Case Study: Display Ads - Image vs. Text Ad for Remarketing

As marketers, we LOVE remarketing.

What’s not to love?

  1. A prospect comes to your website
  2. Your site puts a tracking device on him/her (well… a “cookie”)
  3. Your ads stalk (I mean follow) the prospect around the internet, reminding him/her how great our service or product is. Some E-Commerce businesses get fancy and even show the exact product the prospect was looking at.
  4. Prospect clicks on the ad, comes to your site and converts

So, what’s not to love? But seriously, all jokes aside…

It’s one of the best ways for businesses to increase conversions at a lower cost per conversion.  Every business should be running a remarketing campaign.

There are a number of different ways you can run a remarketing campaign. Take a look at how we chose to set up a remarketing campaign for this particular business with the goal of converting those abandoned visitors.

Search Campaign vs. Remarketing Campaign

So how well does remarketing really work? When we compared our top performing search campaign against the remarketing campaign over a 6-month period, here are the results we saw:  

Additional Leads
0 %
at Lower Cost per Click
Lower Cost per Acquisition

While it’s not an apples-to-apples comparison, we thought it was interesting. We might have missed out on 1,678 leads, if we didn’t have a remarketing campaign in place. Also, the fact that cost per click was 1/5 of our top performing campaign meant that we were able to get those visitors back at very low costs. While 44% cost per lead is great, we saw it as an opportunity to improve. 

To start, we looked at our image ads vs. text ads. 

Image Ads vs. Text Ads for Remarketing


Click thru RateAvg. CPCCPAConv. Rate
Image Ads0.35% $0.94 $67.49 1.39%
Text Ads0.65% $0.84 $86.63 0.97%

Image Ads Out Performed Text Ads

The most important metrics for us were conversions and cost per acquisition (cost per lead). We saw that the cost per acquisition was 28% lower and conversion rate was 30% higher for display ads.  

Lower Cost per Acquisition
Higher Conversion Rate

But we dug deeper, got more granular, to see if this was true at the ad group level.  

We had segmented our AdGroups based on when the person last visited our site. Our goal in using this method was to see if time elapsed since last visit had any bearing on how they engaged with our ads and converted.

Image AdGroups

Click thru RateAvg. CPCCPAConv. Rate
No Conv. 0 - 3 Days0.51% $0.96 $82.59 1.16%
No Conv. 4 - 7 Days0.30% $1.16 $44.40 2.61%
No Conv. 8 - 30 Days0.27% $0.75 $96.97 0.77%

Text AdGroups

Click thru RateAvg. CPCCPAConv. Rate
No Conv. 0 - 3 Days1.06% $0.87 $93.78 0.92%
No Conv. 4 - 7 Days0.62% $0.91 $65.15 1.40%
No Conv. 8 - 30 Days0.50% $0.79 $95.96 0.82%

While the text ads had a better click through rate, the cost per acquisition and conversion rates (the metrics we cared about) were much better with the image ads for those who visited the site less than 7 days ago. For those who had visited the site over 7 days, Text Ads performed slightly better than the Image Ads.  Because of this we left the text ad active for the adgroup “No Conv. 8 – 30 Days”.  We theorized as to why this happened, but couldn’t come up a reason that could be backed with data. 

A/B Split Testing Image Ads

For the remaining two adgroups, we actually paused the text ads and started experimenting with image ads. 

We decided to experiment with a lifestyle image.