Engineering a New World

Microsoft’s newly-branded Bing search seems to be getting faster at indexing & displaying relevant blog content. 

Just a month ago, shortly after the Bing launch, I was doing some tests on how fast Bing was indexing content as compared to Yahoo and Google – especially with respect to blogs. 

I put up a number of blog posts on Bing Maps and KML, and then 1-2 days later, began searching around on Google and Bing for more info on Bing Maps KML implementation.  Within about an hour, Google had already indexed my KML blog posts, and had them available in the results, integrated in to the normal SERPs.  Bing, however, had none of my blog posts indexed, and after a week still had none of my posts indexed.  Now, these were posts I also went and threw links to on forums and all over the web, as I was trying to get a specific question answered – so both Google and Bing had spiders all over my site basically hourly.  However, 2 weeks later, I was searching for more info on Bing Maps and KML, and in Google I almost couldn’t get anything but my blog posts – whereas Bing still was not indexing it.  They’re showing now.

Well, now the situation appears to have changed.  I made a post on one of my other blogs about a new Dianetics DVD that came out.    I haven’t really linked this one from anywhere, so I thought it’d be interesting to see how they were showing up. 

Oddly enough, in both Bing and Google, my posts are showing up on the first page of results when searching for “new dianetics dvd”.   Bing was indexing & showing blog post results after only about 10 hours of having been posted.

In any case, I thought it an interesting observation – and one that’s probably going to get me to use Bing more – as when trying to solve programming problems, the answers are commonly on forums and blogs, something that was getting a poor hit rate on Bing but showing relevant results in Google.

I just noticed that Wordpress comments are marked as “rel=external nofollow”.  I got curious as to what that meant for SEO, as a tactic for SEO’s (and for spammers) has commonly been to deluge blogs with comments, hoping to thereby increase link popularity.

The best explanation I’ve found for this is on the Microformats Wiki, which explains the subtle differences between “nofollow”, “external” and the other “rel=” attributes.

In terms of what actual value and/or impact this would then have on the SEO value of comments in blogs, this is another one that’s a bit up to speculation.

One can first-off point out that an “external nofollow” tag would thereby mean that a link is meant to be marked as an external link that should not be followed and cataloged by any search engine.

However, there’s been rumblings all over the net that Yahoo! and Bing don’t give a damn about nofollow links, and index them anyhow.

So, unless you’re in the know and actually work for one of these search giants, chances are you’re going to have to rely on making your decision the old-hat SEO way – by empirical observation – and then determine what the impact really is.

As a follow-up to the last post I did on foreign-language SEO, here’s an interesting comparison on foreign SEO with respect to page titles, URI paths, and browser-based translation of non-7-bit-ASCII titles.

Take one phrase, such as Googling for “common cold” in various languages:

Japanese:

Googling for “感冒” in google.jp gives you 2 results in the top 10 that directly link to Japanese characters in the URL itself (which will of course be rewritten in the browser), and 3 of the top 4 had the phrase in Japanese chars in the title, all of them in the snippet.

Interestingly, the #1 result has a decoded UTF-8 version of the characters as the URL:

d.hatena.ne.jp/keyword/%B4%B6%CB%C1

Russian:

Googling for “простудой” in google.ru gives you zero results that have Russian characters in the first 3 pages of results. Most are CMS-driven sites that have letters & numbers in the URLs themselves.

All of the top 10 results, though, have the keyword in Russian chars in the subject and most have it in the snippet as well.

Greek:

Similar to Russian, most of the top 10 results for “κοινο κρυολογημα” are CMS-type sites with the keyword in the title and description, but no keyword or transliteration of the keyword apparent in the URL itself.

German:

The word I have for common cold in German is “Erkältung”. May not be right, but it serves me well as it has an a(umlaut) in the spelling.

Googling for this in Google.de gives a very interesting result. The number 1 result, a Wikipedia entry, is for de.wikipedia.org/wiki/Erkältung — with the a(umlaut) in the display URL. The #4 result has this as well — www.gesundheitpro.de/Erkältung.

However, the #3, #5, and #10 results utilize the transliterated “ae” spelling in the title – www.erkaeltung-online.de andwww.netdoktor.de/krankheiten/fakta/erkaeltung.htm andwww.aspirin.de/erkaeltung/index.html — which leads one to believe that it’s just as effective, in the European languages, to use the transliterated version of these characters in SEO-sensitive elements like the filename, title and description.

I’m not sure which conclusions this draws me to, but it’s data. Anyone have insight or suggestions?

Technorati Profile

A passion of my life for some time has been in figuring out the details of foreign-language and foreign-character-set SEO.  How do you do Search Engine Optimization for foreign character sets – and specifically SEO on languages that do not use traditional roman characters, but instead use Cyrillic, Kanji, Mandarin or Greek characters? 

SEO is getting to be more and more a normal thing to do, and less and less of a hidden black art.  Google has made it plain enough times that what they want is good, fresh, updated, relevant content,  and not a bunch of garbage.

Pursuant to that, you’ve got a ton of fairly-well-documented best-practices for SEO’ing your site.  And, if you don’t know the first thing about SEO at all — well — read a good book on the subect.  My favourites are:

• Search Engine Marketing, Inc. – Bill Hunt / Mike Moran
• Search Engine Optimization an Hour a Day

Or you can just hit SEOMoz or SEOBook for some hot tips.

But one unfortunate thing is that most of the best SEO data is coming people who are ignorant Americans like me.  Despite my love of geography and far-off places, I can speak no foreign languages fluently, except for some Korean bad words I learned from fellow soccer players.

What does that have to do with anything?

Take the preceding picture I just linked to where I’m doing a soccer throw-in.

Assuming you could edit that page, if you ask any search engine novice to optimize that page to show up well for its subject matter, they’d probably tell you to hit the easy things first.  They’d tell you to optimize:

- HTML <title> tag
- <meta description=> tag
- <meta keywords=> tag
- <H1> text
- Body text
- text of inbound links
- filename of the page

Ideally your page would have “Soccer Throw-In” or a more unique title and <h1> text, and would have a description and set of meta keywords that followed along.  Ideally, as well, you’d have a filename like “/soccer-throw-in.html” or similar.

Easy, right?  Of course it is — in English.

But, let’s say you have similar items in German, or worse, Japanese, Greek and Russian!

As an example, the Japanese word for “soccer” is “サッカー“.  What do you make as the page title for that?  The filename?

If you do a google.jp search for “サッカー“, one of the first results you get is a Wikipedia article for “サッカー” which has a displayed URL of:

Now, of course, anyone with any technical sense will tell you that you can’t put non 7-bit ASCII URLs into an HTTP request, as that violates the spec.

But of course, pasting such a URL into your browser automatically decodes it to:

http://ja.wikipedia.org/wiki/%E3%82%B5%E3%83%83%E3%82%AB%E3%83%BC

So, it has the benefit of (a) showing up with the proper Japanese term in the search engine result page, improving the apparent relevance of the result, and (b) well showing up at all in the top 10 listings at all — so you’d think it has SOME positive impact in ranking.

European terms are much easier, as there are common transliterations for many of the non-7-bit-ASCII characters that one would use in normal usage.

For example, Google for the beautiful German city of Düsseldorf.  Clearly, one wouldn’t want to have to title all one’s pages as “Dusseldorf” as that would mean “village of idiots” as opposed to Düsseldorf which refers to the small tributary of the River Rhine.   The u umlaut is easily transliterated to “ue” generally, so by Googling for “Duesseldorf” you get an acceptable result – as Google knows what you’re talking about.

Not so easy with these other languages like Greek, Hebrew, Hindi, etc.

I’m very interested for any input or feedback on this, as it’s a massive gray area right now — and I don’t know if ANYONE has this one covered well.

Technorati Profile