Google Pitches TV-Like Quality for Online Ads

BUDAPEST -- Digital TV Central & Eastern Europe -- Although navigating the online video ecosystem is complex, Google can help, according to Google's Strategic Partner Manager Michal Stefanski.

Speaking at the Digital TV CEE event in Budapest Wednesday, Stefanski introduced Google's DoubleClick Dynamic Ad Insertion platform, which he promised would "deliver TV-like ad experiences for online video, at scale."

Stefanski began his presentation pointing out that video consumption is growing rapidly, with time spent on YouTube growing 100% every year. He also stressed that viewing is shifting to mobile devices, with 177 minutes a day spent using a smartphone and the average user taking it out 150 times a day.

Advertisers are following eyeballs, with online video advertising spend growing at a 43% CAGR (compound annual growth rate) while TV advertising is barely growing at 2% CAGR, according to Stefanski. Although it should be pointed out that in terms of total revenue, TV advertising still dwarfs online video ad spend and will continue to do so for several years to come.

Stefanski also stressed that video ad buying is shifting to programmatic, in other words, using an automated way of buying and selling advertising inventory. Google (Nasdaq: GOOG) estimates that 27% of all online video ad buying will be programmatic this year, and that will rise to 44% by the end of 2020.

But the ecosystem for video delivery is very complex, requiring support for a variety of screen sizes, connection speeds and advertising formats among other variables. Google's solution is to simplify the entire process by having advertisers push their content onto the Google cloud, integrate it into its ad server and then use the DoubleClick Dynamic Ad Insertion platform to insert personalized ads to enable reliable streaming on any device.

One of the biggest challenges for advertisers, noted Stefanski, is the delay in switching from the actual video content stream to the advertising stream. This can result in buffering, interrupting the viewing experience (as if the mid-roll ad wasn't interruption enough).

The DoubleClick digital ad insertion stream manager manipulates the manifest in the adaptive bit-rate (ABR) streaming protocol, stitching the content stream and the ad stream together to make its delivery seamless. And all analytics and "ad-decisioning" elements are retained behind the delivery.

"It's a simplified structure," said Stefanski, "and it smooths ad delivery."

Google used this in the field with French broadcaster TF1 during the Euros soccer tournament, delivering three different ads on three different platforms based on the varying user profiles for those devices. Now Stefanski wants broadcasters in central and eastern Europe to embrace it.

Stefanski also pointed to Google's use of machine learning as a tool for broadcasters distributing content online. The ability to gather data automatically and develop more precise insight into viewing behavior and content itself can help broadcasters in a number of ways.

For example, it can recommend specific clips to stitch together to create personalized highlights and promotions. Machine learning also helps create better, more granular metadata, which can enable more precise content searches. It also includes a speech API, allowing for automated closed captioning, and the platform can be trained to learn the specific requirements of customers to address several other use-cases.

"This can remake the media value chain," enthused Stefanski. "Everything can be powered by machine learning."

— Aditya Kishore, Practice Leader, Video Transformation, Telco Transformation