Google FAQ for webmasters

IMPORTANT

As of the May dance/update, Google has been making significant changes to it's algorithms. Many pages that used to show a Toolbar PR value are now showing PR0 (white) and no PR (grey), but the pages have not been penalized in any way and are ranking as normal. Also, the number of inbound links into many websites, as shown by Google, has been slashed dramatically. This has caused many pages to show a reduction in their Toolbar PR values, but the rankings don't appear to have been affected.

The entire web (the part that is in Google's index) is affected by the changes, so there is no need to feel singled out. The best advice is to sit tight and wait until the changes to Google settle down.

Why does a lower PageRank page beat mine?

PageRank on its own does not decide a page's ranking in the search results. If it did, we would have the situation where all PR10 pages would be displayed at the top of the search results, regardless of the search term used. All PR9 pages would follow those, and so on.

Google uses dozens of factors to determine the rankings, such as Title text, body text, inbound link text and PageRank. PageRank isn't even the most important factor. It is often the case that pages beat higher PageRank pages because they are better optimized across the various factors for the particular search term.

One important reason why an on-topic page could beat an on-topic page with a higher PageRank is the inbound link text. Google assigns link text to the receiving page. It is as though the link text is on the receiving page, and it views it as being important to the receiving page. It may be that the link text on most or all of the pages that link to the lower PageRank page includes the actual search term, whereas most or all of the link text on the pages that link to the higher PageRank page may not include the search term.

Inbound links tell Google how important the page is; inbound link text tells Google what the page is important about - from the linking pages' viewpoint.

The fact that a page has a higher PageRank doesn't mean that it has more inbound links than a lower PageRank page. It may only have a few, all from high PageRank pages, whereas the lower PageRank page could have a lot more inbound links, but from lower PageRank pages. The lower PageRank page could have a far greater potential for search term rich inbound links, and could beat a higher PageRank page on account of them.

There are many reasons why pages beat higher PageRank pages in the search results. The important thing to realize is that PageRank on its own doesn't determine rankings.

Why has my PageRank gone down?

When a page's PageRank decreases it is easy and common to assume that Google has issued some sort of penalty. In many cases this is true, but in many other cases, it isn't. Here are some of the reasons why it happens.

Increase in Google's index

This is natural, and yet it is the least known reason for PageRank reductions. The page's PageRank doesn't go down; what goes down is the Toolbar's representation of it.

Google divides the number of pages in its index into 10 sections, and each section is represted by a number, between 1 and 10, in the Toolbar. For simplicity, let's assume that Google has 100 pages in it's index, and let's assume that Google divides the total number of pages (100) into 10 equal parts. Imagine that the pages are sorted in numerical order according to their PageRanks. For the first 10 pages, the Toolbar will show "1", for next 10, it will show "2", and so on. This is where the Toolbar figure comes from. It isn't actual PageRank; it is a representation of which PageRank division a page is in.

Now let's double the size of Googles index by adding 100 new pages to it. The new pages will have their PageRanks calculated and they will take their place in the numerical list. The total number of pages is divided by the 10 Toolbar section, and we find that there are now 20 pages per section. The point here is that the old pages that were in the old 11 to 20 (PR2) section, now find themselves in the 1 to 20 (PR1) section. Their actual PageRanks haven't been reduced but they have moved down a division because of the addition of all those new pages. That is often the reason why a page's PageRank drops a Toolbar point. We shouldn't really refer to the Toolbar figure as PageRank because it isn't. It's a division, and that's all.

In reality, Google divides the number of pages on a log scale (or similar), and not into equal parts. Also, the index grows in much smaller increases, so pages dropping down a division is common but not as widespread as a doubling of the index would be.

Penalties

When a page drops several Toolbar points, often to PR0, it is the result of a penalty, though not necessarily a penalty that has been applied to the page itself or even to the website. Penalties are applied when Google finds things about a website or page that they don't like. E.g. involvement in link farms, doorway pages, multiple domain techniques, hidden text, cloaking, etc.

Link penalties: Google's view is that webmasters cannot control which websites and pages link to their sites, but they can control where their own sites links to. Linking to a penalized page or site, or a page or site that is later penalized, can attract a penalty for the linking site.

Other reasons

It is possible to find a reduction in the Toolbar figure following the update after making internal link changes. I.e. by changing linkages, it is possible to divert the PageRank within the site to different pages. See PageRank Explained for information about PageRank and internal linkages.

Why have my pages disappeared from the search results?

There are a number of reasons why pages disappear from Google's search results. Here are some of them:-

It's a glitch

Occassionally pages, and whole websites, vanish from the search results following an update, but are restored after the next update. Google continually tweak the way they evaluate and rank pages and, sometimes, glitches occur. Sites and pages even get a PR0 as the result of a glitch, although usually PR0 indicates a penalty. Glitches are usually corrected at the next update.

Fresh Pages

New pages are given temporary evaluations and placed (ranked) in the search results. These are temporary rankings and a very unstable. For a new page to ranked correctly, it needs to be in Google's main index. For that to happen the page needs to be crawled in Google Main Crawl, and then go through the following update. If a new page has only been Fresh Crawled (see Fresh Crawl), its rankings in the search results are not based on a full evaluation of the page, and are very volatile. They can appear, disappear, change position, and can even have other pages of the website taking their place for a while.

Until the page has been Main Crawled and undergone the following update, top rankings are nice but should be taken with a pinch of salt. When the page finally enters the main index, the rankings it gets may be very much lower than its "fresh" rankings.

It's a penalty

Googles sometimes issues penalties to web pages and sites. Occassionally, the penalty is for the site to be removed from the index. In this case, the Toolbar PageRank indicator shows gray. Other penalties include the PR0 penlty, which is when a page's PagRank is set at 0 (zero). The penalty is sometimes site-wide. PR0ed pages are not removed from the index, and still show in the search results, but they are usually buried and give the appearance of disappearing altogether.

How can I tell when the Google Dance has started?

Google has 9 searchable data centers. They are:-

www-ex.google.com
www-sj.google.com
www-va.google.com
www-dc.google.com
www-fi.google.com
www-ab.google.com
www-in.google.com
www-zu.google.com
www-cw.google.com

NOTE: Since writing this article, Google have made big changes to their datacenters, so that the above URLs are no longer valid. For information on the latest datacenters, click here.

In normal times, each of these data centers shows the same number of backlinks (inbound links) to any web page, but when the dance begins, one of the data centers shows a different number to the rest. Because Yahoo! has a vast number of backlinks, yahoo.com is the web page that people usually check to know when a Google Dance has started. When the number of backlinks to yahoo.com is different on one of the data centers, the dance has begun.

To check if a dance has started (if any of the data centers shows a different number of backlinks), type link:yahoo.com into the search box on each of the data centers and see how many Results each of them shows.