June 18, 2012
Picking winning PPC ads seems like an easy task. You find the one ad with the best metric(s) such as conversion rate, pause the rest, spin out a variation of that ad, and test again. This can lead you in the right direction fast, or it sometimes can blind you to another value of your ads, gross conversions.
Below are three case studies that use a metric call Conversion per Impression. The set up for all three case studies is the same. We are picking the wining call-to-action based on a keyword or phrase’s impact on conversion per 100,000 impressions compared to the account average over the past 6 months. Why on this metric and now CTR, or conversion rate, or even CPL/CPA?
PPC Metrics: Conversions per Impression(s)
This metric gives the best indicator of how an ad’s CTR and CR impact the gross number of leads or sales. CTR and CR alone paint an incomplete picture, and CPL is too heavily based on bids. By this I mean that if an ad group with one type of ad has high bids because the keywords in that ad group have high ROAS, then the CPL may be high and thus the ads look like they aren’t worth testing across other ad groups. But in reality, they may produce the most leads per impression! This means that in lower priced ad groups those ads could perform at very low CPL’s.
To calculate your conversions per impression for ads you take a sample number of impressions, for the examples below I use 100,000, and you multiply that number by the CTR of the ads you want to compare. That will give you a clicks number that you then multiply by the conversion rate, which gives you the number of conversions you would get for every 100,000 impressions. The sample set of impressions could be any number you choose, as long as you use the same number to compare your ads. I chose 100,000 because the output was a more tangible and thus easier to compare. (As opposed to doing conversion per 1 impression that would have only been .000071 for the account average in case study 3 below.)
In the below case studies, which were pulled from three random clients, we compare the conversions per 100,000 impressions for the top performing ads compared to the account averages. Keep in mind that some of these have Display campaigns grouped together with Search so CTR is going to look pretty low. Because the ads were tested evenly across Search and Display, the numbers are relative though, so the resulting numbers are comparable.
It is also interesting to note that those ads with the highest conversion rates end up all having the highest conversion per impression numbers. It’s the ads in the #2 spot and below where the difference appears to be made. This is especially important for testing because you may identify call-to-actions that you had written off because of low conversion rate but that deserve a second chance with a new approach.
Case Study 1 – eCommerce
For this client, as with most clients, conversion rate matters much more than CTR. If we were only to look at conversion rate, here is what we would see:
Based on this, the winning ads would contain “Shop Now” and “Today”.
But when we take the picture out to include both CR and CTR we see a slightly different picture. Instead of “Today” being 40% better than the account average (.84-.6/.6) it actually produces .34% less conversions per thousand impressions. That’s because the CTR is so much lower than the control that the increased conversion rate does not make up the difference. “Shop Now” remains the best performing call-to-action with ads that contain “Holiday” coming in at second best.
Case Study 2 – Lead Generation
Here’s what the conversion rate looks like for the top performing call-to-actions for our next case study.
Based on CR data only in this example, “Learn More” has a clear edge over everything else with “More Information” coming in second. Here’s what it looks like when we calculate conversions per impression:
Looking at the whole picture shows that “Learn More” is still the top performer by a mile, but the number two performer is now “Apply” which was towards the bottom of the call-to-actions when only CR was considered. Since “Apply” has a higher CTR than “Learn More” something to consider here is to write ads with both learn more and apply in them such as. “Learn More and Apply Now!” The preceding call-to-action combines three terms that all perform better than the account average.
Case Study 3 – Lead Generation
In the last example we see that ads containing “Expert” and “Let Us Help You” perform far better than other ads, and the account average
But when you look at the whole picture it is drastically different.
“Let Us Help You” creates 41% less conversions per 100,000 impressions than the account average. Hitching your wagon to that call-to-action, or testing variations thereof, could have led you down a disastrous path. Ads containing “Expert” still win, but ads containing tell your story are not far behind even though those ads have an average CR. Again, this means that exploring new ways to combine expert and tell your story could lead to your highest gain in gross conversions.
In only one extra step, you can see how comparing your conversions per impressions can help paint a clearer picture of what ads and call-to-actions are most likely to create the greatest gross amount of conversions for you.