SEO proves that Google is using extensive User Behavioral Data as Website Ranking Criteria

I’ve been trying to convince webmasters for so long that Google is very smart and they are tracking user activity on the search index and also on your web pages, but in most cases people laughed at me and said that’s stupid. Guess what, finally a SEO did close monitored experiment and proved that this is true!

First lets start a bit with some background information that I guess some of you already know (whoever reads my blogs and rants on SEO). I’ve posted numerous articles and threads on different forums asking people what they think about Google’s Search Results and especially on the “onClick” javascript that monitors what Users are clicking on the search results page. In most cases people laughed at me saying that if Google is tracking this activity then a lot of people will be able to alter the Search Results and push-up or pull-down websites intentionally. Finally I got the proof on this!

The second thing that really bothered quite a lot, and also being widely discussed on different forums, is the fact that suddenly few years ago Google started offering Free Urchin Analytics Accounts which previously were paid around $300-$400 per year per website! As a businessman I really got amazed how Google suddenly became so generous and offered those really pricy services for free, and I started suspecting that they are doing some serious data mining with all that information collected. Guess what, I proved to be right, Google indeed monitors analytics accounts and website performance and its even adjusting website rankings based on that data!

The third thing that I was slapping in everybody’s faces was my readings from the latest Google Patent Application where I even labeled “VERY IMPORTANT :: USER BEHAVIOUR” in my analysis. Read the whole Google Patent text here.

Now let’s say what the proof is and what was done exactly.

Randy from 1st Rankings, just released the official data on his experiments on Google and User Behaviour Data in Rankings. He conducted two very interesting experiments which clearly show what I was trying to explain to all of you so many times.

Experiment 1 – tracking user activity without Google Analytics

Randy set up two closely monitored websites and asked 100 volunteers to search in Google for some pre-defined keywords and click on the website, then stay in the website for some time and then close the browser without returning in Google. This scenario is supposed to show Google that the User is looking for quality website in those specific keywords and the choice they did (clicking on that website) was that this website is what they were looking for. The situation is that when use clicks on a website from the search results, and then stays for some period, this period can determine how relevant this website was to the specific query keywords. If the period is short and the User returns and clicks other website, this most probably means that he/she did not found what they were looking for and the keep on looking until they find it. On the other hand, if they find a website and spend a lot of time there and even exit the site (not returning back to Google) then this most probably signals that they have found what they were looking for, and in such case, Google can award this website by increasing its ranking. This is often done when the top results are irrelevant to the query and Google automatically corrects its rankings based on the user’s judgment.

He set up two websites, one in non-popular niche and asked those 100 volunteers to click and the website, stay there for few minutes and close the browser. This signals Google that the quality is high and this website is awarded with more attention. His experiment showed that in about week, this website started moving up from position 10 to number 2.

Also he tested his theory in more competitive niche, using the same strategy, but it proved to be quite useless because the changing factor was negligible.

Experiment 2 – Tracking websites using Google Analytics

This experiment was very similar technique to the previous one, but this time he has all the websites participating being monitored by Google Analytics. Also in this scenario, he asked the volunteers to act like they are reading the website and browse throughout the website for some period.

The results proved to be astonishing. In about week and a half, the website popped up from position 80 to about 33 and maybe even more if they continued the experiment (the time frame for the experiment was exactly two weeks). This clearly shows that Google is indeed using the data collected from Google Analytics and adjusts its ranking accordingly to the User Behavior. When users find what they are looking for, the bounce rate decreases, and the time spent on the website increases. This is extremely valuable information to Google, as it’s the primary website quality characteristic.

What do we learn from that

• Yes, Google is monitoring User Behavior and that can adjust your rankings. The user behavior importance for ranking is very low, but still it is considered.
• Google monitors clicks on the Search Results page, and therefore Website Ranking Checking Software can hurt your rankings or at least won’t help.
• Google does monitor your Analytics account and if you have high bounce rate (mostly for MFA and irrelevant websites) you might get punished and loose some ranking positions. If your bounce rate is higher than 75% you should remove Analytics from that website or try to troubleshoot it and lower that rate.
• More content like articles and news on your website can help you increase the time spent and this can help you rank better for both SEO and Google user behavior tracking purposes.
• Now its really debatable if you should even use Analytics, but I believe it’s a fair price to pay for monitoring your websites so if you have good content, you should not worry about that.
• The recent acquisition of FeedBurner from Google might lead to another data mining situation when Google might start looking at your readers, but I guess that will be used in future upgrades on their algorithms.

Google can and maybe already is looking at your website quality using these monitoring techniques to price your AdSense ads and that might hurt a bit Advertising Publishers and MFA’s. We already felt last week’s changes in PPC prices based on web quality both from the Yahoo Publisher Network and Google AdSense. I’ll write on that a bit later when I gather more info.