I’ve been playing with Google Trends lately since they added the relative values to the charts. Trends analysis is a very important part of SEO theory, but not just for picking out queries to target for optimization. Glance at these two charts for just a few seconds and then read my comments below.
Google Trends Capture 1
1) free blog 2) free blogs 3) free link 4) free links 5) google pagerank
Google Trends Capture 2
1) google algorithm 2) google update 3) google sandbox 4) google rankings 5) google pagerank
You may have noticed that I cross reference the trends comparisons through “Google PageRank”, a very popular query that hasn’t lost any steam. I don’t know when the PageRank algorithm first appeared in the mainstream media but these charts indicate that should have happened last year. Regardless of how accurate that indication may be, the sudden appearance of “Google PageRank” in Trends’ news section corresponds with the Fall 2007 Google Update.
Since late 2003 (and perhaps earlier) there have been major upheavals in Google’s search results sometime around October/November of each year and sometime in the February-thru-May window as well. These semi-annual Google upheavals are almost predictable. People all over the spectrum suddenly lose rankings, the search results for many popular queries take a nosedive in terms of relevance and quality, and general panic ensues across the SEO community.
These upheavals may or may not correspond with Google’s occasional Toolbar PageRank updates. These are true Algorithmic Events and even if they only represent a convergence of activities within Google the trends charts show us that search activity for certain types of queries does indeed pick up in the late Spring and late Fall of each year.
There are several correlations between these trends charts and historical events that I and many of you have personally lived through. Take the Google Sandbox Effect, for example. It was first documented around March or April of 2004. However, query traffic for “google sandbox” did not hit the Trends-O-Meter until late 2004.
Oddly enough, it was around December 2004 when I first noticed that Google had begun pushing out data every weekend. The search results have almost consistently changed every Sunday/Monday since then with only a few periods of inactivity. I don’t usually detect much change in SERPs over U.S. holiday weekends like Christmas, Thanksgiving, Labor Day, Memorial Day. It’s like the Googlers close up shop and go home.
These queries actually tell us as much about SEO activity as about Google activity through the past four years. Notice the almost frenzied increases for “google update” queries in the second chart (the red line). There have been MANY updates through the past four years (but don’t look for a correlation between this graph line and Google’s 450 algorithmic changes in 2007). I think we can safely conclude that perceived updates are events which don’t necessarily coincide with actual updates.
That is, due to lag times (which must vary) following minor algorithmic changes, there appear to be nexus points where enough changes start to have an impact on search results that the SEO community perceives a change. The occasional Googler head-scratching in response to these perceived updates is understandable because they cannot predict with 100% accuracy what will happen in the live SERPs.
There are several major spike events for “google pagerank” that I think are interesting. The first occurs (in chart 2) around December 2004, when I first noticed the weekly update.
I would have expected a spike for the query in late Spring 2005 — which is about the time that Google finished its worst update process ever. Starting around February of that year a lot of sites dropped out of the index completely. I mean old, trusted sites. Xenite.Org lost something like a third of its traffic in that period and most of our page listings were URL-only listings. But I found many SERPs were deluged with URL-only listings, so a LOT of Web sites were adversely affected.
Among other bad effects, that update period was characterized by an increase in RSS-feed driven spam site listings, by outdated cache data (in some cases the cache images were two years old for many sites), and by the crazy fact that a site that had lost much of its search visibility could add new content that appeared in the index very quickly. I launched several new sections for Xenite.Org between February and May 2005 and while the rest of the site seemed dead to Google the new sections did just fine.
Nonetheless, there was little to no unusual search traffic for “google update” at that time even though many people were complaining about these problems in various SEO forums. Danny Sullivan turned a blind eye to it, disagreeing with me in private email over whether anything was happening. He did document the Fall update process on SearchEngineWatch, and it would be interesting to see if there is a connection between Danny’s acknowledgement of major update events and search activity.
The next spike for “google update” occurs around the late Summer and early Fall of 2005. That was about when Google started dropping a lot of faux directories (the so-called “SEO friendly” directories that had been created solely to provide linking resources for anyone who wanted them). A lot of the RSS-feed driven spam sites also fell out of the index by the end of the year. As I mentioned, Danny Sullivan wrote about this one on SearchEngineWatch.
A third spike for “google update” occurred starting around December 2005 and lasted for several months, approximately equivalent to Google’s Bigdaddy update. Google redesigned its infrastructure and divided the Web into the haves (the Main Web Index) and the wish-they-hads (the Supplemental Results Index). We’ve been living with the legacy of Bigdaddy ever since, even though Google unleashed Google 3.0 in May 2007.
Matt Cutts issued a couple of advance warnings in early and mid-2006 about changes in Google’s filters and Web spam detection. We started to see those changes hit Google’s index in the summer, when the next spike in “google update” queries occurred.
The first spike in 2007 follows upon the Searchology Update, when Google 3.0 gave us Universal Search (Google was following in the footsteps of A9, AOL, and Ask) but another spike occurred late in the year, about the time when Google’s Toolbar PR penalties hit their second wave and people started expressing some concern over what was happening.
If you look back at chart 1, you’ll notice that the “free links” query erupted at the same time as Google’s war on paid links picked up steam in late 2007. A query that had been gradually dying off suddenly received a new lease on life, and I have no doubt that many link spam resource saw a huge upsurge in traffic. The intense decline in search traffic, however, may indicate that people mostly found old resources that had already been burned out. Since the query traffic has not been sustained, the free links industry is still in decline.
But that could also mean that as the paid link brokers have retooled their services and implemented more stringent security measures to keep Google from exploiting their inventories again that people have shifted their trust back toward paid links.
Nonetheless, the quest for free value-passing links has not diminished in the least and it may be that the paid link brokers are facing competition from free blogging services. Now, queries like “free blog” and “free blogs” are not purely SEO queries. A LOT of people (including students and low income families) take advantage of free blogging services, so you have to understand that these trends are not completely trustworthy.
Nonetheless, there has been a significant increase in activity for “free blogs” and absolutely radical increases in activity for “free blog”. I could not include the data in these charts but I looked at queries like “blog farm” and “blog farms” as well (because people continue to visit SEO Theory for those queries) and the activity for blog-related queries continues to rise.
You don’t have to be a genius to see that people are more and more turning to free blogs for linking resources. If you can solve the content problem for yourself you’re pretty much set. Some people hire writers to create content. Some people republish free articles. Some people republish RSS feeds. Some people republish Google and Yahoo! alerts, news headlines, and other easily obtained information.
The blogging activity took off in the second half of 2005, which was when most free link resources were burned by Google and other search engines. I don’t think there can be any doubt about a correlation between blogs and SEO demand for value-passing links. If you look at the second chart, near the bottom you’ll see the light blue line that represents “google algorithm”. Its activity picked up in late 2005, when the free link resources entered their mass extinction phase.
Since late 2005 the SEO community has inundated the Web with misinformation and bad advice about the Google algorithm. “It’s all about links!” has become the battle cry on many blogs. Clueless SEO experts reiterate the cry at every SEO conference, in their books and tutorials, and on their blogs and forums.
We can take quite a few things away from these charts (and other charts that you can create for yourselves). For example, the SEO community reacts quickly to any change at Google, unleashing an overwhelming response that pretty much renders Google’s algorithmic changes obsolete within a matter of months. That’s not to say that people don’t continue to use the old filtered tricks. It just means that there is a huge amount of inertia driving the demand for links.
The link craze hasn’t let up one bit. If anything it has intensified, and these charts make it pretty clear that rather than simplify things for itself Google has only spurred people on to greater innovation. There is absolutely no incentive in today’s search environment for people to create good, link-worthy content because they have all been brainwashed by both Google and the SEO community to simply pursue links.
When I first saw referrals to SEO Theory for “blog farms” last year I was a bit amused, thinking the concept had outlived its usefulness. But I was wrong. Blog farms are more sophisticated and harder to detect now than they were a year ago. It could be argued (and perhaps should be) that Google’s initiatives have forced people to actually create content but I don’t believe all that content is really worthwhile.
People are figuring out ways to create algorithmically acceptable content more quickly solely for the sake of embedding links in that content. If you can realistically write 100 unique blog posts across as many domains, AND embed a link back to your site in all those posts, why should you not do so? You’re the editor giving yourself editorially chosen links.
These queries show us that there are huge numbers of people who are affected enough by what Google does to want to find out more. They want to know about Google updates, the Google algorithm, Google PageRank, free blogs, free links, and a host of other concepts that have — until now — been the private jargon of the SEO industry.
Truth be told, queries for “seo” and “seo services” have been steadily rising, too. Now, that shouldn’t be news to anyone because SEMPO is predicting more demand than supply for SEO technicians for the next several years. But if you look at “seo services” new reference volume, you’ll see that it really took off this year.
Search engine optimization is making the transition to the mainstream right now. Google is already outgunned and overwhelmed by the SEO community. Imagine what the searchscape will look like in a year after all those new people with their mom-and-pop Web sites have learned the fundamental principles of search engine optimization.
And if you don’t think they will pick it up, just watch the trends for link building queries. You’ll see it coming after it’s already happened.
{ 0 comments… add one now }
You must log in to post a comment.