Unlocking Potential: Harnessing Computer Vision for Media Effectiveness in Human and AI-generated Content

By Scott Hannan, VidMob SVP Corporate Development & Partnerships

As the VidMob team and marketers from around the world converge on Cannes this week it will be a perfect storm of technology, data and machine learning. AI has taken center stage, capturing the attention of industry experts who recognize its immense potential to revolutionize media effectiveness and amplify its power.

Every day, more than 328 terabytes of data are created – and much of this is from online activity. This explosion of data provides us with incredible detail about what content resonates with audiences, including advertising creative. What’s more, with Amazon Web Services (AWS), we have cloud computing that can scale incredibly effectively to help make sense of this data quickly and accurately. And finally, connectivity between systems allows us to harness insights and apply them across many channels.

At the same time, creative teams are buckling under the strain to keep up with the enormous demand. People continually shift their habits, moving from a few top TV channels and websites to thousands of places, from social media to connected TV. Creating the content needed to be “everywhere” is an incredibly time consuming process. And marketing teams are always under pressure to “prove it or lose it” – prove the value of their marketing investments or lose future budget. These stresses span from creative development to media buying to optimization and campaign analysis. Advertisers want to get ahead of the data influx, harness that data, and move away from strictly manual work – and they want insights to prove that it’s worth it.

It’s no surprise that in these strained circumstances, many marketers are excited that AI can help. The generative AI tools today are fast, many are free and scalable, though at the same time, may deliver misaligned or uninspiring content. Nick Law of Accenture Song said it more pointedly, that “mediocrity is now free.”

Computer Vision Unlocks the “Why”

Cannes showcases the opposite of mediocre creative. Here, incredible creative is on display, though it’s not always clear just why one creative is more valuable than another, or what elements within a piece of creative are helping drive a brand’s KPIs.

Enter computer vision. This is a first step and an application of AI that enables computers to understand the specific elements and makeup of images and videos. Computer vision is foundational to powering insights derived from creative that are associated with campaign performance. It can dramatically increase the output and scale of a talented human team through being able to process impossibly large data sets across brands, channels and regions pulling “needles out of the haystack”. These can include factors like product or logo placement, color, setting and more, empowering creative teams to confidently update or modify creative guided by data. And with AWS, they can deliver that content at scale, leveraging automation.
I should clarify one point – creative doesn’t drive campaign success – effective creative does. Ipsos finds that 75% of campaign success is driven by creativity. A Kantar study finds that quality creative and effective campaigns drive four times the performance of other campaigns. And Meta recently reiterated that 56% of all auction outcomes can be attributed to creative (Optimize Creative Strategies to Drive Business Results).

Unlock Creative Value with Amazon Rekognition and VidMob

Amazon Rekognition is an innovative service from AWS that opens new frontiers in image and video analysis, enabling advanced object recognition at scale.

VidMob has leveraged Amazon Rekognition for over 5 years, processing millions of assets that ran against trillions of impressions such that our understanding of performance and quality creative continues to grow exponentially. We initially analyzed simple creative elements such as people, colors and product or logo placement. Through the development and addition of custom models, our understanding now extends to much more complex campaign elements such as emotion, pacing, story arcs and cultural trends which are more in line with an advertiser’s approach to building creative – and extremely powerful levers for creative teams to use.

Through our relationship with AWS, VidMob can provide clients with a rich understanding of “why” creative works – at a detailed level that can be immediately incorporated into a design to help brands stand out in a world of inevitable content abundance. Together with AWS, VidMob delivers a powerful approach to tackling the demand for effective creative, no matter the digital channel or format.

Fresh Eyes on AI

No matter the level of noise, or uncertainty about how you might use AI in the future, don’t lose sight of the basics and how AI enabled optimizations can reliably help you improve and maximize return on ad spend, today. For any new creative being developed it’s worth asking:

• Am I proud? Does all of my creative do a good job representing my product or service?
• Am I aligned? Is the creative on-brand, reinforcing the broader company mission and beliefs?
• Am I set up for success? Is the creative optimized for each unique platform algorithm and formats?
• Am I learning? Am I systematically paying attention to creative data to improve every ad – where one can literally pinpoint what might be helping or hurting the objective?

With VidMob and AWS, creatives can confidently answer “yes” to each of these important questions.