Explore

Ask

Partners

Search Engine Optimization Article

The Logic and Use of Sitemaps for SEO

Site maps (a.k.a site maps, site indices) are often utilized by webmasters as a way to ensure that the most important topics and pages of their websites are included in the Google search engine index and are easily located by Google's freshbots/deepbots (recently combined into freshdeepbot). Typically, a sitemap is a page linked to directly from the index page on a website - and sometimes from all pages within the site. This way, the main pages can be found within just two clicks from the homepage.

A site map can basically be a catalog of simple html links, or it might also include text descriptions alongside the links. Google recommends that webmasters not include more than 100 links on a sitemap. A sitemap will therefore only be able to link to 100 pages or less. If a website requires more than 100 links, the webmaster will need to create a second sitemap. The process of creating more and more sitemaps can eventually get ridiculous, so we usually recommend keeping website link architecture simple and easily accessible, thereby eliminating the necessity for multiple sitemaps within the site.

Let's look at an example. You have a website that utilizes a navigation consisting of 20 main interlinked category pages. These 20 pages are the focal points of your main optimization and your PR efforts. You also have many secondary pages that, due to their keyword and content, are not as competitive; however, you would still like for them to be included in the index and hopefully receive some smaller ranking. In this example, would you need a sitemap? No - you can instead use a small area somewhere on your main pages that will not interfere with the overall look and feel of the pages nor will confuse your visitors.

In our example, you will need 200 links. So, you will only need to place 10 anchor keyword rich html links at the bottom of each of your 20 main category pages! Now all of your 220 pages are indexed and will gain pagerank. How much pagerank you actually decide to distribute to each is always within your control by including or excluding html links to any specific page from other source pages.

Some may not be very familiar with the pagerank algorithm. In its base form, the relationship between two pages is stated thusly: PR (A) = (1-d) + d[(PR{t1})/(C{t1})]. In this equation, d=.85, which is the dampening factor used by Google; PR {t1} means the actual weight of the giving page; C{t1}= the number of links on the giving page. Simply worded, this formula states that when one page links to another page, it passes on 85% of its weight divided by the number of links of that page. Therefore, if a giving page has 100 PR weight points and contains 10 links, the giving page passes 8.5 weight points to each link.

Now, your primary purpose for the sitemap in the first place was to pass along some PR to your important pages. However, what needs to be read into the formula is that with every link a spider follows, the amount of PR passed down to the next recipient is lessened by 15% (since 85% passes on, there is a 15% loss). Because a spider first passes through the sitemap to get to the internal pages, it consequently fives the sitemap a pagerank. So now, each link off the sitemap will receive 15% less PR weight points compared to what they would have if linked directly from the original linking index. It seems that by their very nature as sitemaps, sitemaps are not the optimum or even most logical choice for optimization - they just don't have much on-page content.

Let's look again at the idea of linking with good architecture (placing extra links at the bottom of your main pages. If you do this, because the pages still have to be reached indirectly, these minor pages will receive the same PR points as they would have received from the sitemap model. However, the PR that would have been consumed by the sitemap itself has now been given to your main category pages!

Many times, a webmaster will utilize links pages to attract reciprocal link exchange partners so they can maximize their pagerank to start with. Often, a webmaster will only link with an exchange partner who has a link directly to their links page from their homepage - after all, every webmaster wants his links to be found! But consider another example - let's assume again that you have an index with 100 weight points and 10 links. As we saw above, each link will receive 8.5 weight points. Since 1 of the 10 links is your links page, most of your 8.5 links page's weight points will be sacrificed to your own links exchange partners. ...Or, maybe not!

You will recall that the PR is divided between all links on a given page (up to 100). Therefore, the PR that went to your links page should be divided between your exchange partners and your main navigation, hopefully remaining under 50 per page. For a simple example, let's say there are 10 main pages. Your usual distribution of PR from a links page such as this would be: 1/6 of the PR reinvested into your system, and 5/6 of the PR is passed to your link partners. For the main purpose of PR conservation at this point, it is time to consider the sitemap. What would happen if instead of placing your 10 main page links on the links page, you placed an additional 40 sub-pages on each page of the links directory, bringing you to the limit of 100 links per directory page? Well, now the PR will be split 100 ways, 1/2 going to your link partners and 1/2 getting reinvested into your internal system. That sounds much better than the 1/6 discussed above.

A webmaster with a particularly large site should consider this scenario: one link to the main links.html from the index, 50 links to your links exchange partners, and 50 split between 50 links directory pages (links1, links2....links50.html). Each of the 50 links directory pages will contain unique links to 50 internal sub-pages. This setup will result in 2500 main, secondary, and tertiary pages indexes within 3 clicks from your home page! And don't forget - every links page will recapture 50%% of its PR - that's not a bad deal at all.

20060109