Has the PR update come and gone?

[DreamingEagle]

Originally Posted by obiztek

End result result will be not only good PR but many good sleeps also.

Well said. I could not agree more. I got "filtered" once and went from the #2 position for my travel guide site to disappearing entirely from Google. You would have thought the moon had quit shining in my little world. I fixed the problem and came back six weeks later, but with a firm appreciation of the fact that letting my emotional state wax and wane according to the actions or inactions of some Internet robot called "Google" was a ridiculous way to live.

[Peter (IMC)]

The "Public" update is supposed to happen "once a quarter or every 3 months.

If that is true, they´re late this time. The last visible update of toolbar PR was in the first week of January. We´re now in the second week of march and still nothing.

I did see a huge increase in traffic on a site that I believe has a higher PR than it shows in the toolbar, and it bothers me that the toolbar doesn't update.

IMO - Exchanging links is almost a lost cause these days.

Though I wish that would be true, I have a client that decided to start an inmense link exchange campaign 6 months ago, after I had done just on-site optimization. (I offered link building services to him, but he claimed he found cheaper services for it.) He literally exchanged hundreds of links. In January his ranks jumped up from nowhere to be found to many top10 and some #1 positions.

Though exchanging links probably isn't the most efficient way of link building, it most definitely does work.

[janeth]

Peter,

You seem to be behind the rest of us

We´re now in the second week of march and still nothing.

lol

[Peter (IMC)]

hahaha,. I meant April of course,... can't edit the post anymore,. :(

[ADAM Web Design]

I happened to notice this thread quite by accident (I was in the breakroom and read the rant), and I think I might have at least a semi-reasonable way to solve it.

I apologize in advance for the length of this, but I wanted to post my idea, my prediction, and my logic (which seems to be, in at least some cases throughout this thread, lacking).

Here's the idea, in a nutshell:

1. Pick out a key phrase (or two or three if you want to be really sure) that has a relatively low competition (say, 10,000-100,000 results).
2. Create a single static page optimized for said keyphrase.
3. Do not submit the page to Google, but rather to directories, indices, forums with it in your sig, whatever. Basically, get IBLs for the page, and do so steadily for a long enough period of time to gain an accurate trend analysis (I'd use 3 months myself.)
4. Do not change any aspect of the page (you shouldn't have to if it's optimized anyway, but I digress).

This would create a controlled experiment with only two variables:

1. The number of IBLs Google has to play with and calculate PR from (which is the important variable).
2. The number of results, and the altering of pages and increasing of IBLs and PR of said IBLs to get those results (which you really can't do anything about, but if you run an aggressive enough IBL campaign of your own, then the effect of this behaviour would be relatively nullified.)

Here's my predictions as far as what you'll find:

1. If you get enough good IBLs, you'll end up in Google within 2-3 weeks (in some cases, I've seen exactly 21 hours for new pages, but 2-3 weeks is a conservative estimate.)
2. There may be a few day-over-day drops in ranking for the keyphrase(s) mentioned above, but the ranking for the keyphrase(s) will climb steadily as a whole over the 2-3 month period, as opposed to one massive jump.

tomzo, I took a look at your sites and based on them, I don't see any programming. So I'm going to make the assumption that you're not a programmer. And if you are a programmer, I'm sure there's at least one person reading this thread who isn't, and would find my working theory on how Google works (based on how database-driven sites in general work).

Important database-driven website rule that will come into play later on in this explanation: Open/close the database as few times as possible. In other words, open up your database, read/edit/add/delete your information, close the database, and destroy any objects that may have been used in this process and thus free up the RAM used to open and do your database work.

Now...how does this apply to Google? Well...Google, being a search engine, has a database (or to put it more accurately, a series of them on a series of servers, but for simplicity's sake let's assume it's one gigantic monolithic thing) containing two tables with fields that are structured in a similar manner to the following:

Table 1

The URL of the page indexed
The Page info (meta tags, title, content, if the page is blacklisted, if the domain is blacklisted, if the IP or IP block is blacklisted, etc.)

Table 2

The URL of the page indexed (or another unique identifier joining it with table 1)
The URL of the IBL (or similar unique identifier)
The PR assigned by said IBL.

Now that we've got our structure in place, let's see what happens when a new page is added to Google (or one is updated), step by step:

1. The page contents are read by Google, and the appropriate information is placed into Table 1 (URL, etc.)
2. The links are extracted (which is easy enough to do by finding all the a tags and traversing the href property.)
3. For each link, Table 2 is updated with the new PR from the IBL. This is done one of two ways: in the case of a new IBL, a new record is created. In the case of an existing IBL, the PR is updated.
4. In the case of an updated page, any links that were on the old page and removed from the new one will be in turn removed from Table 2, and their PR will longer be added in.
5. The links from the new (or updated) page are traversed in turn (i.e. Google spiders the links), and Steps 1-4 from above are repeated. In the case of reciprocal links, there are a series of "back-and-forth" iterations between the two pages Google will go through before it eventually stops and says "I've had enough" (I believe this number is 50, but I'm not sure.)

Now, based on those steps above, we can clearly see that the PR would be updated on a continuous basis as Google traverses new (and updated) pages on a continuous basis. If a page were indexed, and then every X number of days/weeks/months, the database record containing the page was reopened and the PR was calculated separately, this would create an extra database call, as well as an extra trip for the spiders to traverse links, and that would be inefficient.

So, by constantly internally updating PR, the PR of a site stays fresh, the IBLs are taken into consideration, and the SERPs stay (relatively) relevant.

Now...why doesn't this information show externally if it's updated on a continuous basis?

The answer to this is relatively simple. By revealing this information, Google provides a means for "get-rich-quick" SEOs to be able to measure their marketing efforts simply by measuring the quality of their campaigns by the increase/decrease/disappearance of PR, thus making an SEO's job easier. And Google, in its rather pandering "don't be evil" approach, will not exactly go out of their way to make an SEO's job easier.

So what's the external PR, and where does it come from?

As we all know, the Google external PR is a 1-10 logarithmic curve designed to show the PR of a site. This would give an online marketer an approximate indicator of whether or not (s)he would be justified in trying to gain either an IBL or a reciprocal link from said website.

In doing so, Google acknowledges that an IBL from an "authority" site is worth more than one from JoeBob's Personal Dancing Jesus and 50,000 Animated GIFs on a Rainbow Background for No Apparent Reason Site. But it doesn't go so far as to say "how much weight", since doing so would simply create chaos among the top sites (which it probably has already, but it would be even worse by showing internal PR).

So...why doesn't this get updated that often? Based on the logic above, it would appear that the external update to the PR would be best served when the internal PR is updated, since it would be the same database call. Or would it?

No. It wouldn't. The difference between internal and external PR is that internal PR is open-ended (since there is nearly no limits to the number of IBLs a site can gain) whereas external PR is closed-ended (0-10), likely based on the internal PR (where the site with the highest internal PR would be 10 and the logarithm would calculate accordingly).

Now...here's where the important database rule comes into play. Let's say for argument's sake that the site with the highest internal PR is amazon.com . Amazon.com is forever gaining IBLs due to its affiliate program, people putting links on journal sites, blogs, news sites, directories, and about a zillion other ways. These links are being traversed by Google all the time. Naturally, this means amazon.com's PR goes up every time a new IBL is discovered as well.

Let's say that Google were to update the external PR of every page in the Google database, based on the new PR of amazon.com, every single time they get a PR increase. A simple search of the word "a" on Google reveals approx. 2,810,000,000 (that's 2.8 BILLION) results. And there are probably more pages out there that don't contain the word "a".

amazon.com gets quite a few IBLs each day. Imagine if they got some combination of IBL PR from new URLs or IBL PR increases from existing URLs totalling 100 per day. (Based on amazon.com's reach, it's realistically a LOT more, but I wanted to make a conservative estimate). That means 100 times a day, over 2.8 billion records would need to have their external PR updated, based on amazon.com's new PR.

For those who don't know, updating that many records at a time is a serious resource drain, no matter what database you're using.

Based on our important databse rule above, we want to do this as little as possible while still keeping things up to date. So what makes sense? An update every 3-4 months seems reasonable enough to get the external update done. It's basically cosmetic at this point anyway.

To summarize:

• Internal PR updated continuously.
• External (hereafter to be referred to by me as cosmetic) PR updated infrequently to save strain on Google hardware.

Again, I'm basing this on how I feel as a programmer of some experience it would make the most sense.

[Peter (IMC)]

Nice "logic" but some practical issues:

Do not submit the page to Google, but rather to directories, indices, forums with it in your sig, whatever. Basically, get IBLs for the page, and do so steadily for a long enough period of time to gain an accurate trend analysis (I'd use 3 months myself.)

From what I understand you created a page inside a website to target a or some keywords. You then want to build backlinks for it. Nice idea, but most directories are not going to accept your page as they want link to sites (home page) and not to specific pages.

Then you ideas on continuous updated internal PR. If Google would be really doing it like that, a google dance would not exist.

Though now a days there don't seem to be specific google dances anymore, in my opinion this is not because of continuous PR updates, but because of periodical updates of different parts of the information being kept on record. Every week a different group of information is updated, and once every so many weeks PR.

Read some information about Google's latest Patent or some interesting speculation on the old Google search engine

[ADAM Web Design]

Those are two good counterpoints, but each of those can easily be explained:

Originally Posted by Peter (IMC)

From what I understand you created a page inside a website to target a or some keywords. You then want to build backlinks for it. Nice idea, but most directories are not going to accept your page as they want link to sites (home page) and not to specific pages.

This is probably true, but I was merely throwing suggestions out for places that might take it. For the purposes of the experiment, it really doesn't matter all that much if directories provide IBLs or not, as long as some pages out there do. Between forum sigs, blogs, online journals, directories, indices, personal websites, news sites, etc., classified ad sites, etc. there are more than enough pages on the Internet to be able to get at least some IBLs.

It would be nice to get IBLs from pages with high PRs, but realistically that's not going to happen. So the counter is to get IBLs from mid-to-low PR pages.

Again though, a good point.

Originally Posted by Peter (IMC)

Then you ideas on continuous updated internal PR. If Google would be really doing it like that, a google dance would not exist.

If Google only had a single database on a single database server, this would be true. But Google has multiple servers (the last time I checked, this number was well over 10,000 but I'm not sure of the approximate count) in multiple locations. With 10,000 or more servers, there are multiple data centers (the last time I checked, 8) as well.

The "Dance", therefore, would be the frequent merging of information compiled within said datacenters, which will vary.

[Peter (IMC)]

:)

You clearly do not know what is/was a google dance..

[ADAM Web Design]

Then please, enlighten me. :)

Actually, I'm very aware of the "traditional" definition of a "Google dance". I simply don't believe it's totally accurate based on the resource consumption that would be employed via the "dance" theory (and for all of the explanations of this, I haven't run across a single explanation I'm prepared to accept as an absolute truth) and my personal experience with databases and replication.

Before you discount my opinion (and no, I'm not posting it as fact either, since like the rest of you I truly don't know since I don't work there), compare it objectively on a purely technical level to that of the traditional Google Dance definition, and consider the resource that would be used in both cases. You know Google will have done so.

This does NOT believe I completely disbelieve it either. I believe there is a "dance" per se. I just don't believe in the prevailing opinion, and think it's been somewhat oversimplified.