What is the true impact of secure Google search, and the emergence of Not Providedinstead of keywords in your statistics, in the management and optimization of your Web activities?
By observing the first Analytics data of a set of sites whose main objective is monetization through Adsense, I would say very low … for now at least.
Unable to communicate precise figures (because of Adsense regulations and especially many of these sites that do not belong to me), here I give you reflections and a method to try to see more clearly in this fog Googolesque.
The irruption of NOT PROVIDED
Bad news for SEO artists …
Following this major change by Google last fall (late October), the international SEO community has reacted, as evidenced by these articles here or there or there .
Encrypted searches ( Google SSL secure search ), initially for users connected to their google.com account, hamper the exploitation of the referrer (in part – the Adwords infos are not restricted) and in particular the use of q = keyword typed by users doing a natural search. It would be unwelcome, as some “bad languages”that the Google Analytics premium option allows you to trace these keywords. What would be a real advantage compared to competing analysis solutions like Omniture, Core Metrics or AB, but could also constitute an abuse of dominant position … Personally I believe more in a decisive advantage to display the best targeted ads, based on of the Internet user’s past searches, to which the competitors will no longer have access.
The SEO exploitation of audience measurement, obviously loses much of its interest, if instead of the keywords used you get a
prettyawful (not provided) ! The data obtained through the integration of webmaster tools are far from complete and more useful to monitor the evolution of your CTRs.
So is SEO based on Analytics dead?
The real impact on your webmarketing and SEO activities
The excellent and inspiring Avinash Kaushik, guru and prophet of web analysis, tells us some perspectives here [in] .
By adapting the spirit of this article to our environment and the sites for which we deal with the Analytics tracking and optimization, I give below a possible approach and a first idea of the real impact for content sites monetized by Google Adsense.
I use an aggregate of statistical data over a period of 9 weeks (to be understood only as the first trend) for a dozen sites (looking first at their English version when available) reporting to their publishers a few hundred dollars of Adsense revenue per month on average per site, in various fields but not really “geeks”. For more information, these visitors sometimes click on Facebook’s Share buttons or “Like” buttons, and more rarely on Twitter buttons and so far completely ignore Google + 1.
First of all, how do you get some insights on this traffic now labeled “Not Provided” (not to be confused for non-Google Analytics with “Not Set” for non-organic traffic, eg direct or from referents )
A profile creation, two advanced filters, and a personalized report later (1 filter to get the complete referrer, 1 filter for the ranking of the keyword on the SERPs, thus obtaining the position of the (not provided)!) And then with a setting in the context of sites you know well, for example by including landing pages, number of visits, etc., you will start by guessing which main words are hiding behind these “not provided” for a significant part of your traffic. Of course possible for a site with a history you know well, where you spent many days and expected many months to climb the SERP; much more difficult for a new site! Or it will only be hypotheses that can prove to be just
Extract from a Google Analytics custom report A
well-known background and context (here landing pages for mini traffic) will greatly help this guessing game, but these are only hypotheses!
Of course you will not get the last query that just came out to feed your traffic “long troll”. With this little manipulation the head will emerge from the fog … maybe … but not the tail . But are you going to really analyze, even knowing them, the tens of thousands of words and varied expressions constituting this tail of troll?
Peanuts … but for how long?
For this kind of sites and visitors the Not Provided represent less than 10% of the organic traffic (average conglomerate) and you will have managed to identify a good half? (your comments for your sites are welcome)
Now if you add in your custom report the Adsense statistics in front of this traffic “Not provided” you will see that this one only generates a really very very good income amount (I ‘ insist!) negligible compared to the contribution of the whole site. Surprisingly this is true for all these sites observed with small variations from one site to another.
Until then, so-called “geeks” who sign in to their Google account only rarely click on advertisements … Google, because they are (too?) Familiar with these ads, there is only a step … that I would not cross!
But here is a trend “that it is interesting!”, Right?
Note: If you have dedicated sites for “geeks” it would be interesting to communicate the% (not provided) and the average Adsense CTR% delta (sufficient magnitude, eg minus 70%)
- By setting up Google Analytics, you can get interesting information about “Not Provided”. Contextualization and in particular the association of the landing page makes it possible to find the majority of the traffic of head, for a well-known site … but in no case that of tail of the long train.
- Visitors to this type of “non-geek” sites connected to their Google account account for a small proportion of overall traffic.
- Moreover, this type of visitors by definition more familiar with the web, seem unlikely to click on the ads! This is at least the trend that emerges on the first statistical data for this particular conglomerate of sites.
- If this initial trend is less alarming than one might have thought, a few weeks Analytics data for a very limited sample of sites is meager … Also with Google’s efforts to promote its social network and application Google.com rules, the trend could be reversed … which could then become problematic for Analytics data. To be continued!
To subscribe to the RSS feed of the new articles of this blog, click on the logo opposite: