awesome, google released a top 1000 list for all websites (well, all not google-websites, they “forgot’ those) for april 2010.

cool list, nice graphs.

sadly google released the list as an ugly multi page ajax thingie i compiled them in one single shareable spreadsheet (and different formats). have fun.

• Google Docs
• CSV
• HTML
• ODS
• XLS
• TXT
• ugly ajax thingie source

Q: häh? what? how can a company go on a sabbatical?

thats the thing, facesaerch is a cool website that takes google images and displays them in a cool and usefull (some argue that usefull stuff) cover flow like design for faster scanning. its not really a company, its a hobby side project thingie. see more about what facesaerch is and what it isnt in the about section http://www.facesaerch.com/about.php.

and facesaerch is a hobby project were from time to time less then <2 people are working on it. and they are going on a sabbatical in a country where they are not so sure about the internet connection. so if you comment, mail or conntact facesaerch one way or the other there could be the possiblity that we can not respons in real time but with quite a few days in between. so please be patient. please see our aout section for any questions. thx a lot.

see you late
yours sincerly facesaerchteam (hobby side project thingie team)

i spotted this today on the beautiful macbook of my new - now not so favorite - colleague. it’s just sick. it makes me want to vomit. a beautiful nearly untainted macbook pro with a windows 7 installation (os x is completely wiped out on that machine.) it’s just so unreal.

www.facesaerch.com reaches 900.000 visits / 30 days - still hosted on a 1EUR webspace - still a hobby project - still for sale!

more infos about facesaerch.com - what facesaerch.com is, what is isn’t (i.e. we don’t host any pictures, they are all managed via the google API, see more about it in the about section) and it’s still for sale.

currently it’s german only - but they will expand to en, es, fr, nl, dk, sl, po, cz, ru,…. soon.

congrats to my sister for her new driving license. she went to klagenfurt to get her driving license this summer. and well, she nailed it on the first try. congrats to her and fahrschule klagenfurt zobernig for the great achievement. finally i have my own personal taxi driver.

RT @enzenhofer http://ow.ly/fuLl framebusters GA referrer tracking fix

thx a lot

URL shorteners are in (thx to http://twitter.com/), not all url shorteners are nice to websites, a lot of them use frames (which are not cool for SEO) oh yeah, and in some cases the frames even break google analytics tracking code in IE6 (yeah, stupid IE6). that’s why it is a viable option to use so called framebusters.

i.e.: searchengine land uses this simple code

<script type="text/javascript">
<!--
if (top.location!= self.location) {
top.location = self.location.href
}
//-->
</script>

ok, the challenge with this approch is: if you get a lot of refereed inframe traffic you suddenly get a lot direct traffic as this javascript mumbo jumbo looses the original referrer.

how to fix this

there is an undocumented GA function
( thx to http://www.prusak.com/google-ananlytics-initial-referrer-update/ )

pageTracker._setReferrerOverride()

we adapt this framebuster script

<script type="text/javascript">
<!--
if (top.location!= self.location) {
document.cookie = "realreferrer="
+encodeURIComponent(document.referrer)+"; path=/";
top.location = self.location.href
}
else
{
function readCookie(name) {
var nameEQ = name + "=";
var ca = document.cookie.split(";");
for(var i=0;i < ca.length;i++) {
var c = ca[i];
while (c.charAt(0)==” “) c = c.substring(1,c.length);
if (c.indexOf(nameEQ) == 0)
   return c.substring(nameEQ.length,c.length);
}
return “”;
}

var realreferrer = decodeURIComponent(readCookie(”realreferrer”));
document.cookie = “realreferrer=; path=/”;
–>
</script>

oh yeah, and you adapt your tracking code to something like this

<script type="text/javascript">
var pageTracker = _gat._getTracker("UA-XXXXXX-Y");
if (realreferrer.length > 0) {
pageTracker._setReferrerOverride(realreferrer);
}
pageTracker._trackPageview();
</script>

now the real referrer is submited (and then deleted) via cookie to frameless page. where it is submitted to the holy google analytics. job done, traffic referrer stats rule again.

• evil
• stupid

Some links: You can read more about the topic here and here, and a great open letter to amazon and you can follow a livestream of amazonfail at all140.com real time news.

Oh yeah, and Amazon Rank!

<link rel=”canonical” href=”http://www.example.com/url/here” />

So cool.

After implementing it on various properties and checking the use of it on other websites there are a few findings:

• It’s immense powerful!
• For Google it’s basically a HTTP 301 permanent redirect.
• Without the actual redirect.
• It’s easy to implement.
• No design or function change for human users.

After implementation you see first effects about 10 to 14 days later, depending on your indexing speed. If you have a grown, messy site it will give you a big plus SEO wise.

Now the downside:

powerful + easy + lack of feedback  = source of possible devastating errors

The minor errors i have seen:

Wrong markup

• <link  rel=”canonical” content=”http://www.example.org/url/here”>

As SEOs are used to deal with meta tags, we write a lot of stuff in meta-tags (robots, description, …). The canonical element is not an meta-tag, it’s a link-tag (like external CSS or RSS referrers).

The consequence is quite harmless: you implemented something without effect. It only becomes an issue if your implement logic on top of it, or look for a reaction and see none. A popular wordpress theme made this mistake, and i got caught up in a serious twitterfigth by pointing it - impolite (sorry) - out.

Typo that leads to an HTTP 404

• <link rel=”canonical” href=”http://www.example.com/urlh/ere” />

The typo (/urlh/ere instead of /url/here) leads to an error (most common HTTP 404 file not found) page.

Google follows this lead, sees that something is wrong, keeps on continuing his work without any consequences. Again not a biggy until you implement something on top of it or wait for a reaction.

Major errors

Typo that leads to an HTTP 200 or HTTP 301 / HTTP 302

• <link rel=”canonical” href=”http://www.example.com/url/hier” />

If /url/hier exists (HTTP 200) or instead of an HTTP 404 triggers a redirect to somewhere else (i.e. the startpage) you communicated to Google that the page the canonical tag is on placed on is not valid, and another page is the right one. Google does not hesitate to kick the original URL out of the index, and goes to the ‘canonical’ one instead.

If the error is systematically (on a lot of pages) you will butcher your search engine traffic in no time (10 days). The shit thing is: mistakes happen, and especially if you don’t see any real time feedback.

Also if you update your URL structure or your redirect logic, but forget to update your canonical you will run into major issues.

Conclusion: The canonical tag is a present from god Google. Due to the lack of direct feedback also a source of possible devasting errors.

Help: A simple JS Greasemonkey script which shows the canonical tag in the upper right corner (if it is sepcified).