As most of you likely know, Yahoo stopped updating Overture’s inventory data in January 2007. Besides the issues this may cause for those relying on it for research on seasonal terms (that boxingdayspecials.com buy isn’t looking so hot anymore, huh?), it means that search marketers have, at best, outdated information in the volume of searches carried out for a given keyword in a given month. I have a solution that will approximate the inventory of searches for a given keyword according to Google’s data.
When conducting AdWords research, Google offers an estimate of the bid required to rank in positions 1-3 most of the time, and the volume of clicks an ad in these positions is likely to attract. For example, a given keyword may require a bid between $2 – $5 to rank in spots 1 – 3 and thus attract 150 – 300 clicks a day.
As a baseline, we know that there are at least 4500 searches a month for our keyword (150 x 30 days). Perhaps there may even be 9000 searches a month. We can get more precise, however. This precision will show that a keyword like the one in the example above will actually have 5 digit + search volume.
Any one Adwords ad is not getting 100% CTR. Actually, more common CTRs for top spots hover (in my experience, anyway) in the 4% – 10% range. Those 150 – 300 clicks a day mean 10 – 25 times that in daily search volume. (i.e. 1500 – 3750 searches a day). Multiply by 30 and you get 45,000 monthly searches or more.
But as CTRs vary widely, I propose another alternative for estimating monthly searches based on AdWords estimated click data. It’s been reported in several places (no time to find sources right now, but I can add them in the comments later, or those of you reading this can do so too) that organic CTR vs Adwords CTR follows the Pareto principle – the 80-20 rule. 80% CTR on the organic, 20% on the ads.
4500 clicks for a top ad spot? Supposing that got the entire 20% of the CTR on Adwords ads (unreasonable, but just for argument’s sake)? That means you have four times more searches (80/20 = 4). In other words, we’re actually talking about 20,000 monthly searches (4,500 x 4).
What do you guys think? Is there any value in these numbers? Would it be worthwhile creating a keyword research tool around such calculations? Perhaps use it for reputation/buzz-monitoring? Looking forward to your comments!
