By Megan Reichelt, Country Manager, SEA at Integral Ad Science
In the dynamic advertising landscape of 2023, the paramount challenge for advertisers is ensuring the safety of their brand within the sea of content and ensuring its alignment with suitable and relevant content. The rapid expansion of content consumption, powered by the influx of millions of videos to the open web each year, underscores the pressing demand for the ad tech industry to devise sophisticated technologies that keep pace with this evolution. In an era of ceaseless innovation, we need to arm marketers with an arsenal to steer their brand safety and suitability, while ensuring productivity.
As we delve deeper into the artificial intelligence (AI) and machine learning (ML)-driven age, marketers have access to cutting-edge tools backed by AI/ML, that reveal largely unused insights by third-party adtech vendors across video, audio, and text content. Being the leading global media measurement and platform that delivers the industry’s most actionable data to drive superior results for the world’s largest advertisers, publishers, and media platforms, IAS wants to help marketers protect and grow their brands by providing them with the most actionable brand safety and suitability data across all digital channels, including Connected TV (CTV). To aid marketers in their quest to protect and grow their brand on digital media, IAS has developed an industry-first advertiser solution powered by our Multimedia Classification technology, providing reporting aligned with the GARM (Global Alliance for Responsible Media) Brand Safety & Suitability Framework.
What is Multimedia classification?
Multimedia classification technology refers to the use of artificial intelligence and machine learning algorithms to analyse and categorise multimedia content such as images, videos, and audio content. It enables the classification of multimedia data based on its content, context, and other relevant factors.
This technology involves training models with large datasets to recognize patterns, features, and characteristics within different types of multimedia content. These models learn to identify objects, scenes, text, and other elements present in the multimedia files. The classification process can be based on various factors, including visual analysis, audio analysis, sentiment analysis, and text analysis. For example, in a 12-minute video, IAS measures over +3,500 visual cues for an accurate brand suitability score or content categorization, allowing you to stay informed and take action.
To ensure advertisers can grow their brands confidently, IAS maps this diverse multimedia content to IAB and GARM safety, suitability, and contextual segments.
How does it apply to advertisers
Multimedia classification technology allows advertisers to target their audiences more effectively. Backed by AI- and ML-based technologies advertisers can gain valuable insights, and this information enables them to deliver targeted advertisements that align with the interests and preferences of their intended audience.
Multimedia classification technology aids in content optimization and personalization. Advertisers can leverage the data generated from multimedia classification to refine their creative strategies, tailor messaging, and improve overall campaign performance. This technology empowers advertisers to deliver more relevant and engaging content to their target audience, enhancing the user experience and driving better results.
It helps ensure brand safety by automatically identifying and filtering out inappropriate or unsuitable content. This ensures that advertisements are placed alongside suitable and brand-aligned content, reducing the risk of associations with harmful or offensive material.
How marketers benefit from it
With the help of Total Media Quality, advertisers benefit from:
- Frame-level Video Analysis – combining image, audio, and text at the video level – providing you with best-in-class brand safety and suitability scoring (*an IAS differentiator)
- GARM Industry Alignment – granular classification with 11 categories and 4 risk thresholds, in line with Global Alliance for Responsible Media (GARM) Brand Safety and Suitability framework (*an IAS differentiator)
- Unmatched transparency to gain insights on the brand safety score of the actual videos/programs next to which CTV ads appear, including show title, by leveraging the IAS Multimedia Tag
IAS’s Total Media Quality for CTV solution is paving the way for a transformation
IAS’s Total Media Quality for CTV solution provides advertisers with the most granular video-level brand safety reporting by analysing the videos frame-by-frame using a combination of image, audio and/or text signals (vs. solely metadata), and scoring them against the GARM categories and risk levels.
Having Total Media Quality on platforms on CTV is a much-needed innovation: advertisers will be able to monitor the brand safety and suitability not just of the CTV apps, but also of the video/programs next to which their ads appear. This gives them a whole new level of data to inform their strategies and budgets, while not limiting scale with an app inclusion/exclusion list approach.
Why shall marketers make note of multimedia classification
Brand safety remains a high concern for advertisers who place their ads on digital platforms. 9 in 10 advertisers said they will be prioritising brand safety in CTV in some way, as reported by eMarketer in 2022.
Amid escalating demands for transparency in advertising expenditure, multimedia classification emerges as a revolutionary technology that promises to alleviate advertisers’ concerns. Greater transparency empowers publishers to thoughtfully curate inventory packages and establish pricing strategies. As multimedia classification technology becomes increasingly refined, not only will it navigate ads away from damaging content, but also position them alongside content conducive to consumer engagement. This paves the way for brands to serve ads that are more in tune with their brand and brand message, reflecting the preferences of their consumers, and ultimately fostering stronger connections.
It is time to look beyond metadata analysis since it only looks at textual information about the video (such as genres, rating, thematic elements etc.); multimedia classification looks at the video file itself and analyses its multimedia elements (audio, images, text). To put it in simple terms, the difference between the two methodologies is like reading a Rotten Tomatoes review and actually seeing a Movie.