Test #011: How GIF Advertising Impacts Clickthrough Rate

EXPERIMENT QUESTION: If we use GIF's instead of static images in our carousel advertisement, can we increase the incentive to click on our content?


Waterloo Sparkling Water is an Austin-based beverage company that produces zero calorie, naturally-flavored sparkling water. Named for the 1830s-era Texas village that would later become Austin, Waterloo has gained national attention and demand (outperforming LaCroix in various taste tests). 

Waterloo has seen strong results on their national 'Store Rollout' campaigns, when it comes to total Cost-Per-One thousand people reached (or, CPM). However, they found that they had room for improvement in their post engagement rate. In other words, they were reaching a wide audience, but few people were actually engaging with their posts. They wondered: Will GIF "videos" increase the incentive to engage and, therefore, increase post engagement rate?

Screen Shot 2018-05-23 at 8.28.25 AM.png


This control version shows the normal 'Available' imagery we've used throughout our national store launch campaigns. There is image of the product and a banner highlighting the store's availability with a 'Now Available' image headline. Here are the observed results: 

Control: 171,085 people reached, 363 clicks (0.21% Action Rate)






COSTCO_1080x1080_GIF (1).gif


This treatment version uses the carousel ad format. This ad format actually contains the 'Now Available' image, as we as three product images. You'll also notice that there is a more prominent headline on each carousel 'card.' The observed results for this version were:  

83,636 people reached, 128 actions (0.15% Action Rate)




This experiment did not reach the needed 95% statistical significance level, and therefore the results from this test are invalid. In other words, it's apparent that the GIF ad format did not decrease or increase ad clickthrough rate.

Even though these results are statistically insignificant, the test does bring to light something important. By using GIFs instead of static ads, we were able to also include video views into our overall engagement rate. Therefore, even though our clickthrough rate did not have a statistical difference, our overall engagement rate experienced an increase of just over 2,000%. If nothing else, this is provides a helpful indicator for the time spent viewing our posts. Facebook counts 3-seconds of a user viewing a video as a 'video view.' Previously, we didn't have a good indicator of how long individuals were viewing our carousel ads. Now we can add a new level of tracking to establish a total time spent on each ad (based on the 'video view' metric').

Concluding thoughts: 

When GIFs became popular again, they consumed the internet. They are a great hybrid of image and videos, and definitely worth testing. But when it comes to using GIFs vs. static images in Facebook advertising, we can't say we complete confidence that it will increase or decrease our clickthrough rate. However, we found in this test that it did not significantly reduce clickthrough rate and added a new tracking benefit in the process: The ability to track total time spent viewing an ad. 

Grant Klembara