When it comes to advertising both on TV and in digital spaces, brands have two major areas of focus: increasing the time consumers spend with their advertising content and making sure that the content is not the subject of unwanted controversy.
But with so much content being produced, and so many places for it to go, there’s a balancing act between maximizing exposure and minimizing the risk of your brand being associated with anything objectionable.
Video image recognition, often referred to as tagging or meta tagging, has progressed rapidly in the last year or so. Compute times are faster, storage is cheaper and the cloud allows for seamless upload. When it comes to tagging video that identifies profanity, sexual situations or violence, it is a matter of minutes if not seconds, and this efficiency is accelerating. But while tagging options are increasingly granular, most companies that offer this service are not going beyond what would be considered a fairly superficial identification process; implied nudity and suggestive content are able to slip by many platforms that aren’t detailed enough to distinguish such issues.
This is where a line can be drawn between surface tagging and true video image recognition.
How tagging works best is by identifying elements and then “scoring” them, which is a way to establish audience segments that increase engagement and enhance brand safety. This is a crucial extra layer of protection. For example, McDonald’s will likely want to know whether or not they are airing next to a suggestive music video. The scoring data also permits the distributor to create custom multiple algorithms within its distribution system to super serve its micro-segmented audiences. Brand safety is valuable as the only function of tagging, but audience segmentation is a critical component of serving up the right videos at the right time.
Video image recognition can also greatly mitigate losses in ad breaks within content. This applies to linear content as well as digital. Tagging by category will permit the creation of an algorithm or practice that permits a network or broadcaster to determine the best category to place in each position within an ad pod to reduce audience losses and increase engagement. From there the traffic system can automate best practices filling each break with the best possible category with the content. This is another win-win for distributors and advertisers alike, if the tagging itself is fine-tuned enough to make sure that the correct elements are being identified which will bring the most value to all stakeholders.
The ability of applied AI and machine learning to enhance the performance and moderation of content with video image recognition is a game changer, increasing time spent with content and protecting brands that advertise within the content. But there is a difference between just using those words and providing a reliable solution. Instead creators must make sure they are seeking out and using the best tools for the job available.