Visualising Google Click Through Rates
Jan. 14, 2009 | by Gregory Lyons
The above heat map (click to see a larger image) shows the results of research we’ve been doing here at iCrossing on Click Through Rates (CTR) for natural, paid and shopping results in Google. The darker red colours represent higher CTR’s while the lighter colours, lower CTR’s. What was found is that the position (on the Search Engine Results Page, SERP) with the highest CTR is the top spot in the natural search results. The next highest, surprisingly isn’t second place in the natural results, but the top position in the premium ads (top left of screen), followed by second place in the premium ads and then position three in the premium ads (which has a similar CTR to position two in the natural search results).
Of particular interest are some of the newer features being introduced into the SERP’s. The above SERP has a shopping result as part of the returned query. This along with other universal search results such as image, news or video results undoubtedly change the CTR distribution on the page and therefore are important to understand. In this case, the CTR for the shopping results changes based on whether the search is item specific (i.e. includes a brand name / model number) or more generic such as in the example above, but in general can be said to have a CTR similar to position four for generic searches. Given this understanding, maybe you’d want to start getting your own results in there!
Finally, analysis of the difference in average CTR for the premium paid spots and top normal paid spots reveals that users are more then twice as likely to click on a premium ad as they are a normal ad, good to know when planning your campaign.
Comments (12)
[...] when you’re showing a client, but what about the data behind it? It’s worth reading the original post and checking out the comments before putting too much faith in the number. Still a nice piece of [...]
Halpern
Thanks for the incite Greg.
Reassuringly it does reflect what we find in the field. It is always good to have some hard stats. I hope you don't mind that I'm going to quote you as a source of the data to my SEO and reputation management clients.Halpern
Thanks for the incite Greg. It
My SEO clients will be pleased when I refer them back to your research that it is still better to go for the organic results rather than the paidSEM Services
I've found that PPC has been obtaining an ever growing number of clicks, although 20% may be a bit high, nearer to 12%. However, it would be interesting to get market average figures for different search terms across a number of markets. When do you think you may do a more in-detail analysis of this?
Gregory Lyons
Hi Affan,
Each click made on the SERP contributes towards the CTR of each result. As for the new Google checkout logo, maybe I'll make that my next research piece :)Affan Laghari
Hello David,
Interesting work and nice to see the analysis and your explanation of AOL faults.
I wanted to know if this study shows the FIRST CTR or the overall clickthroughs. Because many savvy searchers (and there are relatively more of them now in 2009) tend to see a site, note/bookmark stuff, come back to the serps and click on another result. So this means the overall clickthroughs are almost always a lot more in number than the overall number of searches. Ignoring this can greatly underestimate the value of the 2nd and 3rd results especially.
Also, you would have noticed Google now shows an attractive Google Checkout image for Sponsored Links where the store uses Google Checkout. It would be nice to know if that increases the CTR or doesnt affect.David Hughes
Hi Hylton,
Thanks for your comment.
Your points broadly match the results we saw. As you say; when no paid search ad appears; natural clicks inevitably rise; and there are cases where paid search share approaches or exceeds 50% (mostly on very commercial terms). However, there is a huge long tail of queries (http://www.readwriteweb.com/archives/udi_manber_search_is_a_hard_problem.php), where paid search coverage is patchy; and thus the vast majority of clicks are on natural results. The heatmap you see is an average of all of this data.
When viewing the heatmap; it's worth knowing that the colour mapping uses an ordinal scale; that is; 'orange' constitutes a higher CTR than 'yellow'; but not necessarily twice as much.Hylton Donnelly
Hi Greg / Nick.
This is really interesting. Great research.
I've looked at the heatmap and played around with some numbers, and it seems to me that when there are as many as 3 sponsored links above the natural listings then as much as 50% of the clicks on the page must be coming from the sponsored listings. I say this since I can't really see how it could be less and still maintain the order of Min to Max CTR on the various positions as you show in your heatmap.
On the other hand, I expect that when there are no sponsored listings above the natural listings, then the natural listings get the lion's share of clicks, and the sponsored listings on the right hand side generate only a small proportion of clicks.
Does your research reflect this? Please can you comment on whether this sounds right to you, and how this ties in with David's comment that PPC get slightly more than 20% - is this on all pages averaged out, with most of these not having as many as 3 sponsored links above the natural listings?
Thanks
HyltonNick Garner
David, thanks for your response. I'll research these articles. your insight on the aol data makse a lot of sense BTW. As you can imagine here in betfair there is a constant push to come up with business case numbers and as you suggest the aol data is relatively blunt as indicators go - so I'll put less reliance on it!
thanks again
Nick
David Hughes
Hi Nick,
The problem we found with almost all analyses of the AOL data is that they didn't segment the search terms enough, or at all. That's to say; click through rate figures were calculated across all search terms. Within the 36 million searches released by those unfortunate AOL researchers; a great deal were navigational or brand searches. Users are far more likely to click on the top result for this type of query. Therefore, most analyses of the AOL data are skewed towards the top result. Our research indicates that click through rate on the top result is often less than half of that claimed when looking at generic queries.
We found that PPC gets slightly more than 20% CTR.
Although, we try to be open and honest on this blog; we do need to hold some of the detail back for commercial reasons.
However, if you're interested in this type of research; you might also want to look at these other posts by Greg (and Dom)
http://connect.icrossing.co.uk/icrossing-update-google-ctr-x3-higher-yahoo_639
http://connect.icrossing.co.uk/paid-search-position-1-cheapest_816
http://connect.icrossing.co.uk/click-rates-google-shopping-results_1467
I'm also going to be talking about how we use this research at SES in London; http://www.searchenginestrategies.com/london/agenda-day3.html. Its all very well having the findings - it's what you do with it that matters; it can dramatically change your SEO/ PPC strategy.Nick Garner
have you got percentage numbers to go against this? You may remember the aol data that came out showing 1st place getting 42% CTR and descending from there.
ref: http://www.jimboykin.com/click-rate-for-top-10-search-results/
If you were able to give percentages, then it makes it far easier to do analysis on this.
Also do your findings tie in with the generally held idea that PPC gets about 20% of all click throughs?
Chris Eden
Interesting stuff Greg, I think this is a great indicator that users are responding to the bigs G's way of ranking Paid Search ads more by relevancy and Page rank these days, rather than just being so CPC focused.
Post a comment
Stay Connected
SUBSCRIBE
Sign up for email notifications of Connect blog posts.
Guy Weston | Leeds Digital Marketing and Strategy Consultants » Interesting Search Insights