Averages can be deceiving. Gather ten of your colleagues into the conference room and calculate their average income. Now, imagine that Warren Buffett had just dropped by. While your average has just spiraled up, all of your friends remained exactly as well-off as they were before.

Averages can also be misleading when looking at the average frequency of your campaign. On average, two thirds of cookies have a frequency of only one exposure, while only 12% of cookies have a frequency of three exposures. The ‘ideal’ frequency has been under debate for a long time. In an online campaign, the ideal frequency may depend on variables such as off-line advertising exposure, product life-cycle and others. However, numbers that are suggested for an ideal frequency were typically higher than one.

Cookies are used as a proxy for unique users. The first time that a user views an ad served by Eyeblaster, a cookie is placed on his or her browser. Browsers can then be uniquely identified every time they are served an Eyeblaster ad.

In one of our earlier blog postings, we explored the difficulty in increasing average frequency. This provides another angle to this topic. Serving more impressions may increase the number of cookies with one frequency, as well as cookies with higher frequencies—thus leaving the average frequency more or less the same.

Ushering more cookies towards a higher hypothesized ideal frequency should use a balance between increasing the number of served impressions, decreasing the number of sites on which they are served and a prudent use of frequency capping. It is not likely that campaigns would be able to have the majority of cookies circled around the ideal frequency; however, it is possible to have fewer cookies with a frequency of only one impression.