One of the most difficult questions a client often asks is, “How much money can be made with this SEO project?” The answer to this is always a tricky one as there are many variables involved.

One of the Bruce Clay Australia directors, Des Odell, has written a great article about SEO ROI, which outlines ways to estimate potential additional traffic from an SEO project and then put a value on the additional traffic, either by added revenue or by cost savings from traffic replacement through other media.

To estimate the additional traffic volume gained from an SEO project, you have to know what the current rankings are, what the traffic volume is from the current rankings, and then forecast additional volume from the new rankings.

If rankings could improve from A to B, how could the additional traffic from those improved rankings be calculated? The answer is to calculate the additional traffic volume using the estimated click-through rate (CTR) associated with the ranking position.

Once the CTR likelihood of a future ranking can be quantified, it can be combined with the associated search volume to estimate the potential new traffic for that keyword. If this process is scaled to hundreds of keywords for a decent-sized case study, it can then be applied to the entire non-brand organic traffic of the website to get a much bigger picture.

Here is a proposed formula:

Given

F = Future non-brand organic traffic
f = Future ranking
P = Present non-brand organic traffic
p = Present ranking
n = Number of keywords monitored
k = Keyword
V= Search volume for a keyword
c = Function defining the click-through rate for a ranking

Translation:

Future non-brand organic traffic equals the sum of the factors of all keyword volumes by their future ranking position click-through rate, multiplied by the present non-brand organic traffic, divided by the sum of the factors of all keyword volumes by their current ranking position click-through rate.

Assumptions:

• The search volume for a keyword remains the same in the future.
• Google search volumes are accurate.
• The sample of keywords selected is representative of the entire website.
• Google is driving 100 percent of your non-brand organic traffic.

Note 1:

You could take the inverse of which would give you a number between 0 and 1, essentially showing a potential for improvement of your current rankings. The larger the number of keywords monitored (n) is, the more accurate that number will be.

Note 2:

This sum would give you the potential of Non-Brand Organic traffic you could receive from the n keywords monitored. A low performance could mean your listings are not compelling enough.

One key missing part of this forecast methodology is the CTR for each ranking. Luckily, at Bruce Clay Australia, we have been running CTR research for some time now, using the enormous amount of data available to us, and we thought it would be a good idea to offer these results to you guys, so here are the results:

Some of you will compare these to the good old AOL CTR stats from 2006, and as can be seen, the CTR for the top 3 is much lower here. These results are comparable to those which Optify and Enquiro published, although the universal search results were not as prominent back then, which should explain the difference in CTR between the top 3 and the fourth position onwards. We have actually noticed the same difference when comparing the latest data (October 2011) to the first data set (August 2010).

There are a number of other studies out there, some interesting. One problem is that few would publish the methodology followed or the sample size. Also, most of these studies did not account for branded queries, which skews results towards a higher CTR. Other studies also used rather unusual methodologies, sometimes over-doing it to the point of reducing their sample size to a few hundred keywords, which impacts the quality of the results with such a low base.

For those interested in the methodology used, here are the different steps followed:

• Collected search queries in Google Webmaster Tools across a large number of websites across various industries every six months for a year (three data sets).
• Removed all branded queries to avoid any brand bias in the CTR.
• Removed all non-clicked impressions to avoid any bias in CTR.
• Aggregated all impressions and clicks for each ranking position.
• Calculated the average CTR for each ranking position.

Sample size:

• 100 millions impressions
• 650,000 unique queries

Additional interesting facts:

• 97 percent of the queries drove less than ten clicks.
• 91 percent of the impressions were driven from queries with more than ten clicks.
• Here is a breakdown of the clicks per SERP:

Finally, assuming that the “reasonable surfer” surfs reasonably the same way across search results pages, you could apply the stats in the table above to establish the potential traffic for keywords past the first page.

We hope this research can be used to help you value your organic rankings and establish the ROI of your SEO project.

If you want to discuss this topic further, please feel free to contact Bruce Clay Australia or me, Martin Orliac.