Google Blogoscoped

Monday, June 14, 2004

Google It

This is a rude way to point out to people they could have used Google to get their answer.

Recent Blogger Images

I like the way you can see the ten most recent images from LiveJournal, so I wrote this for Google’s blogging framework: see the most recent Blogger images (not always safe for work).

Nigritude Ultramarine Round Two

The second month of the Nigritude Ultramarine competition is still going. The first month was the player award (the iPod), the second month is the stayer award (the big flat screen, which if I win goes to someone who linked to me). So what has happened lately?

The new kid on the block

Anil Dash’s blog entry on Nigritude Ultramarine went straight up as predicted, and firmly landed on the number one position. But not for long as it’s now back on rank 25. This is a phenomenon known as “sandbox effect” (which has nothing to do with backlinks posted to Wiki sandboxes). This theory states Google highly favors fresh links over old ones, but will not attach this bonus permanently to a new page – which would explain the sudden drop of this particular competitor (who like me set out to find proof content is king).

Responses on the web

The Nigritude meme last week kept spreading on the Web and had what I suspect to be its all-time high. This humble blog was covered on Slashdot twice, on (including different languages other than English), and the English Guardian.

For the first time in this blog’s life I realized what it means to be slashdotted – ten-thousand visitors from Slashdot within just some hours. (Good for me I have VisitorVille installed because I could immediately see what was happening on my server.)
But all in all, though I believe this month’s interest (and interest for further competitions of this kind) will ultimately decline, I still get a steady hit flow from Google users who searched for “Nigritude Ultramarine” and briefly end up here.

The Wiki Sandbox strategy which got me in the Slashdot news is still being discussed at the forum.

Current Results for Nigitude Ultramarine

Now when I look at the Google results for Nigritude Ultramarine, I see this site and the Merkey forum trading places constantly. Sometimes I am on number one, sometimes the Merkey forum is. And this change even takes place within the seconds I hit the refresh button. I assume Google employs some sort of quasi-random ranking algorithm when two sites would otherwise rank similar.

Do titles matter?

One thing to note is that I changed my blog homepage title from “Google Blogoscoped - The Nigritude Ultramarine of the 21st Century” back to just “Google Blogoscoped”. Google indexed this new change yet they do not rank me lower than before. Once more we can see how little on-page optimization matters to the position one holds within the SERPs.

Still, a title has merits of its own far beyond the ranking position. A page title is what the Google user will see and click on – or not. Undescriptive, unrelated page titles will not be followed as often as those in tune with the search query. This competition might be a difference, because the page title stands out in the top 10... it’s the only one not using Nigritude Ultramarine at least once.

Word order and Yahoo

Another interesting observation one could make from the top results are that if you enter the single word “Nigritude”, and “Ultramarine” respectively, there is not a lot of change in Google. The same pages are ranking well for the terms used on their own, as well as the terms used in combination. Also, the results for the phrase search “Nigritude Ultramarine” (in quotes) mirrors the one without quotes.

Searching for the reverse “Ultramarine Nigritude” (with or without quotes) on the other hand does not resemble other results. Again we see how much Google treats unquoted searches like phrase searches. If we go to Yahoo, searching for “Ultramarine Nigritude” will result in almost the same as the “Nigritude Ultramarine”. Yahoo still puts emphasis on phrase search with actual quotes.

Google is doing good optimizing for the average “lazy” user (aren’t we all as lazy as our tools allow us to be?) who uses whatever comes to mind to search, then reformulate the query, then search again (most users rather adjust their search terms than check the the second ranking page). Google is using everything they can as indicator on what to deliver back, and this specifically included word position.

In the meantime at MSN

Microsoft’s search contender MSN* in the meantime shows results similar to Google. Anil Dash is on top. Merkey’s number two. Only my blog is missing from the top ten.

*Is that a search engine, evil portal, or plain big ad? I can see lots of animated Coca Cola bubbles covering the left-hand navigation. I guess the MSN search beta test, which has ended in the meantime, is not yet implemented on the front page.

Google Groups, Froogle and others

So who’s first in Google Groups, the Usenet part of Google, and what can you do to score well here? I’m afraid I can answer only the first question. A post titled “Re: GOSPEL MONEY | Nigritude Ultramarine” in newsgroup is on number one. How it got there I don’t know.

Same goes for Froogle’s number one, a $15 product titled “The Nigritude Ultramarine Search Engine Optimization Contest”, being “automatically extracted from web pages. Price and category information are uncertain”.
As you can see, Froogle suspects a shop everywhere.


If you want to see how your web site looks on a different set of browsers (including those you can’t install from, give the free trial a chance. You can choose a set of browsers, operating systems and screen resolutions and provide different URLs. After a while, you will receive an email leading you to a screenshot gallery.

Why TinyURL is Harmful

You may have used TinyURL before. It’s a little Web service which lets you provide a web address, to then give you a shorter version of it. There are other services similar to this.

Why would anyone want to do shorten the web address? Well, some people do not like to post long URLs. Often this is the case for newsgroups. However, you can enclose your link like this:


But why should you not use TinyURL? Because, as Jukka Korpela puts it, “links want to be links” – and URLs want to be URLs.

You probably heard “Click here” is a bad link text. Because you won’t know where you will be taken to, and also, Web tools won’t be able to draw any conclusions from this link (Google would not work as it does if all the world would use “Click here” as link text). At least with “Click here” you can hover your mouse over the link to find out where it might lead.

Obfuscating URLs with TinyURL not only makes it impossible to make first guesses about the web address and where you will be taken to, even when hovering over it. It also comes with the risk of destroying Web archives – like Google Groups, for example. Who knows if TinyURL still exists some years from now? (Even if Google Groups ceases to exist I might be able to extract something from by providing the original URL, or find Web quotes on the site.)

And who knows what TinyURL might turn into? For all we know it might turn into a porn site.
If it sounds unlikely for a service to change so drastically, consider what happened to a research site answering questions online. This site once had researchers put their files on a third-party server, and linked to it for customers. One day, this free file server changed to become a porn site. Now to anyone arriving at the research service, and following up on archived answers and links, it looked like the researchers were pointing to porn sites.

There are many other problems with TinyURL, such as lack of expected link color (visited will not show as visited). But the fact remains it’s nearly impossible to predict all the problems that are caused as soon as you do something online not quite the way it’s intended to be done.
Same holds true for JavaScript – writing a quick hack takes 5 minutes. Supporting this quick hack might be a life-long job... what with new browsers, different settings, and unexpected user behavior breaking it. The easy solution of course is to not use JavaScript for such basic things as linking.

On the Web, only simple means reliable. Introducing complexity, even when it looks like a small hack, never solves a problem without introducing new problems.

Google Tracking Clicks

Whenever you click on a result page on Google, your action will be recorded. This is done by loading an invisible image which exists solely for the Google statistic. The function Google uses has three parameters (the second one just recently being added, as Chedong points out):

function clk(el, ct, cd)
    if (document.images)
        ( new Image() ).src =
                "/url?sa=T&ct=" + ct +
                "&cd=" + cd +
                "&url=" + escape(el.href);
    return true;

And every link on a Google result pages contains this event listener:

onmousedown="return clk(this, 'res' , 7)

"This" is the link link itself to track the URL you click on.
"Res" is speculated to be the content-type and could mean "resource". (I did not find other parameters than "res" in typical results.)
"7" is the position of the result within the ranking.

Other information can be read out by Google as well by using referrer string information, or information stored in cookies.

Is this tracking behavior a bad thing? No, Google has a right to do so in order to provide better results*. And they might after all make use of this information, even when that opens up the results for further misuse (just think of an organization doing nothing but clicking on their pages in the results, thus trying to push them up the ranking).

Google in fact is one of the only search engines implementing this click tracking right; i.e. they leave the actual URL they link to intact, and put scripting on top. Other search engines like Yahoo track more openly. Unfortunately they change the URL they link to**.

*If for some reason you do not want your clicks to be tracked, you can still turn off JavaScript and cookies. Your IP and whatever you enter in the search box will still be tracked by someone, even if you use a proxy like Google-Watch. Webmasters can also track what "you" entered into Google by checking their referrer logs. Note however the "you" is often nothing more than a dynamic IP which doesn't reveal your real name.

**Not directly linking to the right URL brings with it some usability issues. Like the extra step to go to a redirect server, or the fact you can't right-click the link to copy its URL to the clipboard. Services like TinyURL do their part to destroy the usefulness of plain Web links.


Blog  |  Forum     more >> Archive | Feed | Google's blogs | About


This site unofficially covers Google™ and more with some rights reserved. Join our forum!