Google Conversion Optimizer Observations

Back at SES Chicago in December, I heard a lot of good things about Google's Conversion Optimizer, which (in a nutshell) allows you to set a desired cost per action instead of cost per click on your PPC campaigns. You only pay for conversions (not clicks), and theoretically, Google's algorithm adjusts your ad serving in order to show the ad at times when it's most likely to convert. Sounds great, right? Why aren't we all using it, then?

Well, I tried it a few months ago on a client campaign that wasn't converting as well as I thought it could. To me, this is the ideal campaign for Optimizer. However, it totally bombed. Campaign trafffic dropped to next to nothing, and we got 0 conversions - that's right, 0 - in over a week of testing. I canceled the test and decided Optimizer was flawed.

But I've been mulling it over in my head ever since. That particular client had a limited PPC budget which was well below the available search volume on their keywords. And the recommended cost per conversion that Optimizer suggested was about double what their daily budget was! I had to set the cost/conversion at their daily budget limit, which severely limited Google's ability to generate conversions & traffic. I suspected this was the fly in the ointment.

So, last week I decided to try again with a different client. This client has been doing PPC for years, and they get very high conversions (both volume-wise and conversion %-wise) because their goal is lead generation (email signup). But there was one campaign that I thought could do better. So I decided to try Optimizer again.

It's only been a week, so the jury is still out, but early results are promising. Here are some early observations:

• The campaign is getting higher impressions, clicks, and CTR with Optimizer.
• Average ad position is slightly higher (but only very slightly).
• Conversion rate is 10% worse with Optimizer, and cost per conversion is 22% higher.

I know, you're thinking "Cost per conversion is 22% higher? Stop the test!" That's what I thought, too. But there's a caveat. The client drastically reduced their budget for this campaign earlier this week (a planned reduction). The lower budget has been in effect for 2 full days. (Yes, I know my test is now skewed. Keep reading...) Observations over those 2 days:

• At the lower budget level, CTR and conversion rate are nearly double what they were at the higher budget level.
• Cost per conversion is about half what it was at the higher budget.

Of course it's too early to draw any real conclusions, but what I suspect is going on is that now that the budget is lower, Optimizer is forcing ad serving on only the best-quality searches. Impression volume is about 1/6 of what it was before, but I don't care about that. I care about conversions.

Conclusions?

• Optimizer seems to work best for campaigns with lots of conversions (i.e. way above the 30 per month minimum required to use Optimizer) and a decent conversion rate (i.e. 10% or better).
• If your budget is enough to max out impressions & clicks, you may not get as favorable a cost/conversion with Optimizer as you do with a more limited budget. However...
  
• If your budget is too restrictive (i.e. impression share less than 50%) and your conversion rate is low, Optimizer is unlikely to work well for you.

I'll be running the test for another week or so to see if these observations hold true. I'd love to hear other observations about Optimizer - post them in the comments!