Why Multimedia Classification Holds the Key to Unlocking the Advertising Potential of CTV and AVOD
Consumer demand for streaming services is growing rapidly and shows no signs of slowing down. We are no longer awaiting the arrival of AVOD (Advertising-Based Video On Demand) streaming services and CTV devices: such platforms are fast becoming the media of choice for consumers.
Across the EU5 markets, between January 2020 and May 2021, traditional TV viewing rose by 17%, while CTV consumption increased by 55%. Just like the shift to mobile before, the rush of consumers to CTV and AVOD is entrenched and will continue to grow into the future.
However, while ad money is following consumer habits, there remain concerns amongst advertisers about the quality of ad placement.
Multimedia classification solutions that are able to classify video content - in the same way that is done for text on web pages today - hold the key to solving the impasse of transparency and brand safety for advertisers.
The Rise of AVOD
Think of a snapshot of a sports match or news segment: ‘reading’ video that contains countless frames with multiple layers on screen and includes audio is the first challenge. But as the technology develops and the uptake increases, advertisers will be better placed to make the most of changing consumer habits and to gain their attention.
The exciting potential for advertisers looking to spend in CTV and has been underscored by the abundance of streaming platforms launching an ad-supported tier. A consensus is growing that the long term health of streaming will be sustained, in part at least, by advertising.
The numbers support this logic: across Europe, streaming is preferred by 71% of audiences over linear TV, and brands are increasing their spend on CTV and streaming. According to a report by eMarketer, global CTV ad spend will double in the next five years from $16.6bn in 2021 to $32.6bn in 2026.
At the same time, however, agency professionals have identified brand safety and CTV transparency as a leading challenge of buying CTV inventory programmatically.
Advertisers Demand Better Transparency
Streaming platforms are committed to delivering a surpassed TV experience for advertisers, a claim that speaks directly to advertisers’ concerns. For all the attractions of AVOD and CTV, advertisers are still drawn to the transparency of linear television. AVOD does not yet offer the clarity of linear spend, as uncertainty lingers over when and where ads will be served.
Speaking to a panel at New York’s Advertising Week, Omnicom Media Group’s Kelly Metz precisely articulated just what advertisers felt they were missing when compared to linear spend:
“The beauty of traditional linear television was we knew where we were running, it was on a schedule, we could see the programs, we understood exactly what we were buying on behalf of our advertisers.” In contrast to this transparency, Metz believes that streaming services are lacking. “The way the ads are being sold and packaged [on streaming services], we do not have the visibility we need, we do not have the trust that we need across these different environments.”
In Focus: How Multimedia Classification Works
Thankfully, multimedia classification presents a framework by which this so-far absent trust might be provided.
AI-based Multimedia Classification promises a transparency and granular detail to advertisers that can change the game, providing the tools to fully harness the power and reach of CTV and AVOD.
Multimedia Classification will allow advertisers to unlock video-level transparency, emboldening brands to enter the AVOD space with confidence, brand safety, and better media quality.
To provide this trust, it is necessary to go beyond app-level and work towards building content-level brand safety solutions at scale, and this is exactly what multimedia classification facilitates.
It maps video-level data to brand safety categories, which gives advertisers granular analytics to see which video channel and content are brand safe and suitable for their ads. More accurate than still image or text-based solutions, which are overly reliant on metadata, multimedia classification can track video files itself and audio frame-by-frame.
This tracking provides granular analysis of the content that brands are being served alongside, providing a foundation for transparency and trust between advertisers and publishers.
Take the example of a football match.
Multimedia classification is able to perform an analysis of the scene, scanning the players’ faces, the logos on their shirts, the adverts encircling the pitch and the commentator’s speech.
In this way, multimedia classification can put together a comprehensive picture of the match’s content, and a clear environment for brands.
Frame-by-Frame details are all collected, as even the letters in a logo can be identified with Optical Character Recognition (OCR) technology. Furthermore, audio track analysis can detect a range of information from language to sentiment of the commentator(s) during a clip.
Providing the Transparency and Flexibility Advertisers Need
Multimedia classification will not be a one-size-fits-all technology.
The information that the technology gathers will offer unparalleled granularity, but also flexibility: brands will be able to set the information against the IAB and GARM brand suitability categories associated with the video, or choose to follow their own specific criteria. The technology can be adapted to meet what advertisers need and expect.
So, the brand advertising on a comedy channel that might want to be more lenient towards obscenities in a standup routine will be able to set their own limits. A brand will be able to avoid a competitor or any negative feeling regarding their specific industry.
In this way, brands have access to transparency and control that will provide the security and safety that has been so far absent.
The importance of this granular detail control for advertisers cannot be overstated. In their report into CTV in Europe, the IAB noted that ‘audiences have become more fragmented and elusive than ever before.’
For this trend to abate, stakeholders from across the advertising ecosystem will have to collaborate. Multimedia classification will help: providing advertisers with in-depth understanding of video content will bring more of them on board, opening the door to more collaboration and cross-industry solutions.
With Multimedia Classification, Everyone Wins
The wins to be gained through the granular control facilitated by multimedia classification has already been demonstrated in the end-to-end roll out of the technology on social platforms.
A three-tiered machine learning system, analysing video, audio, and text, provides advertisers with comprehensive awareness of the material that their ads are being served next to as well as a targeting pre-bid solution.
As multimedia classification technology becomes more refined, not only will ads be served away from harmful content, but next to relevant content that makes consumer attention more likely. This contextual targeting will increase the performance of ads, whilst also improving consumers’ experience.
Magnite’s ‘CTV Is for Everyone’ report found that ‘80% of CTV viewers said they are more likely to pay attention to ads from brands that are in line with their lifestyle and opinions,’ a statistic which affirms the gains on offer for advertisers.
Contextual targeting predicated upon multimedia classification data will serve ads that align with consumers’ tastes, boosting performance.
We have reached a crucial milestone for CTV and AVOD, as consumers and content providers make the space their own.
However, in order to deliver the surpassing ad experience promised whilst keeping advertisers and publishers on side still requires a great deal of work and collaboration across the industry.
Brands are demanding transparency, security, and trust, and to achieve a healthy AVOD ecosystem, solutions to these demands will have to be found. With increased transparency, publishers would also be able to curate inventory packages and dictate pricing strategies.
Multimedia classification can form the foundation for these solutions, providing the technological framework within which the potential of advertising on CTV and in streaming may be harnessed.
It is the technology that will provide precise detail, assuaging advertisers’ concerns whilst also boosting the value of publisher’s inventory. There is still much work to be done across the industry, but with multimedia classification, utilised correctly, everyone can win.