This is part 2 of a 3 part SEO series by Wil Reynolds, founder of Seer Interactive, an industry leading SEO consulting firm. If you didn’t happen to catch part one of his series, you can catch up and have a read here.
So let’s say I have a 9 page web site about widgets
• Homepage (widgets)
• Management team page
• Category Page (all widgets)
• Product page #1 (blue widgets)
• Product page #2 (red widgets)
• About Us
• Contact Us
With that as a given structure for my site, I should first take an inventory of what pages are likely to have content that my audience would be searching for and is important for me to rank well for, i.e. keywords. Pretty quickly I can eliminate pages like management team, about us, and contact us. I would recommend slapping a no follow on those in my navigation, that’s easy. I’d likely have a sitemap that has links to all sections of the site with fully followed links, this will help you rank for management team names and other non critical keywords.
Looking at the remaining pages, I then need to figure out if all of them need to be linked from the homepage (or main nav) to be able to rank well, to start making this assessment just go to Google and type in your desired keyword. If you see over 50% of the results returned are the homepages of your competitors that should help you prioritize which words you may need to target on your homepage as well. For a more scientific way to determine how competitive a keyword may be try the SEOMOZ keyword tool. I like this tool much more than the seochat keyword difficulty tool which doesn’t give you much background into what makes a word more or less difficult than another. If you are already thinking, WOW Wil, you lost me already, I’ve got a link for you to get some background, go read this now then come back.
Let’s say that my big keyword is “widgets”, I might use the homepage and my widgets category page to tackle that term. If I find that blue widgets and red widgets when searched for in Google bring back mostly sub pages in the SERPS, I may be able to architect the link juice flow internally to better help me rank for those words.
My first option now that I’ve eliminated the pages of low value is: Linking the widgets, sitemap, red widgets, and blue widgets pages from the homepage.
In this scenario I am basically taking my homepage value and spreading link juice to 5 pages (all widgets, red, blue, sitemap, and the blog). If I take an arbitrary homepage value of 5, what I have now done is just give each page linked from my homepage a value of 1. You with me? Good.
Creating a hierarchy by only allowing link juice to flow through to 3 pages from the homepage, the sitemap, the blog and the all widgets page. Keeping that same arbitrary value of 5, what I have now done is almost DOUBLED the internal link value I am assigning to my blog and to my all widgets pages, allowing me to target more competitive terms there. What you are likely going to find in this scenario is that the red and blue widget pages can achieve top rankings without being directly linked form the homepage (wasted juice) that may have prevented the blog page and the all widget page from ranking well.
This could be a big mistake…What if the keywords red widgets and blue widgets need more link juice to rank well? My hierarchy has just prevented me from doing that.
Now it’s time to determine how much juice a page needs to be able to compete. This is where the rubber meets the road, as an SEO you need to be playing with terms to see which pages need more juice to rank well. The one thing to consistently keep an eye out for is bounce rates. If you put widgets on the homepage instead of the widgets page you may see a higher ranking, but you’ll also likely see a higher bounce rate. So you need to calculate if getting more traffic at a higher bounce rate good or bad? You may find that it is better to rank lower and drive the user to the 100% correct page, then to drive the user to a page that is CLOSE to what the user is looking for but allows you to rank slightly higher.
Rookie SEO’s too often make the mistake of stuffing too many keyword links in a nav or a footer. What they aren’t realizing is that by linking everything from your homepage you are basically giving the search engines no way to value your internal content as anything other than a homepage or just another page. Obviously some keywords are more important to rank well for than others, you need to be linking your most IMPORTANT and most competitive words directly for your homepage. Create a hierarchy to your site.
Sitting on a panel with Rae Hoffman at IM spring break she mentioned how she allows “juice” to flow from her blogroll to other sites but on the homepage of her blog only, which is a GREAT IDEA for a couple reasons:
1 – She passes link juice to the sites in her blogroll. She has a great site, so that link is a good one to get.
2 – Then she no-follows the blogroll on her blog subpages, which is 100% good for her to not leak juice on every single page of her blog, but on the homepage of her blog (the most important page on a site typically) she is passing juice. This just makes a ton of sense, especially given that getting a link from every page on any site is not that much more valuable than one link from the homepage.
One of the biggest culprits to linking to everything is hidden divs used for nav items, ajax effects, DHTML, and other effects on a single web page. If you have a feature box on your site and that feature box has 10 links, but you have other feature boxes show up what you may not realize is that you are loading your homepage up with a LOT of links. We had a client who at first glance looked like they had about 35 links out from their homepage, but once we asked about the tabs and nav, they had close to 150 links…definitely needs a no follow strategy.
Before you go rushing off to start no following, advanced SEO’s know to test everything, and to evaluate the downsides of everything. This video (also at SEOMOZ) will help you understand how using the nofollow can come back to bite you and have the opposite impact of what you hoped for.
This is an advanced tactic, keep a LOG of everything you are changing and limit the number of things you are testing at one time.