3 Subtleties Of Google Conversion Optimizer You Should Master

3 Subtleties Of Google Conversion Optimizer You Should Master

Lots of advertisers have turned over their bid management in AdWords to Google’s Conversion Optimizer.  Some find it scary, some find it a relief, but everyone probably acknowledges that Google processes incredible amounts of data to optimized bids, using very complicated formulas.  Those complicated formulas make for a complicated tool that requires control with subtlety that’s very different from CPC bidding.  Here are 3 aspects of Google’s Conversion Optimizer that you need to master that require some forethought.

1) How A Conversion Algorithm’s Definition Of “Relevant Data” Can Burn You.

With the Conversion Optimizer that Google offers in AdWords, Google’s conversion algorithms  need relevant historical data to calculate their bidding strategy, ad distribution and positioning. Optimizer looks into the history of your account, but focuses mainly on the last 2 weeks of data. So if there is a recent bad week with inaccurate or incomplete conversion data or unrepresentative CPAs because of novel marketing initiatives, news stories that impact your business or extreme holiday seasonality, Google still will use that recent week’s while determining strategy.  If something is going on with your data that’s not going on with those keywords for your competitors, it’s an especially volatile situation for your advertising campaigns.

What do you do if this happens? Turn off the Conversion Optimizer and let another couple weeks of solid, good data enter the system before you turn it on again.

2) How To Add Keywords With Zero History Into Optimizer’s Equations. (They start out with NO history? So what happens, and what’s the best practice…)

Keyword expansion is absolutely possible with conversion optimized campaign, but again the process requires proceeding slowly.  The overall campaign performance may take a small hit once you initially add the keywords, especially if they’re high traffic keywords, at least until Google gets data on the new keywords. Normal stats shake a little (higher costs, possibly higher CPAs), but once Google can analyze the keywords the performance should stabilize.

My suggestion: Expand in waves.  Think expansion in stages, folding new keywords into the overall campaign mix over time.  For example, if you have an online book store and you want to add keywords add your keywords by themes, or authors first.  Let the system digest, and proceed the next week with more.

3) How To Tune An Intricately Running Conversion Machine While It’s Running. (The Answer: Delicately)

AdWords campaigns on Conversion Optimizer with CPA target bidding do NOT bounce back from dramatic bid adjustments like in traditional campaigns with CPC bidding. You lose the ability to just revert changes, and resume like you were in the past. The danger isn’t just the expense of high CPAs, it’s the possibility of a stranglehold on your traffic as Google’s permutations cut the traffic to your site while reacting to dramatic CPA shifts.

Remember to view your account holistically keeping in mind that changes other than CPA bids can affect the overall performance, too, such as concurrent account wide ad copy changes, or large keyword pruning/deletion or expansion projects.  If you see a dramatic decrease in performance, turn off optimizer and let data run through the system again.

Drop CPA Slowly – Gradually work your CPA down, and by gradual, I mean at a snail pace. You should work down the maximum CPA by 0.10 at a time until you are seeing the right results.

Summing Up:  Act Slowly And Deliberately

Big Changes – Bad Idea : The most important strategy  I can reiterate is the importance in optimizing campaigns governed by Google’s conversion algorithms, is act slowly and delicately. Be conservative. Don’t make big changes, especially dropping the CPA drastically.  Google’s ability to predict bid strategies will be compromised, and this will result in a huge loss of traffic.

By Nicole Nerad

Paid Search Manager

This Post Has 2 Comments

  1. Dan PPC PROZ


    I second Glenn’s comment. I learned so much from that short webinar, thanks.

    A couple questions:

    1. would you recommend merging/deleting adgroups with little to zero conversions? If you would leave them, would you recommend a higher cpa?

    2. To clarify… you would not recommend enabling conversion optimizer on campaigns that are actively and aggressively being ad tested? For example, I routinely copy the old winning ad, delete the old ad, with history, and run two new ads, so that they are competing on equal footing. Would this totally screw up conversion optimizer?


Comments are closed.