Tonight, as I was writing another blog post, I started working on a graphic that probably deserves a post and discussion of its own. Below is my personal opinion of how some of the key factors in Google’s ranking system have changed through the years I’ve been in SEO:
A brief dissection of the criteria:
- Domain Trust/Authority – I think this was a factor that most SEOs did not seriously feel until after the Florida update (November of 2003), after which it skyrocketed into consciousness, crept along for a bit, and over the last couple of years, has become the dominant factor in the success of rankings at Google. That’s not to say that things like exact-match domain names + lots of anchor text from diverse root domains can’t still overwhelm the occasional page from Wikipedia, Amazon or the BBC, but the preferential treatment has reached new heights.
Just in the last couple days, we’ve started seeing authority sites like Technorati re-ranking for their tag pages on virtually every SERP they target. Google’s “Brand” or “Vince” update also points in this direction, as does the collection of user-data and usage metrics that are potentially being applied or could be leveraged in the future.
_ - Anchor Text in External Links – While this is still a very powerful ranking tactic, it’s not the powerhouse it once was, and before ~2004, I really felt that it didn’t carry the same power it did for those years afterwards. Today, my belief is that anchor text has come to be regarded much as PageRank was after its dominance in the algorithm – as a technique that SEOs have focused on gaming to such an extent that much of it has become noise, and it’s only really valuable when found in conjunction with other positive signals (or at least, this is how Google thinks now, and their algorithm is still moving in transit towards that destination).
_ - On-Page Keyword Usage – There’s little doubt that when I initially started doing SEO, even Google was more susceptible to keyword stuffing. Incidentally, I think this gave rise to the myth of keyword density as a ranking factor (or at least, didn’t help slow it down). Today, it’s not lost all of its ranking power, but it still sits in the middle as an essential element, but one where “more” won’t help you. This is in direct contrast to the other elements I’ve included in this diagram (where more does equal better rankings).
_ - Raw PageRank / Link Juice – In the early days of my SEO career, PageRank was everything (or nearly everything). Manipulating rankings was as easy as getting a few high PageRank links, and this exploit, along with Google’s display of PageRank in the toolbar, built industries of link sellers and buyers we still see today. In 2003, PageRank was already on the decline as a ranking factor – a decline that has continued to this day. My feeling is that now PageRank can still make some difference, but it’s much more effective for Google as a determining factor for inclusion in the index and comparison against duplicates & scrapers.
BTW – When I say “PageRank” I’m referring to the original, egalitarian concept of links as votes, and the idea that every page and every link passes link juice in a similar fashion to help calculate the raw popularity of URLs in Google’s index.
Now I’d really like to hear your thoughts. Where am I right? Wrong? Totally off track? And maybe even a question of where you see these lines heading in the future.
p.s. For more detailed coverage of these factors, see My Personal Opinion – 90% of the Rankings Equation Lies in These 4 Factors (from 2007, but still shockingly relevant)
p.p.s. Yes! We have changed the comments format and the thumbs up/down visuals a bit, so if you have feedback, feel free to leave it here.
