Get started Bring yourself up to speed with our introductory content.

Computer vision holds the key to disrupting TV advertising space

For the last several years, the television industry has experienced a significant decline in advertising spent as marketing budgets have shifted toward online advertising. This decline is fueled by several factors, including the decrease in subscriber growth as streaming content over over-the-top devices continue to become more popular and the decline in cost-per-thousand dollar amounts advertisers are willing to spend on TV ads.

Online advertisement currently offers advertisers several main benefits over TV advertising:

  1. User profiles — By using cookies, online advertisers are able to gather significant data on the user and offer the most relevant ads.
  2. Real-time targeting — Based on the profiles created, online ads are delivered immediately to the relevant audience to maximize conversion. For example, if the user is in the market for a new car, relevant car ads will be shown.
  3. Transparency — Online tools offer advertisers clear reporting of the amount of exposure of each ad, which allows for clear ROI decision-making.

As a result of this shift, TV advertising must go through an evolution of its own. The industry needs to adapt so cable providers and networks can offer advertisers a better and clearer return on their investment to win back advertising budgets.

Historically, cable and satellite providers have relied on rating companies to learn about the consumption habits and preferences of their customers by using a sample of homes that represent the market. This audience measurement data is used to help determine which ads to place at which times and their cost based on exposure. While this data has proven to be informative, it is still subpar compared to online. First and foremost, ratings are a sample and provide an indication of viewership, but not actual numbers. It is also not real time. Moreover, ratings do not reflect accurate viewer statistics, such as the user’s level of attention on the content consumed or how many people were in the room at the time the content was displayed or their demographics.

The TV industry should move toward personalized, targeted real-time advertising and abandon the traditional way of gathering viewer analytics.

Computer vision and AI-powered personalized advertising

Embedded computer vision technologies can accurately provide viewer analytics and ratings to cable and satellite providers, TV networks and advertisers.

These technologies have the potential to offer the following insights:

  • Real-time audience measurement — Detecting the presence of a user, or group of users, for reporting real-time viewership and ad exposure.
  • Viewer recognition — Recognizing the face of returning viewers to enable customization based on past behavior and predefined preferences.
  • Viewer demographics — Recognizing viewer age and gender for demographic segmentation and content personalization.
  • Viewer attentiveness — Tracking the user’s head pose direction to monitor the viewer’s attention on the displayed content.

For advertisers, computer vision allows for the most accurate reported viewership numbers. With computer vision, the user’s head pose is detected, indicating not only presence, but also the user’s attentiveness. This is a game changer for advertisers looking to get the most out of their ads as they can easily determine their targeting based on age, gender and user viewing history, and can pay for actual exposure. This is a transition from passive impression to actual viewership, including how many people viewed an ad, their age and gender, as well as their viewing histories. It would also provide more accurate information about an ad’s cost per impression, allowing advertisers to increase the relevancy and cost-efficiency of their targeting.

Computer vision also offers the consumer an improved real-time viewing experience, for example, showing child-friendly content/ads when a child or family is detected. Face detection using computer vision can offer a similar experience such as logging into your account on Netflix. The face detection technology can automatically detect an individual and allow him to pick up a show exactly where he left off or offer more relevant content to watch based on his previous viewing preferences.

How the technologies work

In order to display truly personalized content, data is gathered through a sensor in the user’s home, which provides cable operators and advertisers with analytics about which ads and programs viewers are watching in real time. The sensor also benefits the user by allowing personalized content and gesture-based interaction with the TV. The touch-free gesture control allows users to change the channel, control volume and more using simple hand gestures for a natural and seamless experience with the TV which can replace the use of the remote control.

Additionally, because it’s an embedded technology, viewership data is fully secure and private as it is processed using custom-designed, cloud-free processing that does not store images, nor does it send them to the cloud. All data runs locally using proprietary embedded computer vision algorithms, allowing for full privacy and minimizing latency.

Bringing computer vision to TV

With these technologies and the right back-end infrastructure, advertisers can work together with cable operators to slice and dice the information and automatically deliver truly personalized experiences with real-time reporting. While the computer vision systems exist and are ready for market, there is still work to be done by the operators and advertisers on the back end for these technologies to be implemented. Cable operators’ systems and the advertising delivery ecosystem need to be redesigned to be able to deliver this type of real-time personalized, targeted content and ads. While it will require major effort to change the traditional systems, the long-term financial benefit will be well worth it in the end, as it brings TV advertising in line with web and in-app advertising.

Transform TV to a world of “cookies”

Websites already gather data from users to allow advertisers to target their chosen profile using cookies, and this has proven to be successful. If a user visits a clothing website, she’s likely to see ads for it the next time she logs onto her personal social media accounts or visit another website. This happens because of artificial intelligence-based programs pulling from data pools to automatically make ad buying decisions for brands based on demographics and cost-versus-benefit.

Targeted TV advertising would work similarly, and in some ways bypass internet advertising with the amount of insights on the user, and with specific ads appearing based on the viewer’s viewing preferences, their age and gender. This means you and other people may be watching the same show but receive different ads based on your viewing history, demographic data and more. The more personalized ad delivered, the higher the conversion and the return on investment.

While addressable TV advertising is still in its infancy, it could be the technology that fuels the TV industry. In order to do so, cable and satellite providers need to embed computer vision technologies as an end goal to their advertising partners. With targeted advertising, the days of advertisers relying solely on program ratings could be history.

All IoT Agenda network contributors are responsible for the content and accuracy of their posts. Opinions are of the writers and do not necessarily convey the thoughts of IoT Agenda.

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.

-ADS BY GOOGLE

SearchCIO

SearchSecurity

SearchNetworking

SearchDataCenter

SearchDataManagement

Close