Kidoodle.TV Uses AWS to Scale Ad-Supported Video Streaming 3,200% in 2020

Kidoodle.TV, a children’s streaming service, needed to scale up quickly to meet a global spike in demand during the COVID-19 pandemic. Stay-at-home orders worldwide meant that more children were accessing home entertainment. To keep up, Kidoodle.TV wanted to expand its streaming capabilities while maintaining high availability and low latency across a variety of devices and regions. And because Kidoodle.TV is ad supported, the company also needed to expand ad delivery while providing age-appropriate material and maintaining high standards for quality, security, and privacy.

In 2020, Kidoodle.TV scaled from 400,000 monthly active users to over 12 million. To support this growth in content and advertising, the company expanded its use of Amazon Web Services (AWS), including Amazon Simple Storage Service (Amazon S3), an object storage service that offers industry-leading scalability, data availability, security, and performance. “For most practical purposes, Amazon S3 has effectively limitless parallelism,” says George Belotsky, senior software architect at Kidoodle.TV. For ad delivery, the company used AWS Elemental MediaTailor, a channel assembly and personalized ad insertion service for video providers.

Kidoodle.TV streams thousands of hours of shows through all kinds of devices across 235 countries and territories. The service has over 100 million app downloads worldwide and more than 1 billion ad seconds for sale every 2 days. As a streaming service for children, the company emphasizes child-safe content and Children’s Online Privacy Protection Act compliance. And as an ad-supported service, Kidoodle.TV has human reviewers vet every ad to confirm both appropriateness for children and that messaging gets conveyed in the advertisers’ intended manner. Once reviewers deem the ads appropriate, which takes minutes, Kidoodle.TV transacts the ads to the open auction exchanges in real time, at which point the ads are ready to integrate with the streaming content.

Kidoodle.TV has used AWS in some capacity since the streaming service’s founding in 2012, originally storing video content in Amazon S3. Over the years, the company has consolidated more of its stack on AWS, making a concerted effort to rebuild the service on AWS starting in 2015. “Scale, reliability, ease of use, and the frequent release of new features were some of the reasons we chose AWS over other cloud providers,” says Daniel Riddell, chief technology officer of Kidoodle.TV.

In 2020, the COVID-19 pandemic brought a spike in demand for home entertainment. Kidoodle.TV also discovered that parents, now working from home, were coviewing Kidoodle.TV with their children up to 70 percent of the time or were within earshot of Kidoodle.TV ads at least 90 percent of the time. This sudden demographic shift meant that Kidoodle.TV needed to customize its ads to appeal to parents as well as children. Looking to meet this new challenge while keeping ads appropriate for the whole family, the company added a new creative review process to its workflow. This process broadened Kidoodle.TV’s ability to moderate a wider range of ads, but it also meant more compute-intensive workloads.

Because Kidoodle.TV already used so many AWS services, it had little trouble scaling up to meet the increase in viewership and ad-related workloads. “AWS enables growth in the sense that we don’t have to worry about it or even think about it,” says Riddell. For example, the company could simply ramp up its use of Amazon Elastic Compute Cloud (Amazon EC2)—a web service that provides secure, resizable compute capacity in the cloud—and make sure it had the compute capacity for its increased workloads using AWS Auto Scaling.

AWS Auto Scaling monitors applications and automatically adjusts capacity to maintain steady, predictable performance at the lowest possible cost. “The AWS Auto Scaling group is a big thing because we have some of our own custom software built on Amazon EC2 instances, and we can just scale them up and down as demand fluctuates throughout the day,” says James Purdy, senior vice president of development at Kidoodle.TV. In addition to seeing a 3,200 percent increase in customers in 2020, Kidoodle.TV tripled its staff from 24 employees at the start of 2020 to 75 by the end of the year.

Load times for content have a big impact on viewership, so providing a short time to first frame, or the time it takes the first frame of content to reach the screen, has been a focus for this team. “Every time a kid looks at a black screen, that’s an opportunity for them to walk away,” says Purdy. Looking to keep the time to first frame down as it scaled up, Kidoodle.TV turned to Amazon CloudFront—a fast content delivery network service that securely delivers data, videos, applications, and APIs to customers globally with low latency and high transfer speeds, all within a developer-friendly environment—and was able to decrease time to first frame from 3 to 1.5 seconds.

On the ad side, Kidoodle.TV made robust use of AWS Elemental MediaTailor. “AWS Elemental MediaTailor is a core service we use,” says Purdy. “We can set up the configurations and point the AWS Elemental MediaTailor ad decision server to our home-built ad decision service, and our code takes over. We deliver industry-standard responses through video ad-serving templates, and AWS Elemental MediaTailor gives us a server-side ad insertion solution. It’s hands off: no maintenance, no headaches.” The speed of AWS Elemental MediaTailor also helps Kidoodle.TV maintain continuity of content, helping prevent viewers from losing patience as ads load. “It feels and works like Saturday morning cartoons, where you sit down and there’s this seamless transition between content to ad and back to content,” says Purdy.

Having built a strong, globally scalable solution on AWS, Kidoodle.TV is poised to continue expanding to emerging markets all over the world. And no matter where people stream Kidoodle.TV shows, the company will use AWS to maintain its commitment to user privacy and security. “Before we analyze data, we thoroughly anonymize it,” says Belotsky. “We can do that on AWS because it has all of these tools to process large amounts of data.”

To aid with the analysis of anonymized data, the company plans to explore using Amazon SageMaker, which helps data scientists and developers prepare, build, train, and deploy high-quality machine learning models quickly. Kidoodle.TV expects to use machine learning to collect ad engagement data, such as when a viewer turns off an ad, and improve ad performance for viewers on aggregate without resorting to collecting personal data from viewers and targeting them specifically.

As Kidoodle.TV continues to grow, it knows AWS will continue providing the right tools for its needs. “If you’re a great designer, you can build a beautiful sculpture with a set of building blocks,” says Riddell. “I think we have a great design team, and I think AWS is like a set of blocks. And I think Kidoodle.TV is a great product that we built with it.”