Sometimes internal links are enough, sometimes not

Every site has to have some form of internal information architecture; some kind of navigation for visitors to find the information they need. Sometimes this can be set up in a wildly eccentric manner, such as in some flash sites (one site I analysed had a splash page that directed visitors one of two ways – the complete opposite to being search engine friendly), or it can be strategically laid out, designed carefully for both fluent usability and search engine optimisation.

From what I‘ve gathered, “siloing” a website involves sculpting the Google Page Rank trust to filter through the site in a hierarchical fashion, from top tier category pages to their child pages. Dylan discussed the economy of siloing back in March. I’ve also picked up a distinct residue of general opinion leaning toward the view that siloing is not always worth the effort it takes to construct such a map of links and “nofollows”.

My perspective is gradually telling me that what is most important in this case is definitely linking relevance, as this is what Google places most emphasis on most of the time. Of course it helps if the source pages of your internal links carry a high page rank, in fact, I’m sure many webmasters apply the rule with near religious adherence.

While working on one site recently, I have been tossing links and rel=nofollow attributes back and forth like it were child’s play. I haven’t obeyed a strict regime of hierarchical siloing, but rather tried to simply link from very relevant internal pages, some with a little pagerank available. This has caused the site, which is in a competitive field, to squeeze one or two positions up the results in the space of a week or two. I’m hoping for more, but of course, with a slow progression of accruing offsite links, the pages are probably going to stay put for a while.

I came across the desperately troubling test that SEOMoz’s Rand Fiskin performed earlier in the year. The test was rather simplistic in its attempt to prove that only the first link on one page to one other URL counts in Google ranking.

To explain the theory, the anchor text on my second link isn’t going to count as a vote for one page twice. Rands theory was sound, but his test relied on one single site. It’s only when I came across it being tested again by Shaun Anderson, that I was convinced. This algo feature is probably in place to stop blackhats from stuffing a page that has pagerank with dozens of keyword rich links to a targeted page for SEO.

So one page can vote for many different pages with whatever anchors it feels like, and two or more pages can vote for a single page with the same or different text, but one page cannot vote for one other page more than once.

Well, how kind of Google to make all my in-text cross-links that link to important pages completely valueless to those pages. Why is this? Because the navigation links that occur before my cross-links in the HTML sequence have already voted and passed pagerank along. A second link to that URL is devalued, and won’t do anything of more use for the target page.

I work on another site that uses a popular content management system, which totally eliminates the option to place the copy above the navigation in HTML, so my poor cross-links are stranded. Of course, the pages I’m targeting are important enough to find themselves in the main navigation. This sucks, to put it bluntly.

I have done my homework and found a way to “nofollow” the nav links by replacing a few lines of code in the PHP block that generates the nav. This can be done universally, once off, and it will effect the change site-wide, which is nice and easy. The problem is; do you go and “nofollow” all your nav links on a site that has been building trust over time? The site has so many pages indexed and ranked. Would it be wise to “plug” the nav and disallow link juice to flow, except through my own little excuse for a link structure, now, after all the site has been through!?

Well, it’s not my site, so I’m not going to risk losing all the work that’s helped the rankings so far. I should get to work on more creative offsite ideas. I’ve always wanted to understand more about the nature of viral content and more importantly, monetising it. The drawing board awaits me.