Ad Extensions: The Paradox of Choice

In the PPC industry the prevailing wisdom is that ad extensions are vital to success because they increase the amount of real estate you occupy and they affect Ad Rank.

During a recent #PPCchat I saw the following tweet:

Since 140 characters isn’t nearly enough, here is your response Mark.

The Paradox of Choice

In The Paradox of Choice: Why More Is Less by Barry Schwartz, he cites a set of studies done to determine if the mantra “the more choice, the better” was actually true. These studies were published in the Journal of Personality and Social Psychology and here is a summary:

A display was set up in a grocery with high-quality jams. Samples were available and people were given $1 off coupons if they sampled. One display had 6 flavors available for sample and the other had 24. The larger selection brought more people to the display, though people tasted about the same number of flavors. The big difference was in purchase behavior as 30% of those who encountered the 6 flavor-display purchased whereas only 3% of those who encountered the 24-flavor display purchased.

In this example, it’s possible that the number of options available increased the amount of effort needed to make a purchase decision. Most of them just didn’t buy rather than put in the mental energy to determine the best option from so many choices.


Perhaps the most frequent and consistently aggravating group decision is where to go out to eat. Whether it was my family as a kid, me and my wife now, or out with colleagues this always ends up being a stressful decision. However, I’ve found a method that works more often than not.

You structure all decisions as A/B choices.

Do you want Mexican or Italian? Do you want to eat at KFC or Arby’s?

My experience is that people are very good at making a simple A/B decision. They can determine which is better quickly and decisively. Add a third or fourth option and the whole thing blows up.

Which brings us back to ad extensions. Consider these results:


Put yourself in the mind of the customer here. You’re looking for an auto insurance quote and this is what you see. Here are my thoughts:

  • Google – Okay, Progressive, GEICO and AARP. I’ve heard of all of those. AARP has a 4.7-star rating. Click!
  • Bing – Free instant quote? Click.

The above reactions are based on a couple of assumptions. First, I assume that people don’t read unless they have to. Default is to skim. So all that text that gets added from callout extensions and review extensions and annotations and social extensions, etc. isn’t likely getting read. Second, people are in a hurry. Hence the Bing thought process where our customer clicked as soon as he/she saw a link titles “Free Instant Quote” (which just so happened to be an extended sitelink of the AARP/Hartford ad.)

I know that experienced PPC professionals are going crazy after reading that. Whether it’s thoughts of “that’s too simplefied” or “[insert single example that proves otherwise]” it likely rubs you the wrong way. But think a little deeper. Does it bother you because it’s false or because it’s not what you do?

Can We Test It?

Ideally we could set up a test to see if a vanilla ad with no extensions could out-perform the same ad with all eligible ad extensions. Unfortunately, AdWords and Bing don’t offer the control needed for this type of experiment. Sure you can load up your ad extensions and reporting shows you if people click on the extension itself or just the ad when the extension appeared, but there is still an exceptional level of fuzziness in ad extensions. Many of them are automated by the platforms even. You don’t even control what those ones say and you never control when an extension will or won’t show up.

So there you have it. My hypothesis is that in an environment where competitor ads are showing multiple extensions, a non-extended ad with a simple, clear value prop and CTA could see better performance. If you’ve got ideas on how this could be tested I’d love to hear them. Drop a comment below or reach out to me on Twitter @robert_brady.

Comments (8)

  1. Yes, testing would be great. It boggles my mind that there is not more texting or rotations with ad extensions. Recent testing of the new Structured Snippets indicated that they operate the same way as Sitelinks & Callouts, with the top/first listed ones getting served first.

    I documented it here:

    Someone should start a petition, kickstarter or something to get Goog to move on this..

  2. Hi Robert,

    Good thinking – the paradox of choice is certainly at play here. On the other hand, AdWords is all about CTR and if extensions would hurt CTR, they would use them more carefully.

    My personal hypothesis: Many people behave the way you described. They scan and click but don’t pay much attention to sitelinks or other extensions. Still, extensions matter in the sense that they make the ad bigger.

    Google may talk about more helpful ads, but in the end a bigger ad gets more attention, even if the attention is mostly on the headline. The most helpful parts are still the big blue headline and the spacing between the ads.

  3. I’m curious on seeing some eye-tracking studies on SERP/Ad density.

    A couple factors that would be interesting to note is how often users bother to scroll down the page to see all the ads vs. the ones just above the fold.

    At the very least, this effect should be seen less on mobile since 2-3 ads basically take up the entire screen.

  4. I’m actually piecing together an eye tracking study on serps at the moment. I’ll look to publish findings in the next couple of months. Hopefully it’ll help draw some conclusions from ad density and level of detail that users find helpful.

  5. “My hypothesis is that in an environment where competitor ads are showing multiple extensions, a non-extended ad with a simple, clear value prop and CTA could see better performance.”

    What about an ad that has a simple, clear value prop and CTA along with all eligible extensions? Why does the idea of using all available ad extensions and having a clear CTA and value prop have to be mutually exclusive?

  6. The idea is that an ad WITHOUT extensions would stand out because it is different. If you also accept the premise that people don’t really read the extensions, this is a more plausible hypothesis.

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.