The Effects of Keyword Insertion on Ad Performance
Keyword insertion can be a powerful tool in pay per click campaigns, and its ability to increase click through rate is well documented, but how does it affect conversion rate? Recently, one of my clients wanted to run a test to see just what this effect would be. To conduct this test, we took a subset of ad groups selling similar products and populated each of them with two versions of the same ad, one with keyword insertion in the headline and one without. A month later, we bundled the data for each type of ad and compared them against each other for different time periods.
The results we saw were quite interesting:
As you can see from the chart, the keyword insertion ads typically outperformed the regular ads on click through rate by about 13% over the 30 day period. On the other hand, the regular ads dominated the keyword insertion ads on conversion rate to the tune of about 19% over that same stretch.
Overall, this means that the regular ads had fewer clicks, but the clicks they did get were of higher quality, leading to a much lower cost per conversion and a better return on investment. The keyword insertion ads were able to bring in more traffic, but it was typically of a lower quality. This was probably due to the keyword insertion ads pulling in search queries that were not relevant to our site, and thus bringing in searchers for whom we had nothing to offer.
Now I’m not suggesting that you pull down all your keyword insertion headlines immediately, as every campaign and business will see different results and have different needs. For this particular client, we are able to bid on keywords that are somewhat broader than what we might use in an average campaign. For this reason, it’s important that we control our ad messaging to prequalify our clicks, and with keyword insertion we evidently lost some of that control.
I would encourage you to run a similar test to see how keyword insertion might affect your campaigns. Be sure to let it run for a long enough period of time to establish statistical significance, and once you have the results, let us know what you found in the comments!
Written by Kevin Hill (kevin.hill (at)channeladvisor.com)