A common problem for running search campaigns in niche industries is the typical low volumes of data to optimise against.
However, it's not just restricted to areas such as B2B, but also applies to long-tail campaigns where some keywords might only get triggered a few times a day.
It’s no secret that click-through rate is the dominant factor for quality score in Google Adwords. At the SMX Advance London sessions earlier this week, I argued that, due to this in particular a lot of long-tail campaigns are not doing any favours for their quality score, and ultimately campaign success.
The problem has all to do with split testing and statistical significance. Since the dominant factor is CTR, it’s key to continuously split-test and evaluate new creative in order to stay ahead of the competition.
Someone utilising a long-tail approach by creating a 1:1 campaign structure (one Adgroup containing one keyword and a tailored creative) and at the same time is relying on Google to optimise by CTR of split-tested ad copies, will have to possibly wait a very long time.
Assuming that we want to achieve 95% confidence that the new adcopy will outperform the original, the following table highlights the number of days we have to wait to achieve this if our creative is only getting 100 impressions a day.
Most search marketers would find it difficult to wait 40 days to merely know if the new adcopy is performing better or worse. In case it’s worse (with the same percentage point dip), the consequences could be wider reaching than just the loss in clicks, but also lowered quality score, higher CPC and ultimately negatively impact campaign ROI.
Note that more dramatic changes in performance will lessen the time needed for this test.
In order to properly CTR optimise these campaigns, one solution is to use a system that bring together data from different keywords and adgroups on a semantic level and use the aggregate numbers to find the winning combination of a creative.
Alternatively, if there is a reliance on Google, the impression levels need to be considered when structuring the campaign, to make sure that the ad groups receive enough volume for efficient split testing.
The size of the circle denotes the number of keywords in each Adgroup (real or semantically created through reporting), and the number inside the total daily creative impression volumes: