SEO study by Compete, whitepaper comments review

Compete.com, an online research company, has recently published a whitepaper stating that 53% of organic search clicks go to the first organic result. Feel free to download this paper. It presents some interesting results about user’s search behaviour based on their 2mm sample group. Miranda Miller, Compete’s marketing person, posted an article with findings at Search Engines Watch. Make sure you take a look at it for more details.

However, in this post I wanted to summarize some of the most interesting comments from that article. Also, I wanted to share a few resources from that SEO study that Miranda and dweigner, another commenter, shared.

SEO study comments

Frank Orman
While it makes perfect sense for higher rankings to get more clicks, the bit that always interests me is the ROI of the clicks. So, for example, if we take paid listings and compare the cost per conversion at postion 1 versus the cost per conversion at lower positings, what would be the optimum position for ROI? I know this metric also depends on site conversion capabilities, average value of sale, etc but have you carried out any research into this?

dweigner
Hi Frank, You might be interested in this study (beginning slide 24), where we found that position does matter in terms of conversions: http://www.thinkwithgoogle.com/insights/uploads/940793.pdf/download/

anilopez
I’ve downloaded the whitepaper looking for explanations on what methodology was used to perform the analysis.

I could only find those two sentences: “Compete analyzed 10s of millions of search engine results pages generated by actual consumers in our U.S. panel in Q4 2011.” and “Looking at all of the listings displayed with the 10s of millions of SERPs generated by Compete’s panel we found…”

I’m sorry but this does not tell me much about the procedure so “Compete’s research shows” is not a valid statement, specially knowing that Compete numbers (regarding organic traffic) are really far away from what you see at your favorite analytics tool.

dweigner
Hi Anil, Compete has a panel of 2 MM consumers that we use to track all of their clickstream behavior on the web on a longitudinal basis. For this data, we used a new capability we have where we capture the full HTML of every SERP generated by our panelists (among 2 MM people you can imagine that is 10′s of millions of SERPs). We then parse the HTML data so we can see what listings are on the page, where on the page the listings are, and for what advertisers. Because it is clickstream data, we can also see if the user clicked and on which ad. Let me know if this helps, you can also find more info here: http://www.compete.com/us/about/our-data/

Larry Chrzan
Thanks for this articl about this study. I’d be curious to know what is meant by “”tens of millions” of consumer-generated search engine results pages”. Are we talking about just consumer/retail type searches? As I understand it, there is a wide variety of results for different industries, with B2B searches not so heavily focused on the #1 result.

MirandaM_EComm
Good question. As I understand it, Compete used a panel of 2 million U.S consumers for their Q4 2011 analysis, so the tens of millions of SERPs are based on the searches of those people. It’s a larger sample size than many studies we receive, but marketers always have to keep in mind that no one study is perfect. You can take these insights and use them with your own data and other available studies.

It would be great to see queries segmented by query intent, but if all this does is start the conversation about it so someone else with access to this volume of data is inspired to look deeper, I’m happy with it :)

I know that Compete did work with “Think with Google” to produce a number of reports on consumer behavior within specific verticals, such as toys or footwear, based on research in this period, using this panel. There may be more information there that could be of use.

Again, make sure you get the actual whitepaper and read the post on Search Engines Watch to get a whole picture of the discussion. I hope these select comments drive the point of importance of methodology in any research. Finally, this post should give you a good idea on how you engage with users and value you provide in your comments. Really good example!