AdWords “Optimize for Conversions” Ad Rotation Could Be Killing Your Conversions

Having worked in the PPC industry for over 10 years, I’ve built a lot of campaigns. The AdWords default setting for ad rotation when you first setup a campaign is “optimize for conversions”, but every time I go through the process, I change ad rotation to “rotate evenly” so that Google is forced to split impressions between ad variations. It’s not exact, but it allows me to be the judge of which ad I want to keep or pause in future ad tests.

I’ve also talked to various AdWords reps over the years and they’ve often recommended I change the setting to “optimize for conversions” so that AdWords can serve the ad that is most likely to convert. The same advice shows up frequently in the Opportunities tab. The entire AdWords ecosystem, from default settings to automatic recommendations to in-person advice, tells people to let AdWords do the optimizing.

I was recently auditing an account and came across a case study of why you should be making the decisions yourself instead of letting AdWords do it. Here is 3 months of ad performance data for an ad group:

AdWords Optimize For Conversions

First, look at the % served column and see that AdWords is vastly preferring ad #4, showing it 52% of the time despite having 4 active ad variations. Next, look at the Cost/Conv. column and see that ad #4 has the HIGHEST CPA of all variations. You can throw out ad #1 because it only has 261 impressions, but look at how much better ad #2 is performing. The CPA is more than 25% lower and the CTR is 4.75 times higher.

This is why you don’t trust AdWords with the important decisions affecting your bottom line.

But Why?

The only possible explanation I can muster is that AdWords has a built-in bias. It’s obviously not looking at conversion rates (which would seem like the obvious metric) or ad #4 would be getting ignored almost entirely.

I’ve noted the type of ad in red so that you can see ad #4 is a new Expanded Text Ad while the other 3 are old Standard Text Ad formats. The bias is so strong that it ignores the fact that ad #2 would make Google more revenue because of its much higher CTR. I repeat, AdWords is favoring ETAs at the expense of Google’s revenue.

In short, I still recommend using the “rotate evenly” setting and regularly reviewing your own ad tests. If you’re too lazy, get a tool like Adalysis to help you (full disclosure, I’m a customer myself).

What ad rotation setting do you prefer? Have you had a similar experience with the “optimize for conversions” setting? We’d love to hear your thoughts and experiences in the comments!

Comments (14)

  1. How does the data look when you compare conversion per impression or CTR * Conversion Rate? That’s how they claim the algorithm makes the decision.

  2. The default is optimze for clicks, rather than conversions, as you say in the second sentence, but I know what you’re saying. If you’re actively managing a campaign, then rotate evenly is the only option I consider.

  3. Doing the math, CTR * Conv Rate would be this:

    Conversion/Impression would be this:

    I don’t see any mathematical machinations that can justify the serving without putting some kind of bonus on ETA format/discount on older formats.

  4. thanks! was just thinking about this. also thanks for the adalysis idea, ive been looking for something like that

  5. Hey Robert,

    Just out of curiosity, have you since moved to using “Optimize” ad rotation and only ETAs? If so, what kinds of data have you seen since making that changed, compared to what you were reporting in this article?

    If you have time, check out my post about how the new (Sept. 17) “Optimize” ad rotation works and let me know what you’re seeing…

Leave a Reply