actually, if you read the forumula for PR, relevance isn't a factor at all. The questions really are, how much is PR a part of the total outcome, and is real PR anything close to what the little green bar says it is?

PR was never meant to be a way to determine relevance. Instead, it was meant to judge a "popularity contest".

And while we're knit-picking about stuff:



Actually, this is a little mis-leading...If you remember in the early days, Hotbot/Inktomi either beat or was tied with Google in every major independant SERP study for relevance. Google's rise had more to do with other marketing factors, as Hotbot was every bit as relevant, and a case could be made, it was more relevant.

And a trip to "back in the day", Google was the easiest SE ever to spam. All you had to do was create a page using the keyword in the anchor text in tons of links....As a matter of fact, I had sites that literally held 18 out of the top 20 spots in Google for some competitive, money making keywords.

But their algo has changed a lot since then.

Hi Kurt,

Thanks for replying.

Now to summarize what I think you said:

• There is more than one search engine each with a unique algorithm
• The definition of an algorithm as a set of rules is false
• SERP listings are based on randomized variables (in other words: pot luck)
• Live data is dynamic
• algorithms are continuously updated
• I don't understand what PR is and it isn't used to weight other factors
• I don't know how to do math or follow logic
• Spread my efforts across many different techniques just in case it works
• You liked Hotbot more than Google and Google's lack of marketing in the early days was a superior strategy to HotBots very aggressive marketing

All this time I thought I was learning, turns out I just got lucky on a consistent basis. Thanks for schooling me. I guess I should not change routines so I don't jinx my luck against those randomized SERPs. Does that make me a supersticious person?

[jayguy]

Quality backlinks are key inyour quest for better ranking. Relevance is key here and whilst link juice is welcome, the right link in the right place develops quality traffic.

[Kurt]

Originally Posted by dburk

Hi Kurt,

Thanks for replying.

Now to summarize what I think you said:

• There is more than one search engine each with a unique algorithm
• The definition of an algorithm as a set of rules is false
• SERP listings are based on randomized variables (in other words: pot luck)
• Live data is dynamic
• algorithms are continuously updated
• I don't understand what PR is and it isn't used to weight other factors
• I don't know how to do math or follow logic
• Spread my efforts across many different techniques just in case it works
• You liked Hotbot more than Google and Google's lack of marketing in the early days was a superior strategy to HotBots very aggressive marketing

All this time I thought I was learning, turns out I just got lucky on a consistent basis. Thanks for schooling me. I guess I should not change routines so I don't jinx my luck against those randomized SERPs. Does that make me a supersticious person?

You demand a complicated answer, saying I over-simplified things, then get sarcastic when you get it. I tried to keep things simple.

Yet which part of my post do you actually disagree with? You didn't even get your juvenile "summary" correct.

I didn't say I liked Hotbot more than Google. If you actually read where I dominated Google in the early days, your "logic" should tell you I liked Google more than Hotbot. I cited independent studies which contradict your statement, not myown opinion. You stated a fact, I stated a contradictory fact.

How do you come up with I said Google "Google's lack of marketing"? My point was just the opposite.

Another point you missed was that I gave a general strategy for changing algos, but you ignored this. What's your strategy? Other than merely criticizing others?

And what did I learn about SEO from your latest post? Or don't you have anything to offer except "relevance! relevance!"? Care to explain how Google determines relevance? Of course Google wants relevance, thanks for stating the obvious.

Another point you twisted, which you seem to have a habit of doing, is that I use different techniques that I "hope" will work. I said use every "reasonable" technique, mixing up things, weighing those that work more heavily. This gives one staying power. It also gives one traffic and multiple rankings.

And if you do everything the same, how do you know if a new linking strategy works or doesn't work? If you use social sites to build link wheels today, what will you do if Google decides it doesn't like link wheels any more?

What do you do if you're into article marketing and one day Google gets tired of all the spun content and downgrades the authority of article sites?

And if one builds an entire "empire" based on one strategy that works now, what happens if Google changes this in the future? Are you saying you are against contigency plans when doing SEO? Sure, if something is working well now, pound away at it until the well runs dry.

But that doesn't mean one can't create tons of other resources that aren't dependant on that strategy and not have sucess with them, either now or in the future. At the very least, these "other" sites are linking assets.

PS. You should have also learned from my post that some advanced SEOers theorize Google doesn't treat popular and unpopular queries the same. This alone will explain a lot of discrepancies in the SERPs. And if it is true, then multiple SEO strategies should be used, if one wants to be most effective.

And if you don't agree, the intelligent thing to do would be to consider it and not discount it simply because you haven't heard of the theory before.

Care to debate this in an adult fashion, assuming you have something of substance to add? Preferably something original, as I've read tons of EO stuff and get bored reading the same things over and over.

[dburk]

Originally Posted by Kurt

You demand a complicated answer, saying I over-simplified things, then get sarcastic when you get it. I tried to keep things simple.

Yet which part of my post do you actually disagree with? You didn't even get your juvenile "summary" correct.

I didn't say I liked Hotbot more than Google. If you actually read where I dominated Google in the early days, your "logic" should tell you I liked Google more than Hotbot. I cited independent studies which contradict your statement, not myown opinion. You stated a fact, I stated a contradictory fact.

How do you come up with I said Google "Google's lack of marketing"? My point was just the opposite.

Another point you missed was that I gave a general strategy for changing algos, but you ignored this. What's your strategy? Other than merely criticizing others?

And what did I learn about SEO from your latest post? Or don't you have anything to offer except "relevance! relevance!"? Care to explain how Google determines relevance? Of course Google wants relevance, thanks for stating the obvious.

Another point you twisted, which you seem to have a habit of doing, is that I use different techniques that I "hope" will work. I said use every "reasonable" technique, mixing up things, weighing those that work more heavily. This gives one staying power. It also gives one traffic and multiple rankings.

And if you do everything the same, how do you know if a new linking strategy works or doesn't work? If you use social sites to build link wheels today, what will you do if Google decides it doesn't like link wheels any more?

What do you do if you're into article marketing and one day Google gets tired of all the spun content and downgrades the authority of article sites?

And if one builds an entire "empire" based on one strategy that works now, what happens if Google changes this in the future? Are you saying you are against contigency plans when doing SEO? Sure, if something is working well now, pound away at it until the well runs dry.

But that doesn't mean one can't create tons of other resources that aren't dependant on that strategy and not have sucess with them, either now or in the future. At the very least, these "other" sites are linking assets.

PS. You should have also learned from my post that some advanced SEOers theorize Google doesn't treat popular and unpopular queries the same. This alone will explain a lot of discrepancies in the SERPs. And if it is true, then multiple SEO strategies should be used, if one wants to be most effective.

And if you don't agree, the intelligent thing to do would be to consider it and not discount it simply because you haven't heard of the theory before.

Care to debate this in an adult fashion, assuming you have something of substance to add? Preferably something original, as I've read tons of EO stuff and get bored reading the same things over and over.

Hi Kurt,

Whoa now, I only meant to make fun of myself, no offense toward you was intended. I had hoped the smileys would have conveyed that, but apparently I failed to make that clear, sorry about that.

You bring up some interesting points. I do agree that Google's algorithm checks a list of pre-compiled data stores (popular searches) and if the keyword is not in this list then does a sub query against a larger data store, compiles this new dataset and stores for future queries.

It appears that each of these small pre-compiled data stores include a maximum of 1000 records and this is likely done this way to facilitate speed and scalability. Since the less popular queries are created on the fly, it's likely that relevancy scores use a less complex algorithm for the sake of speed.

Since the search volume for these unpopular searches would not be sufficient to focus our attention on, I generally don't. Have you had much luck targeting those unpopular queries?

As far as the early years of Google, all I can say is that my memory seems very different then yours, I must be losing it. I seem to recall that Google had virtually no consumer marketing in the first couple of years, relying instead on partnerships with AOL and Yahoo to provide the technology for their massive search volumes. Things really took off for them when they introduced the Google toolbar which went viral. Perhaps I am remembering this all wrong.

Regarding my linking strategies, I have somehow managed to outrank many competing pages with a relatively small number of backlinks even on highly competitve terms. I attribute this to something I learned a fews year ago while studying Jim Boykin's service.

I do test various linking strategies, but realizing that Google has imposed certain constraints on their ranking algorithm for the sake of scalability, it has allowed me to take advantage of this knowledge. I have noticed a few other folks that have come to the same realization, and I like them, have never discussed this in public.

Finally regarding relevance, I have noticed that relevance is a central theme in virtually every service Google provides. It has become apparent to me that they focus on relevance in nearly every facet of their services including their SERP algorithm. Due to massive campaigns designed to manipulate PageRank they have evolved their off-page ranking factors to partially compensate for these manipulation efforts. Not surprisingly, relevance attributes seem to have been incorporated into this evolution.

Most notably to me and few others in the industry has been the innovative way they seem to have adapted PR as a weighting factor rather than a direct ranking factor. PR seems to have an important influence on the relative weight of an IBL but only as it applies to relevance. This technique tends to mitigate much of the current PR manipulation efforts used by the SEO industry.

My own interests in these topics has been to streamline my efforts to what I know works. Nothing is more frustrating than toiling for hours and seeing scant benefit. I have learned the hard way that much of the free advice that circulates on forums is completely useless dogma that has no basis in reality, it just sounds good and keeps getting repeated.

When I can root out something that seems to have potential merit I like to setup a series of tests and add the results to my collection of data. I feel compelled to challenge folks when they are asserting a plausible argument to see if they possess a theory worth testing. My poking and prodding is meant to tease out these revelations. Apply stimulus -> collect data -> analyze results.

If I seem to be picking on you it's probably because I am interested in what you are saying. Please forgive my clumsy methods, no malice intended.