Image Testing’s impact on homepage engagement: What we’ve learned so far

As more publishers add Image Testing to their optimization toolkit, our Data Science team is beginning to see the impact on homepage performance. Therefore, we wanted to analyze our early data to see how key audience metrics, such as Click-Through Rate (CTR) and Bounce Rates, are impacted as a result of testing images.

More on our findings below.

How testing images impacts CTR & Quality Clicks

To begin, we analyzed more than 7,000 image tests run by our partners. We found that, on average, our partners are testing about 3.5 images per experiment (users can conduct an unlimited number of tests). Those running image experiments produced a notable lift in both the story’s CTR as well as Quality Clicks.

Other interesting findings included:

  • There was a 38% lift in click-through rate and a 37% lift in Quality Clicks in tests where a winner was chosen. 
  • The most engaging headline is the original headline 35% of the time, but the most engaging image + headline combination is the original only 15% of the time.
  • Varying images and headlines simultaneously results in 247% higher lift in engagement than varying headlines alone.
  • CTR and Quality Clicks increase 37-38% on average with Image Testing when a winner is declared. 

Analyzing CTR and Quality Clicks by segments

We also wanted to look at differences between segments, comparing Arts & Entertainment sites and News & Media sites. 

In image-based tests across both segments, CTR and Quality Click lift hovers near the 40% mark, a significant boost to organizations trying to find new ways to engage audiences for longer periods. The Arts & Entertainment sector also tends to experiment with nearly double the images of News & Media, our analysis shows.

Overall, our data suggests that customers seem to get better at image testing over time, with increases in CTR and Quality Click Rate serving as the reward for those efforts.

What our data says about Image-based Testing and Bounce Rates

Another question we had was, “Can image experiments reduce Bounce Rates?” When we looked at the Image Testing data related to the instances where audiences load and leave the homepage without reading any stories, we found that:

  • Our Image Testing partners have seen their Bounce Rate decrease by about 14% in their first month of testing.
  • The positive effect on Bounce Rate increases over time, as editors become more proficient at testing. By the fourth month of Image Testing, our partners have seen their bounce rate decrease by about 27% compared to their average bounce rate before testing began. 

We also get asked, “Does testing just shift clicks from one piece of content to another?” Our research suggests that it does not. In fact, our partners are decreasing their Bounce Rate through experimentation — fewer readers load and leave the homepage.


  • As the number of Image Tests increase, the Bounce Rate decreases, as shown in the table above. 
  • The greater the number of  both image/image and headline tests conducted in a given day, the lower the bounce rate is on the homepage for that day.

Similar to CTR and Quality Clicks, our data suggests that Bounce Rates improve with Image Testing. Our partners also appear to be getting better at Image Testing over time, which decreases bounce rate as well.

Takeaways: Our Image Testing findings thus far

Our early findings on testing images bodes well for content creators across the board. Three key metrics — CTR, Quality Clicks, and Bounce Rate — are all improved with the help of visual experiments.

What’s more, the more you test pictures, the better the results. We’ve long emphasized the importance of making few, yet impactful adjustments given the nature of fast-paced news or content cycles. 

Our data suggests that allocating resources to image experimentation as part of your homepage optimization strategy is well worth the investment.

Contact us for a demo of Image Testing and our other homepage optimization tools.

More in Research