Image Search

The most important thing for publishers is obviously driving visitors to their website. The Internet is full of articles dealing with this, so I'm not gonna go into great detail here. Suffice it to say that one easy way to make sure Google can index all pages containing individual photographs is to generate them statically. A database driven website may be smaller, cleaner and easier to modify and maintain, but may also cause headaches when it comes to being indexed by search robots. I believe solutions do exist, for instance listing all relevant URLs in a sitemap (unfortunately this doesn't work for Google Image Search), but in the case of my website, generating all pages statically was what I had been doing for a while and did not seem like worth changing.

Now when it comes to photographs, most of the traffic comes not through Google's normal web search, but rather from Google's Image Search. For Hulubei.Net, the statistics for August 2005 look like this:

Temple Of Bacchus, Bekaa Valley, Lebanon
Search Engine Pages Percentage
Google (Images) 62643 92%
Google 3383 4.9%
Yahoo 1124 1.6%
AOL 211 0.3%
MSN 182 0.2%

Sure, some visitors will come through Google's regular search, but as you can see, a good ranking in Google's Image Search index is paramount - that will bring in most of the traffic, and probably most of the revenue. Unfortunately, Google seems to be updating its index very rarely - the following Slashdot article points out to an official admission that they are indeed very slow at that. This is an important fact to keep in mind when making changes to a website. If the site's structure, the location or file names of the photographs and/or enclosing HTML pages change, Google's Image Search index may not catch up with these changes for a long time. Earnings will obviously drop if visitors are no longer able to find the photographs because Google directs them to pages that no longer exist.

Google's slow Image Search index update is obviously no reason for publishers to stop developing their websites. However, if preserving AdSense revenue is important, a redirect from the old location of each photograph to its new location must be provided for as long as necessary. This requirement is greatly alleviated by the fact that Google's Image Search results point to the enclosing HTML page of an image, rather than to the image itself. Once people click on a thumbnail in the search results, Google presents them with another thumbnail and the option of viewing the full-size image (which may have moved) in the top frame, and with the original web page containing the image in the bottom frame. If a redirect to the new location of the web page exists, people will be transparently redirected to the new location. Better yet, code like the one included below can be added to all pages in the album to simply get rid of the small top frame in Google's Image Search results page as soon as the visitor clicks on the thumbnail, making the lack of a redirect for the actual image largely irrelevant.

Batroun Village Club Amphitheatre, Lebanon
function DiscardAlienFrames()
    if (self != top)
        if (document.images)
            top.location.href = location.href;

To see an example of what I am talking about, take a look at the following image search query. Click on the first thumbnail pointing to Hulubei.Net that you see, then try to click on others and notice the difference in behavior.

An important, albeit obscure thing people setting their own domain names should be careful with is that search engines may find both the "www." and the non-"www." names and index them as if they were different sites. This can cause trouble in many ways, the incoming link count in Google's index may be lower than it should be, certain Web APIs (like Google Maps, see Content & Metadata) may not work properly, etc. My solution to the problem was to redirect "" to "" with an Apache Redirect rule, which at least solved the problem with the Google Maps API. Unfortunately, incoming links pointing to are still lurking around. Anyway, the Apache redirect code must be placed in a .htaccess file located at the root of the web server's document hierarchy and should look like this:

RewriteEngine on
RewriteCond %{HTTP_HOST}
RewriteRule ^(.*)$1 [L,R=301]

«Previous    1  2  3  4  5  6  7  8  9  10  11    Next»