We are all data. Whether we like it or not in today’s internet user data equals money, and so we are all profiled based on our browsing habits and our data is used to serve us future search results, show us new adds, suggest new friends and contacts we should make… ultimately the goal is to drive action, since companies are willing to pay for those actions, clicks, sign ups, sales… Actually the only reason why so many services can be offered to us for free is because companies who do offer free services can profit from the data they collect from us while we use those services service, whether it’s directly or by selling that data to third parties.
Now if you are ok with Facebook knowing who your friend are, Microsoft knowing how clumsy you are with the Office package (despite what your CV says) and Google basically knowing what your wishes are before even you do yourself is a debate for another time. However, accepting that user data is so powerful these days brings up another interesting question for us webmaster, which is; Does Google use Analytics metrics as SEO signals?
Why would Google want to use GA data as SEO signals?
If you think about it Google using GA metrics as SEO signals to rank a site higher or lower on the search results page would make a lot of sense. The only thing Google cares (or at least that’s what they tell us) is serving the most relevant result to each search query for the specific user who is performing that search.
In the beginning of times (Google time that is) this was determinate by objective factors; Number of back links, anchor text of those back links, title tags, keyword density, etc. However as Google’s algorithm got refined they started to introduce more subjective signals too which has lead to search results being different for each one of us based on things such as our geographical location, the time of the day, or our browsing habits.
Now imagine how much subjective data can be find in your Google Analytics! Bounce rates, time spent on the site, time spent on each particular page of the site, page views per visit… and all that can be filtered by country, city, month of the year, week of the month, time of the day… in other words, a gold mine! Google could apply these number to its search algorithm and serve even more refined results to each particular user.
The Official Answer?
So Google is using this data then, right?
Well the official answer from Google is NO. Google has always claimed (and still does) that they do not use GA data as SEO signals. If you don’t trust me and want to hear it for yourself watch this short video
Even though Google’s official answer is NO there are still plenty of rumors out there of people claiming that GA data does indeed affect SEO. Many even claim to have performed experiments that prove that by having sites using GA and others using a different analytics software and studying the differences in ranking.
If you want to read more about one of these theories you can check out Michael Gray’s article on the topic.
What’s most likely happening?
So here is what I personally think is going on here.
I honestly don’t believe Google is using GA data as SEO signals. Why? Well for one they’ve told us so (call me naïve). And then there’s also the fact that even though GA is the most common web analytics system and many sites use it there are also many others that don’t, so if Google was in fact using this data as ranking signals then they couldn’t do it for all sites equally, which I don’t think will benefit the end goal of providing the most relevant result for each search.
So this means people who think Google is using this data are just paranoid? No, probably not.
What I think is probably going on is that Google is “guessing” all this data from other metrics, metrics they do admit to use in their algorithm, for instance Dwell Time. Dwell Time is the time between a click on a search result and that user going back to Google to search for something else. This of course isn’t very accurate, you could click on a search result, close it after 10 seconds and then go check your email for 10 minutes before you go back to Google. In this scenario the Dwell Time would be 10:10 seconds, which does not mean you spend that much time on the site. However, if you click on a search result and go back to Google after 5 seconds that probably means you didn’t find what you were looking for in that page. You could see this as an equivalent of a time on site or bounce rate metrics. Google watches out for this sort of behavior and probably penalize sites which get a very low Dwell Time (in the past they even experimented with allowing you to block a site right after you hit the back button if the Dwell Time was low).
This is very similar to what happens in motor sports such as F1 racing. If you are an F1 fan you’ll know this, but let me summarize it for those who aren’t. One of the F1 most protected secrets of each team is their telemetry. The telemetry is basically a report of everything that happens while the car is on track; when is the brake pedal pressed and how hard, how many horse power is the engine delivering at each point of the circuit, were does the balance of the car goes in each corner, etc. This data helps teams figure out where there is room for improvement. Having this information from a rival team would be like gold, since you could see where they are doing better than you (breaking later for instance) and adjust your car to match their performance. This is why this data is so carefully guarded.
Teams however are able to figure out most part of these reports without actually getting their hands on it just from one public record, which is the GPS coordinates of every car at any given second. If you have the coordinates of a car at any given moment in time you can figure out where it started breaking, were it shifted gears, their top speed, etc.
The first teams who started applying these methods were accused of espionage, until they explained their methods which now all teams use. Sounds to me like a pretty similar situation to what is happening with Google in this case. The only difference though is that F1 teams were forced to reveal their secret (their algorithm) if they wanted to stay in the competition (as F1 is a privately managed competition), Google on the other hand can’t be forced by anyone to make their algorithm public (at least not at the moment).