Audience

Click Through Ratio, and How It Can Be Improved

Most online advertisers have to deal with improving their CTRs, and likewise, publishers have the challenge of improving the CTRs of the advertising that they sell. The way it is calculated is simply dividing the number of clicks a campaign (or an ad) has generated over the total number of times the ads were served (impressions).
What about the users who saw the same ad several times and clicked? Do we still expect them to click every time they see the ad. Worse yet, the same user might see the same ad on different websites, and therefore, for that user the CTR is 0% on the sites where she did not click.
I think the healthier way to calculate this is first by going back one step and actually understanding what the objective of CTR is.
Do we really care about this metric, or is the reality that we want to know the percent of people who responded to our message?
Easy question...
Then why don't we focus on measuring this thing exactly instead of just dividing two metrics over each other? 
How this can be calculated is very simple, but needs to take into consideration the number of people (remember we are advertising to human beings).
Let's agree on some simple definitions: 
Impressions: the number of times an ad has been displayed during a campaign.
Unique impressions: the number of unique visitors who were shown the ad.
Clicks: the total clicks an ad has generated.
Unique clicks: the number of unique visitors who have clicked on the ad.
Dividing the number of unique impressions over the number of unique clicks gives us the effective response rate.
Let's call it the effective click-thruogh ratio eCTR.