The Value of Google Result Positioning
May 25th, 2010
How much is the top spot on Google actually worth? According to data from the Chitika network, it’s worth a ton – double the traffic of the #2 spot, to be precise.
In order to find out the value of SEO, we looked at a sample of traffic coming into our advertising network from Google and broke it down by Google results placement. The top spot drove 34.35% of all traffic in the sample, almost as much as the numbers 2 through 5 slots combined, and more than the numbers 5 through 20 (the end of page 2) put together.
“Obviously, everyone knows that the #1 spot on Google is where you want to be,” says Chitika research director Daniel Ruby. “It’s just kind of shocking to look at the numbers and see just how important it is, and how much of a jump there is from 2 to 1.”
The biggest jump, percentage-wise, is from the top of page 2 to the bottom of page 1. Going from the 11th spot to 10th sees a 143% jump in traffic. However, the base number is very low – that 143% jump is from 1.11% of all Google traffic to 2.71%. As you go up the top page, the raw jumps get bigger and bigger, culminating in that desired top position.
| Google Result | Impressions | Percentage |
|---|---|---|
| 1 | 2,834,806 | 34.35% |
| 2 | 1,399,502 | 16.96% |
| 3 | 942,706 | 11.42% |
| 4 | 638,106 | 7.73% |
| 5 | 510,721 | 6.19% |
| 6 | 416,887 | 5.05% |
| 7 | 331,500 | 4.02% |
| 8 | 286,118 | 3.47% |
| 9 | 235,197 | 2.85% |
| 10 | 223,320 | 2.71% |
| 11 | 91,978 | 1.11% |
| 12 | 69,778 | 0.85% |
| 13 | 57,952 | 0.70% |
| 14 | 46,822 | 0.57% |
| 15 | 39,635 | 0.48% |
| 16 | 32,168 | 0.39% |
| 17 | 26,933 | 0.33% |
| 18 | 23,131 | 0.28% |
| 19 | 22,027 | 0.27% |
| 20 | 23,953 | 0.29% |
Numbers are based on a sample of 8,253,240 impressions across the Chitika advertising network in May, 2010.
Contact:
Daniel Ruby
Research Director, Online Insights
Chitika, Inc.
+866.441.7203 x966
press@chitika.com
Become a Fan on Facebook
Subscribe to Chitika Research
Hi Daniel, nice to meet you. What a fantastic article. This also supports what we’ve seen on our company website. We never broke down the percentages but we’ve noticed similar patterns with our landing pages. Thanks for the post.
Jim Adams
This is great data. Thank you for sharing it. It would be even more interesting to see how that traffic converts. So, do conversions from the traffic segments follow the same or similar trends as the traffic itself? That is, are conversion rates from traffic from the #1 result (magnitudes) higher than conversion rates from traffic from the #2 result, etc?
Any information for when more “universal” type data is returned. For example a local 7 pack, news results at the top, videos or images on the page? Most of these numbers look pretty similar to the AOL data leak a few years ago.
Thanks!
Mike
I agree with Jeff Scott (you can find something of mentioned study on AOL’s search query logs here: http://www.webuildpages.com/jim/click-rate-for-top-10-search-results/).
It was August 2006… so what has changed since then? NOTHING?!
I’m sure that everything has changed in communication (i.e. social networks), multimedia sharing (i.e. images, videos, songs): so we have more opportunities to reach the top.
Nothing, instead, has changed in Google’s or others’ s.e. core algorithm: you have to provide the best user experience, to collect quotes and reach TOP ranks.
Whatever some SEO say about positioning, it is definitely not dead !!
Whilst this study does yield results that we’d expect, there seem to be a number of issues with the methodology (which should be published in greater details when making claims like this).
For example – is the study seasonally normalised? There’s a snapshot in time, but there’s no clue as to whether certain sites naturally perform better / worse in May. This could skew the results.
Are the results normalised for keyword search volume? Again, keyword search terms with higher / lower search rates could well skew results.
Is there an accounting for the presence of wikipedia / similar sites in the results? Intention of search is definitely something to consider when considering clickthrough rates.
Are the search terms quoted as ‘exact match’ or not?
I think, while interesting, the results as is don’t really give us much in the way of insight beyond ‘higher is better’.
#2 is the first loser. But in this case you get rewarded at least 17% of the time. I’d rather #1 anytime.
Its interesting to see that people do go to second page of Google to find what they are looking for. Quite surprise by the 2nd page results. I also thought there would be a huge difference between 1st and 2nd position.
Nice one Daniel.
I knew there was difference between 1 and 2 and of course between 10 and 11 considering it page 1 versus page 2 but honestly i did not know it was that much.
great post and info. thanks for sharing the info.
That’s not the pattern I expected – I thought there’d be a far more substantial drop off between first and second.
Great article though, very helpful.
Interesting – Great motivator for seeking out niche search terms with longer tails
Interesting to see these numbers in comparison to the 2006 AOL dataset which showed the following (42% for 1st position, 12% for 2nd, 8.5% for third, etc).
One of the key take-away’s is that ~90% of traffic never makes it to the second page. Search behavior research continues to show that people refine their search (ie. do a search with more keywords, long tail) instead of scrolling to the second page. When you overlap this with the “highest converting keywords by length” studies, it shows that the keyphrases with 3-4 keywords tend to convert much higher than generalized shorter-tail keywords. Managing several ecommerce clients, I can tell you that the volume (along with competition) is in short tail, the conversion is in the long-tail!
Another thing to consider is the that first position in organic results is competing with other aspects of Google’s Universal search (live feeds, Base, Places, Video, news, images, etc). The” #1 spot in search” can often actually be the 10th, 15th option (3 Adwords ads, 7 Maps ads, images/news/videos….and THEN organic). So while the #1 spot gets ~35% of the clicks of organic traffic, that same organic traffic gets 65-75% of total search volume (with 25-35% going to PPC depending on which study you look at).
Great to see people utilising the data available to them, Daniel. Though I do have a few reservations over the accuracy of the conclusions being derived from this data.
Simon makes some good points. The most important point is a point I would like clarification on.
“Are the results normalised for keyword search volume? Again, keyword search terms with higher / lower search rates could well skew results.”
From the blog post it sounds like this is not the case. This means if I was to rank first for “pens” and second for “red ballpoint pens”, I could assume that first place rankings drive a huge amount more traffic than second place rankings.
Clarification on this point would be awesome, to give this study some credibility.
Good questions, and with regards to the results, it’s a sample of all Google search traffic – we did no normalization of keyword search terms. Given the scale of data we broke down, the pens vs. red ballpoint pens discrepancy should even out. Our goal was to get as accurate a picture as possible of the overall impact of Google result position.
I wish I had access to compare these numbers to Google’s paid results, but unfortunately I do not. It would be interesting if Google came out with a comparison of top paid vs. top organic result across their network.
The methodology was pretty straightforward – we looked at all Google traffic that came into our network and broke down what percent was from what results location.
And while it’s not seasonally normalized (it is certainly a snapshot in time, although it’s over a full week so should even out any weekend vs. weekday shifts), I’m not entirely sure that would be overly relevant. The data set is large enough that individual sites’ seasonal shifts should even out.
Daniel – good stuff here… I’ll second Mike Belasco’s question though – what impact on the #1 listing did you see when a 7 pack or other vertical search results were shown vs. not? Intuitively I’d have to assume it was lower than the 35% or whatever, but I’m very curious as to how much lower.
I’d be very interested to see if/how this pattern varies if the data is broken down by types of searches, such as name searches (typing “facebook” into Google to find the Facebook web site — don’t laugh, a *lot* of people do this) as opposed to subject keyword searches (typing “shoes” into Google to find general information about shoes), etc.
I’ve seen a lot of behavior to suggest that name-searches are much more likely to select the first result than are most other types of searches, and would be interested to see whether or not the data supports that. Filtering the data to just subject-keyword searches (and perhaps other classifications of searches) might provide more interesting insight to most SEOers…
As Thogek mentioned: branded search terms get even higher CTR’s for a nr 1 results because its a navigational search. Searchers are looking for only 1 result. Opposed to non-branded informations search keywords that can have lower CTR’s due to the fact that people search for information whatever website. So the clicks are more spread over the top results. Our studies show that the last get max about 20% if the site is listed nr 1.
So the summed ctr is over-estimated for non-branded and under-estimated for branded keywords.
@Danielruby: could you clarify about the mix of keywords used in the research? Both branded, non-branded? Biased?
Keesjan, it was a mix of branded and non-branded – we applied no filters to the keyword mix.
@danielruby Would be nice to see this research segmented by brand-non brand and 1 term versus 2 term versus 3 term analysis!
In our tool SEO Effect we use as benchmark figures 20% for non-branded fist position and 50% for branded nr 1 possiton. This helps the tool to predict on what terms the highest gap is in current traffic and possible traffic.
As Thogek and Keesjan mentioned, branded search terms get a lot higher CTR which must have skewed the result. On the other hand, when Wikipedia is the #1 result, it will never receive 34% of search traffic (in my opinion).
Great article. The fact that 60% still goes to spots 2-10 gives hope to the rest, plus demonstrates that Google’s results don’t always put the most relevant results at the top.
Where are you getting your impressions data from? You don’t have direct access to how many searches took place for each keyword, only google themselves have this. So I’m guessing you’re either using Adwords keyword tool estimates or Webmaster Tools impressions data, both of which are seriously dubious in terms of accuracy.
Don’t get me wrong, I’m very glad you’ve done this study and we need more like this but I need some clarification on how you came to the CTR data before it can be taken seriously.
@jaamit, our impressions data is from search traffic coming into our ad network. Essentially, we can see tens of millions of searches each day, spread across a huge number of keywords and a wide variety of sites.
The research doesn’t make difference between Sponcored links and organic results. Is there any? I guess so…
ProClau, we didn’t look at sponsored links at all. All traffic in this study is purely organic.
thought #1 spot would get 50% click through.
but its less than 40%.
Many thanks for publishing the above data. Even with the caveats raised by other comments, it is very useful to have data made available like this.
Hmm…thanks for the stats! However, in my mind I think I remember that the #1 spot was up just above 40% in the last survey I read. Course, as I didn’t bookmark that survey, I just can’t remember where I saw that…dang it! But your numbers while lower, are more recent….so thanks!
Jim
Great works. Very helpful stats and i have a question, How can I find Google impressions values?
This data has to be slightly off, do the math. If you add all the top 20 listings percentages they equal 100.2%!! I think you guys made a boo boo somewhere.
Great article and thanks for sharing the data.
It’s a shame to see that #1 gets less than 40% of clicks. It just goes to show that it would definitely be better to have multiple pages of your website found on the search results.
I think its self fulfilling and demostrates the skewed linear usage patterns from teh page layout. If teh links were not stacked as they are now but circular (would work well in chinese) might see something different. I reckon different users would developn different habits and that in turn would affect the ranking.
thats my tyheory anyhow, and when the browsers start displaying the data differently just remember i said it first
@Austin Texas Notary Public, that should just be rounding issues. I rounded all the numbers to two spots past the decimal, so there was a bit of excess when you add them up.
Great research! I’ll translate it and publish on my site. Thanks for share this information.
This information, whether exact to the last impression or not is very helpful to those of us in SEO. You don’t know what your are missing until you find out what’s available to you. Obviously, the key here is to be sure that the keywords you are ranking for is the keywords that sell for you and not something like “oatmeal” which for some reason was a big hit on one of my clients business websites.
Thanks for doing the math.
This information it’s very important, because is a proessional study. The difference between first and the second position.
If you appear in the first position you can win the double of the money instead the second.
I examples of sites who appear in 5th position instead 8 position, and the traffic value increase a lot. Maybe the double.
Paulo
I agree with Jeff Scott (you can find something of mentioned study on AOL’s search query logs here: http://www.webuildpages.com/jim/click-rate-for-top-10-search-results/).
It was August 2006… so what has changed since then? NOTHING?!
I’m sure that everything has changed in communication (i.e. social networks), multimedia sharing (i.e. images, videos, songs): so we have more opportunities to reach the top.
Nothing, instead, has changed in Google’s or others’ s.e. core algorithm: you have to provide the best user experience, to collect quotes and reach TOP ranks.