At Thing Towers, we’re all in favour of ranking analysis, particularly when it comes to Google updates and understanding how sites are affected. But Penguin seems to have triggered a lot of speculation about changes that are not based on data at all. Understandably, a lot of people have started analysing rankings, but without using any historical data. Make sure you don’t jump to the wrong conclusions, and if you've cited one of the examples below in connection with Google’s recent updates, it’s time to re-evaluate your data sources.
A frequently cited example is that the search for [something] returns something.com in first place. This is true enough, but when you take into account that something.com has been consistently #1 of Google for “something” since at least 2010, then it’s time to look elsewhere for clues to what Penguin is doing. Here's a graph of rankings over time:
If you check something.com at the internet archive you’ll also see it’s been the same content (one word) for many years. It ranks because “something” is a fairly unusual query, and because it has a mountain of links from people using “something.com” as an example on forums and suchlike.
“Python hosting” strikes
The empty site (with just the word “OK” as content) python-hosting.com ranks on the first page of Google for “python hosting”. It’s another example of Google Penguin favouring exact match domains and not caring about content, right? Wrong. Python-hosting.com has been top 5 in Google for at least 6 months:
Anyone can “make money online”
“Makemoneyonline.com” is another frequently cited example of a site benefiting from Penguin. We didn't have any data for the keyword "make money online" to hand, but when you look at the example site's search visibility over time, it paints a pretty clear picture:
Subsequent Google updates have basically killed this site’s rankings - presumably, what the people posting this example would have liked Google to be doing.
Google is not a relevant “search engine”
Worse still, Google doesn’t even rank itself for [search engine]. Something's up! But anyone in the search world for a reasonable period of time is likely to have come across this one before. Google has never ranked particularly well for “search engine”, although it often jumps around a bit on the low first page:
Google's homepage SEO sucks. They have no content, no keywords, nothing. So it's wrong to even expect them to rank.
There are plenty of other examples doing the rounds – some of them genuine, of course (scraper rankings one of the most concerning examples). But be careful that you don’t use “auto repair manuals” (problem site 10th in Google in January 2012) “mexico real estate” (problem site bouncing on and off first page since at least 2010) and to be blunt, most of the other examples out there.
Yes, these are examples of irrelevant, spammy or plain crazy Google rankings. But if you use them as examples in connection with Penguin or other recent updates, your data is bad – and your conclusions based on that data are going to bad too.
The graphs in the examples here are taken from the good folks of Search Metrics.