Target CPA is a smart bidding strategy offered by both Google & Bing that optimizes data to get as many conversions as possible within the set goal of cost per acquisition. For example, if you set your target CPA (tCPA) to $100, the engines will make optimizations to get as many conversions that have a CPA of $100 or less. One thing to note is that even if you set your tCPA to $100, that does not guarantee that your CPA is going to be $100 since there are other factors that could impact this number. We have found in past tests that it can be under, over, or spot on.
Google states on its support page that when making optimizations, it uses historical data and contextual signals at each auction to find an optimal bid.
Things to consider before you launch target CPA
- It is recommended to have at least 15 conversions in the last 30 days. This allows Google and Bing more data to optimize. If you have less than that, the engines have a more difficulty deciphering when to make adjustments.
- Is it your busy season? If so, it is best to wait until after that is over so that you don’t take a risk that may negatively hurt your main KPIs. With Target CPA it is recommended to let it run for around 30 days before making adjustments to the set CPA. If it is your busy season, and the Target CPA test is not working well, you could have a serious impact on your bottom line.
- Are you about to make major changes to your landing pages? If so, this could play a large role in both CPA and conversion volume. It is important when reporting on performance on Target CPA to note any other major changes made to those campaigns.
- Don’t make the Target CPA unrealistic. If your average CPA over the last month is $200, setting the Target CPA to $50 could be detrimental to conversion volume. The hope with Target CPA is that it doesn’t have a major negative impact on conversion volume.
What campaigns should you test target CPA with?
The first recommendation would be to test on your highest conversion driver campaigns. The more historical data and previous contextual signals the engines have to use for optimizations, the better they are at making those adjustments.
We tested Target CPA in a campaign that had 90 conversions in the last 30 days. We set the CPA 20% below the last 30 days CPA. After running the test for 30 days CPA declined 30%, conversions increased 19% and CVR increased 8%.
You should also test it on campaigns that have an ROI below goal but sufficient conversion volume. For example, one of our clients had a strict 2.0 ROI goal. There were campaigns in the account that generated strong conversion volume but were consistently missing the mark on the 2.0 ROI goal. For one month Target CPA was run 20% lower than what CPA had been in an experiment against the campaign running Manual CPC at a 50/50 split. At the end of the 30 days the experiment had a CPA $30 lower than the campaign not running Target CPA and it drove 94% more conversion volume. This resulted in a 2.0 ROI
When should you not use target CPA?
It is best to not test it on campaigns that have less than 15 conversions in the last 30 days. With less than 15 conversions, you are not providing the engines with enough data to make the most informed decisions and optimizations. We tested Target CPA in a campaign that received 11 conversions in the last 30 days and set the CPA 20% below what it had been in the last 30 days. After the 30 days test, CPA increased 64%, conversion volume declined 55% and CVR declined 58%. The 11 previous conversions were not sufficient enough in providing Google data to make the optimal adjustments.
This is a reminder to not use the set-it-and-forget-it tactic with tCPA. Complete a 30 day test and follow up with a check-in. That way you can easily check to see if the smart bidding strategy is not working.
If you receive low conversion volume but still want to use a smart bidding option, we would recommend testing eCPC on the low volume campaigns in an effort to lift monthly conversion volume.
To put it simply, we recommend testing tCPA if you have over 30 conversions in the last 30 days. It works sometimes but not every time. If it doesn’t, you should test other automated bidding strategies.