After a couple frustrated attempts to get this post published in some of the great guest post SEO sites I’ve decided to post it here where it should have been published originally. They all told me that an exact translation was duplicated content and that they only can publish original content. Some of them asked me to revamp my content, others to complete it with an analytics specialist point of view, …, but I don’t really want to do this, I can’t imagine myself “spinning” my own post or writing a copy-paste tutorial to make it “really valuable for the community”.
I don’t want international fame, I don’t need +6000 views a day, I just want to let the non-Spanish speakers to read something that can help them to optimize the news, video and images strategies and to encourage people to keep searching and sharing.
So, here you have the duplicated content:
Since I started working as an SEO in mass media, one of my biggest frustrations has always been being unable to measure Google News Snippets performance in the Universal Search. For those not familiar with the term, it refers to the News summary that appears blended with adwords, maps, the authors and other trifles, and maybe some organic results in SERP's.
I know I'm not the most qualified person to talk about web analytics, and I hope not to offend the authorities (although, truth be told, it is always interesting to do so) but my natural curiosity and my obsession with metrics has led me to do some research to find out what's going on behind the News in Universal Search.
Before I begin, I must say that this information doesn't guarantee success when used, so if you have trouble applying it or the side-effects are not expressed in this article, ask your SEO or web analyst.
The main problem in measuring its performance is that Google in its large commitment to "transparency" (lately this term has been used to refer to other concepts not related to transparency,so that’s what I’m doing here). Google marks these url's as source=web, as well as when a user searches within the news search engine, Google marks it as source=newssearch. Therefore, when we want to see in our reports how much traffic comes from "News" (referer news.google.com news.google.es, …) it does not include these two cases in which, according to their shrewd logic, this traffic always appears mixed with all the other organic results, making the delight of all media analysts who go crazy looking for a way to split hairs, using filters, advanced segments, causing headaches, requiring ibuprofen, …
I've always had the feeling that since Universal Search pages include news summary, few people use the news section to search and they normally click on the “selected summary” offered in Universal Search. Despite not being an analyst, and knowing that when we SEO's talk about analytics it "sounds" like when we SEO's talk about content, I decided to share the results of a small experiment that allowed me to measure, with the help of many people, in which situations a click comes from the news block or not.
If you look at the url that lies behind a click you can see that Google "encodes" this url and measures many more things than the regular user could imagine. There are a lot of posts in other blogs speaking about measuring the position of a result tracking the cd parameter. When talking about news and other nonsense, we have to focus on the ved parameter, which hides much information related to Universal search.
For results in news snippets, you can “always” find the pattern QqQIw immediately after the fourth character of the ved parameter, for example, ved=0CDkQqQIwAA, and the source parameter fixed to web.
Usually, in these snippets there is a thumbnail that refers to another piece of "top news" (and that’s the reason why I have quoted the word always in the above paragraph), this link has a different pattern in its ved parameter, showing QpwI instead of QqQIw, also from the fifth position, being its source the same, but without marking any position via parameter cd.
I know some of you may expect a "copy-paste" tutorial on how to implement this "trick", and I would certainly get more comments, RT and links, just as if I had named this post something like "ved parameter exposed" or "google news behind the scenes" or "Everything You Need To Know About …". I am not in favour of writing such posts and I'd like to encourage people to investigate, test, break things (carefully, because afterwards you’ll have to fix them again). If you think that this information can be useful and you get to find out more, I ask you to please share in the comments below a link to your blog, so we can move forward and learn every day a little more about Google's "transparency".
If you feel lost with analytics and what I've written sounds a bit messy, do not try it, start with easier things. If you know what I'm talking about but you need more information:
- Filter organic traffic
- Filter Google traffic
- Filter source=web from the reference
- Filter the ved parameter pattern inside the reference
- And that's all folks!
We now have in this profile all the traffic that comes from the News snippets in Universal Search.
Moreover, as a control factor, I have located the position of the click in the filter, via cd parameter, to detect errors. This way I realized that when I was performing my first tests on this profile I got results in impossible positions because of my “great” skills with regex.
In addition to these patterns, I've been able to detect others, such as:
- Organic Videos (ved = 0CCgQtwIwAA for links and ved = 0CDkQuAIwAzgK for thumbnails)
- Search within News (ved = 0CC4QqQIoADAA and ved = 0CDcQ-AsoAjAB both with newssearch source parameter)
- Images (ved = 0CF8Q9QEwCA)
I have not been able to isolate either authors or maps, but you have to understand that, when working with mass media, I'm not really worried about maps, so from these lines I encourage you to keep searching and sharing.
I could not finish this post without thanking everyone who has helped me out with this little experiment:
- Jaime Cuesta ("if you need some help with that shit you're doing, just let me know!")
- José David Pascual Traver for letting me use the car news section at motor.es for my tests and for his support with regular expressions
- Jorge Mesa, Web Analytics Specialist at Kanvas Media
- Carlos M. Lebron for helping me to implement all this mess cleanly in an account as complicated as RTVE.es
- and especially to Luis Pinheiro, project manager at RTVE.es, for allowing me to spend some time doing research and move forward in certain small details that will allow us to analyze data in a simpler, quicker and more reliable way.
Finally, thank you very much for reading and I hope that, if you are interested in continuing this little research and you reach other conclusions, you will share them with us.