What’s the difference between the highest performing publishers and the lowest performing publishers in Australia? How about 363%?

Data from MediaMind shows that the top 25% of publishers achieve Click Through Rate (CTR) of 0.07%, compared to 0.02% for the lowest performing 25% of publishers. Therefore, choosing the right set of publishers is critical to an advertiser’s success. The best way to achieve this is with a media plan that takes into account historic performance data. Historic data can prevent costly mistakes and optimize performance.

Since publishers are not all cut from the same cloth, it is wrong to assume that as long as they can attract your desired audience and demographic, these publishers will perform as expected. In fact, many of the variables that can influence your online performance such as site layout, engagement with the content and average time on page can be more detrimental than the site’s demographic.

src=http://www.marketingmag.com.au/web_images/grarphg.png

To examine how publishers fare in delivering performance to advertisers, MediaMind looked at a random sample of 47 US Australian publishers, and divided them into four quartiles with roughly equal number of publishers sorted by their performance. To make the comparison fair, only publishers with over a million served impressions of Standard Banners were included. The comparison was based on Standard Banners, which have less variability in creative implementation than Rich Media. Therefore, the results are less likely to be skewed by variables such as ad format and use of video.

Results were staggering, with the top quartile performing nearly four times better than the bottom quartile, and significantly better than the second quartile. Large differences were also observed when examining publishers’ performance for Rich Media, using Dwell, an engagement metric. Using historic performance data to choose publishers in the top quartile boosts performance significantly. This analysis uncovers the tremendous impact that a good data-driven media plan can have on the results of a campaign.

There are three caveats to this analysis. First, clicks may only be a partial measure of an online campaign’s effectiveness and some advertisers may rather use conversions. Second, this analysis did not include costs, so low performing publishers may still have lower cost per click. Last, there is a large variation between site sections within every publisher, so well performing placements can be found within low performing sites.

As online advertising becomes more quantitative, marketers and agencies should immerse their operations with data driven processes. Data driven planning should start from the top, adopting planning and buying tools that influence data-driven decisions, all the way down to each member of the planning team by providing them with training and education.

Every media planner can create better data-driven media plans starting today. It does not mean drowning in Excel spreadsheets of previous campaigns to calculate publisher performance averages. That’s what an ad server is for—aggregating historic performance data, so all you have to do is pull it out and use it.

While it is hard to pinpoint exactly what makes some publishers better than others, the stark differences in performance mean that the days of buying media by audience, by feel, or just by doing what was done last time should be over. Data-driven planning will improve the performance of online advertising by putting ad dollars where they have the highest impact and incentivizing low performing publishers to spruce up their site.