Are you a Google Analytics enthusiast?
More SEO Content
Getting Database Content Indexed
Posted 17 February 2005 - 01:51 PM
Posted 17 February 2005 - 02:00 PM
Of course, that depends on what you mean by 'searched' .. If a user must fill in a form to find the content, then no, that won't get indexed - Search engine spiders don't fill in forms.
However, there's dynamic content and there's dynamic content. A good dynamic content system will contain pages that appear to search engines (and users) as if they were static when really their content is being fed by a database.
Of course, this may not apply to you, as I have no way of knowing whether your specific situation would lend itself to that.
Posted 17 February 2005 - 02:08 PM
The provisos are that too many parameters confuse the engines, so minimize the parameters, and try to ensure that you minimize the potential for duplicate content, especially through session ID variables creating endless URLs that lead to the same place.
Posted 17 February 2005 - 04:12 PM
Do we have to physically add a link to a resulting URL or can the search engine somehow find it on it's own?
Posted 17 February 2005 - 05:17 PM
Physically add the link, incorporating it into the site. There's lots of ways of doing that, as I'll come to in a moment. The reason for it is simple - spiders don't want to fill in forms because some of them are order forms, so instead they'll stick to chasing links for safety.
Ways of incorporating those dynamic links into the site naturally include the standard navigation. In this way, you create a URL that searches the database for all products of a certain type and voila, there's your Category page. Then create the urls for each product within the category and there are your individual product pages, all easy to include in the navigation of your website, and ready for the spiders to follow and index.
You can also provide deeplinks from the homepage, highlighting this month's special offers, or most popular items, etc. In each case, these are links based on asking the database for exactly that kind of thing, and this helps spiders find those deep-content pages far sooner and more easily.
When you're ready to get more advanced, you can emulate Amazon with a database query to find which other products were most often bought at the same time, allowing for automated upselling and cross-promotions with the old "People who bought this item also bought items a, b and c".
Posted 19 February 2005 - 12:47 PM
If you got a huge database and find it a bit too much to present hundreds of links to your visitors on category pages you can for a brief moment use a semi-dirty trick.
The idea is to present the links to google (& Co) so it learns about your virtual product pages just once. When the links has found its way into the index and is kept valid you can drop the pseudo category page again.
One way of doing this is to sniff out the visiting robot and presenting a nice page with say 50 dynamically created links at a time to the beast. You may wish to include some good keywords from your meta data. When the virtual product pages that the pseudo category points to has been indexed you stop the game before anyone can say the word cloaking twice.
Remember your are not presenting content specific for the robots with the intend to get a better ranking. You are just helping the robots to index interesting content from your database.
Posted 19 February 2005 - 03:38 PM
I don't know if you've actually done this, but it's not something I'd recommend.
A few pages of categorized site maps that are accessible to both users and spiders should accomplish the same thing, I'd say.
Posted 19 February 2005 - 03:52 PM
Really, you don't need hundreds of links. With just 10 links per page, within three tiers you have over 10,000 links. With four tiers you have over 100,000 links.
Posted 19 February 2005 - 05:14 PM
Bruce Clay's approach to the problem has been to take the actual search queries used by site visitors, and create site map pages on the fly that link to the search results. This approach has its own problems, and Bruce's tool has other issues like all the backlinks it creates to his own pages.
If you look at the Gimpsy directory (www.gimpsy.com), it's sort of organized into categories, but what you're really doing is drilling down into search results from their database, and there can be more than one path to find a site. That's another way to approach the problem.
The ideal method, in terms of making a site that's search engine friendly and usable, is probably to bite the bullet and organize the content into a category structure as others have already suggested.
Posted 19 February 2005 - 11:15 PM
I can tell you for a fact that the trick isn't semi anything - its cloaking, and is one of the higher-risk techniques available. The blackest of the blackhat black arts.
To sniff out the robots correctly, you're probably going to use User_agent detection, which means half of the high-level SEOs will be seeing what the spider sees. I have my Firefox setup to identify as Googlebot, as do many other SEOs I know. I can set a user-agent to that of any other major spider almost instantly.
Of course, advanced cloakers use IP addresses instead (since these are far, far harder to 'spoof') but this means you have to be committed to building and maintaining a full list of all possible IP addresses that spiders can use. Even then, any Google engineer could use the same IP to hand-check the pages if anyone were to even suspect there were a reason to do so.
Finally the fatal flaw is that if you cease the cloaking after the pages are indexed, there will no longer be links pointing to those pages, and the pages get dropped from the index as soon as a couple of rounds of PR calculation are done. So you have to keep the links in place, which means you have to keep the cloaking in place too.
If you think being indexed isn't better ranking than not being indexed we differ greatly in our idea of better. I absolutely guarantee you that this excuse will not prevent a ban if the cloaking is discovered. The cloaking will be easy to spot via the Google cache, which if disabled, is a known flag to check out whether cloaking is going on.
Posted 19 February 2005 - 11:23 PM
You guys clarified it very nicely, and made it obvious to anyone reading here that the technique Clintorious espoused, is certainly not the best way to go.
Posted 20 February 2005 - 12:44 AM
New pages in the index that have all their links disappear, disappear as well.
It seems that older, established pages at some point do stay in the index, even if there are no longer links. I've seen this happen on a few sites I worked with where we didn't delete the old page off the server, we just created a new page and changed the links.
We'd get the odd SE referral to those pages every now and then, and upon checking, they were still indexed.
Lesson: If you want the page gone, delete it and let your 404 error handle it.
If you want orphan pages in the index (which is what Clintorious is talking about), they have to be well-established over many months before that happens... once- or twice-indexed pages will quickly be dropped. If you are going to the trouble to organize it to establish the pages, you might as well go ahead and create a permanent sitemap for them.
And FYI- although orphan pages are still in the index, they really don't rank for much of anything except the most obscure of terms.
0 user(s) are reading this topic
0 members, 0 guests, 0 anonymous users