This is the sixth in what has become an ongoing account on some of the little tricks people try to improve their rankings in the various search engines, particularly the all mighty Google. If you wish to start at the beginning, simply follow the link. It isn't necessary however and you may be wise to skip the majority of part four as I got really off track due to ongoing personal problems stemming from my time as an MBA student.
Although I've been saving up material for the sixth instalment I've also been actively blogging. One category in my blog is "Search Engines", the archive for which may also be of interest to anyone reading this article. Having already covered a lot of the most common techniques to improve your rankings earlier, I've tended to concentrate lately on how my webpage has performed for a few search terms as well as new developments in online marketing and alternative search engines that continue to crop up. One particularly interesting piece of news that isn't slated to be focussed on in this article is the fact the Japanese government has made internet search a priority and is going to throw money at it to try and produce a Japanese champion, something of a Japanese Baidu perhaps.
Here are some recent search engine stories and links I've collected but have yet to write about in my blog. First of all I discovered another interesting meta search engine: Search Tuna. This site will by no means replace Google in fact for everyday searching I can't recommend it. However if there is something you really care about and need to know what are the most authoritative sites on that subject and you are not in a hurry, Search Tuna may be able to help you. It takes a long time to return its results but they are organized well and I was pleased to discover my site was number one for Nurgle in yet another search engine.
Secondly, A List Apart has gotten into the search engine optimization game. Actually they are still more about usability and accessibility and web standards, but as most white hat SEOs will tell so are they. Of course White Hat vs. Black Hat is a contentious topic in and of itself.
Lastly someone has set up the Pepsi Taste Test Challenge of search engine results. They ask you for a keyword or search term and submit it to the big three (Google, Yahoo, MSN) and ask you which set of results you like best. To do this test properly you need to pick a keyword you often search for. I chose Nurgle. I then looked over the results and selected the one that corresponded to Google. This didn't surprise me, Google is still the best and Yahoo is number two. I almost never use MSN and it isn't an anti-Microsoft thing this time, they just don't have better results or innovative features that compel me to use their service.
Some of the more common keyword phrases found in my logs are those belonging to names of MBA classmates. In fact I've even reported which MBA classmates are the most popular at least on my website. No less than four of my classmates have setup their own blogs while they went on exchange. Although I'm not actively competing against them, in fact the opposite I've gone out of my way to link to their blogs, but I was curious to see whether my website or their blog would rank higher for their name. I've decided to use quotations in these searches to improve accuracy.
|Classmate||Their Google Ranking||My Google Ranking||Their Yahoo Ranking||My Yahoo Ranking||Their MSN Ranking||My MSN Ranking|
|Darlene MacNeil||Not Ranked||21||1||2||1||6|
|Chris Kerluke||Not Ranked||1||Not Ranked||1||1||2|
|Gary Lau||Not Ranked||10||22||3||1||75|
Moral of the story: Don't name your kid "Dan Robinson" if you want them to rank well for their own name in search engines. Also I think Blogspot being blocked by the Chinese government at the time I did the searches affected the Google results. I had hoped just by knowing the URL I could verify the rankings but I think Google has removed "blocked sites" from their results for people inside the Great Chinese Firewall. Another thing to note is the latest MSN search index really likes blogs at least BlogSpot blogs as Muschamp.ca got beat for every name with the exception of the Dan Robinson popularity contest in which both Dan and my webpages got blown out of the water.
One area that is not written about as much in SEO circles is how to do better in the various image searches. I think the reason for this is there isn't as much profit in it. The Porn Guys in particular probably don't like people being able to browse their pictures even as thumbnails without first going through their gateway page. To this end any porn webmaster worth the title has probably blocked or in some way limited the image search bots ability to index their site. However for the end user Image Search can be a boon, that is why all the major engines have one now.
As the spiders can't actually see the images they rely on other information to rank them. The ranking used to be heavily based on file name, too heavily based in my opinion. Lately I think Google in particular is giving more weight to the text around the image which is how I would have done it all along. One thought I was struck with is, whether the file size matters? From my own limited use of the image search feature I've noticed that most of the images indexed are relatively small. Furthermore in the case of galleries both the thumbnail and the full size image are indexed. Ideally only one would be indexed and that would be the full sized one. Currently the search engines don't appear to be clever enough to do this. They do not even always successfully rank the full size image ahead of the thumbnail. I'm not exactly sure how I'm going to do it and whether or not it'll be done on my blog or as static pages, but I plan to make some sort of test of this hypothesis. Hopefully it is as clever and effective as my quadrangle juniper test.
Another factor that is often cited as affecting search engine rankings is directory depth. It is a commonly held belief that the deeper your directory structure the worse a page will rank. To this end some people throw all their pages in the root directory. The original purpose of directories in file systems was to help organize files more intelligently and on some operating systems there was a limit to how many files could be in a single directory. Modern operating systems are unlikely to have hard limits that the average webmaster is likely to exceed but thought should still be taken in setting up a website's directory structure.
One school of thought holds that you should keyword stuff in your directory structure. This can get too ugly and shouldn't be too hard for a search engine to flag as gaming the engine. I advocate using directories to ease the webmaster's burden in maintaining and updating a site rather than worrying about cramming in keywords separated by dashes. One popular alternative to deep directory structures is sub-domains. About.com and Blogger.com both make extensive use of this option. I have several on Muschamp.ca. It takes a tiny bit more work than a normal directory but it makes for a cleaner URL and I think bots and more importantly end users like them.
If you've been a regular reader of my ramblings you should remember that the three keyword phrases I'd decided to concentrate on recently are "Sauder School of Business", "Tsinghua School of Economics and Management" and "Muskie" but without using quotations. Beyond the big three I decided to throw in Clusty as it gives a numeric ranking to sites unlike Kartoo. The most interesting of these is the first one because I have competition. Previously on October 18th, 2005 I ranked 83rd for that keyword phrase in Google. I'm now up to 28th.
|Keyword Phrase||Google Ranking||Yahoo Ranking||MSN Ranking||Clusty Ranking|
|Sauder School of Business||28||90||39||26|
|Tsinghua School of Economics and Management||9||37||28||53|
An astute observer would notice that I went up in the rankings in Google for all three terms. So whatever I'm doing must be working, unfortunately I haven't done much. Mainly I've continue to blog which has increased the frequency I've used phrases like Sauder and Tsinghua. I've also used internal links intelligently and that is it. Adding well written content and thoughtful organization of files, directories, and links that is it. That is all there is to SEO apparently.
As I've previously written I've been actively trying to do better in Google Image search and find out how it works. I'm happy to report that many more images from Muschamp.ca have been indexed by Google. However the ones I specifically included to see how important file name was are not currently showing up for the obvious search terms. Rather than dwell on my failures I'll instead give some examples where I and Google have succeeded. I think in addition to increasing the number of images in their database Google as altered their algorithm for the better.
Here are some specific examples of where I've had success. When searching for Marlene Lau you also get to see a picture of just me, as well as several of Marlene and others. In fact since I first noticed this it appears even more images from my website have been indexed by Google, including several that weren't only a few days ago. I held off writing this article until after my last exam. Several images that are in search engine optimization articles are now returned by Google Image Search. I'm sure Marlene will be overjoyed when she learns this: The first eight pictures returned are now from my site, with all but one of them containing Marlene, good work Google.
Searching for Tracy Yang this time also results in images of her being return from Muschamp.ca unlike last time. Searching for Anni Cao results in pictures of her but also results in two pictures that she's not in. In one Wendy Chang and Ines Biedermann are smiling at some party and the other is of Chi-Wei. This is quite perplexing as those files are named "WendyAndInes.jpg" and "Chi-Wei.jpg" respectively.
One of the reasons I began to investigate all this is over on my Nurgle sub-domain I have a lot of pictures and although they are indexed they didn't use to do particularly well in Google Image Search. I haven't had any time at all to work on that portion of my website and I just don't see why I should have to alter my perfectly good file names so I have files called "nurgle.jpg" or "plagueMarine.jpg". I have so many Nurgle and Plague Marine pictures that I have to give them different and more elaborate names. Another reason I can't change the file names is I used Apple's iPhoto to generate many of my galleries and it names the images clever things like "0.jpg". Currently in iPhoto there is no way to change this behaviour. How I discovered all this was repeatedly searching Google Image Search for Nurgle over a long period of time. I found some cool models and links, stuff I couldn't find with the normal Google search but I was dismayed that although I ranked first in Google for the keyword "nurgle" I did so poorly in the image search. I blamed it all on file names and set about to test this theory. I also didn't like the one photo Google would choose from the hundreds on my site to represent nurgle.muschamp.ca
I appear to be doing better, I now have the 28th ranked picture for Nurgle but I would of course like to be number one or at least on the first page. And in my case Google has truly chosen a random image. Further digging reveals another picture from my domain on the third page. Google Image Search is still difficult to predict, probably moreso for competitive keywords.