The Inside Scoop On AdWords Ad Rotation

It’s important to put your best foot forward. If you’re at a job interview, you’re going to try to say the things that hiring managers generally like. How do you know what other people like? You use data gathered through your entire life up until that point — you have life experience, and that helps you decide what you say to the world.

Just as people put a lot of thought into what they wear and say at a job interview, advertisers put a lot of thought into their ad text. A crucial part of the ad text equation that people over-think, though, is ad rotation. There are times when people experiment too long by rotating their ads indefinitely, which is like trying out a line that you think might be a loser when you’re interviewing for a job that you want. It’s in your best interest to use all of the information you’ve already gathered in the past and put your best foot forward.

The Options

As a refresher, there are four different ways to rotate your ads in AdWords:

• Optimize for clicks. As expected, this is all about driving clicks. Whichever ad has gotten more clicks in the past is more likely to show in the future. (There are some edge considerations here, which we’ll cover below.)
• Optimize for conversions. Ads that are more likely to produce conversions are more likely to show. Note that I said conversions and not conversion rate (more on this below).
• Rotate evenly. Each of your ads in an ad group receive equal preference, regardless of performance. Note that “equal preference” doesn’t mean “equal number of impressions.” Each of your ads will enter a similar number of auctions, but that doesn’t mean that each ad will be competitive enough in each of those auctions to actually get an impression. This can be especially true with new creatives, when the system is still learning about their performance. After 90 days, if your ads are untouched, they’re optimized for clicks (or conversions if you’re using Conversion Optimizer or Enhanced CPC).
• Rotate indefinitely. Same rules as “rotate evenly,” only there’s no auto-optimization that kicks in after 90 days. It’s each of your ads taking their turn at bat forever into the future.

Indefinite Is A Long Time…

…so long, in fact, that it has no end (did I just blow your mind?). Endlessly rotating your creatives can lead to lost impressions, clicks and conversions. Really, it’s just lost business.

Depending on the scale of your account, staying on top of ad tests can be a pretty laborious process — especially if you’re running those tests indefinitely and depend on some sort of internal system to know when you have enough data to determine a winner. Luckily, there are options available to make that process easy, yet a lot of advertisers don’t use those options. Why is that?

I’ve seen presentations by incredibly smart people at conferences that say rotating indefinitely is the only way to go. For some, there’s an inherent distrust of Google (my employer), and I hope to clear up a couple of misconceptions that lead to that distrust. Plenty of advertisers don’t want to optimize for click-through rate (CTR) since they’re focused on conversions, not clicks. That makes sense to me. I love conversions, and I’m sure you do, too.

So, why not use the optimize for conversions setting in AdWords? Often times, the challenge here is not the results, but how they are interpreted. It’s not uncommon, using the % served column in AdWords, to find cases where ads with a higher conversion rate are shown less frequently than other ads. While that may seem counter-intuitive, I want to explain why it’s almost always a good thing for your account.

Why Isn’t My Ad With The Higher Conversion Rate Showing More Often?

There are a few reasons that this could be the case:

1. Your test might be too fresh. As the Help Center says, “If there isn’t enough conversion data to determine which ad is likely to provide the most conversions, ads will rotate to optimize for clicks.”  The readers of Search Engine Land are probably savvy enough to spot that issue, though, so I’m assuming this one doesn’t apply to you.

2. The “optimize for conversions” setting is focused on conversions, not conversion rate. The winner of your ad tests under this setting is going to be whichever ad has the highest CTR*conversion rate. This isn’t a column in the AdWords interface, but it should be easy enough for you to calculate on your own.

   Basically, if you have 100,000 impressions, which of your ads is going to drive the most conversions? An ad with a CTR of 1.6% and a conversion rate of 4.5% is going to drive 72 conversions. An ad with a CTR of 2% and a conversion rate of 4% is going to drive 80 conversions. The ad with the lower conversion rate is going to drive more conversions for you. It’s counterintuitive at first, but once you get used to that way of thinking, the results of your ad tests will make a lot more sense.

3. Ads need to win impressions before they drive conversions (or clicks).  You might even notice a few instances in your account where CTR*conversion rate is higher for the ad that’s serving less often.  That’s a byproduct of Ad Rank — if your higher-conversion-driving ad has a wide enough CTR/relevance disparity to the other ads, it might be losing impressions that would ultimately result in you losing out on conversions.

A Couple More Considerations

While those are the three most common issues if you see data anomalies in % served, there are a couple of other things to consider.

The first is the time frame that you’re reviewing.  In most cases your ads aren’t created at the same time.  If you’re reviewing a date range that extends to a period where one ad was running but the other was not, this can confound rate metrics.  Previous ad rotation winners may be losing in the present, but the accumulation of their past performance may cloud that fact.  Make comparisons at the right date range and you should see stats that back up your ad rotation option.

Additionally, I’ve heard people say that the automated rotation options decide tests too quickly.  That they launch a new creative and it apparently loses an ad test and stops serving before it has generated enough impressions.  Sometimes ads are losers right out of the gate.  Even if CTR is similar at the start, Ad Rank might be a different matter entirely.

CTR isn’t the only way your ad is judged (remember relevance and landing page experience).  The system is going to prefer the similarly performing ad (in terms of CTR) that has better relevance and landing page experience.  I personally think of this as a benefit – if your ad text is clearly losing then you’re better off having those impressions go elsewhere.

Last thing I’ll mention — remember to investigate your segments. Performance on things like search partners might skew your results. Also, if you have somewhat disparate keywords in the same ad group you might see some noise in the data at the ad level (we talked about this a bit earlier in the Optimize for Clicks portion).

If you see something that isn’t behaving as it should, do some investigation into your segments.  The system should be working for you, and you might be able to pick up some other insights into performance along the way.  If you’ve done all this and you still believe your rotation settings are in error, let your rep know or drop me a line.

Conclusion

Rotating indefinitely is a noble thought, but I ultimately think it can hurt your account.  Optimizing for conversions should be a better setting for you.  It’s an easy enough thing to test.  Try it out in your account and see how it compares to campaigns that are rotating indefinitely.  And if you still are determined to rotate your ads, at least rotate them evenly and not indefinitely.  You have to pick a winner at some point.