Cost Savings, Shorter Development Times, Better Performance
With DoiT International’s technical advice and AWS’s flexibility, Minute Media now has full control of its platform’s development roadmap. It also experienced significant savings compared to its previous solutions. Weiss estimated using Amazon S3 and Lambda is saving the company around $10,000 to $12,000 per month on infrastructure costs, compared to its previous provider.
Minute Media has reduced development times for products, too. “AWS Elemental MediaConvert offered every feature that we wanted—all we had to do was configure it,” says Weiss. “As a result, we were able to get our video player in the market quicker.”
The new platform performed better, as well: it enabled the delivery of 70 percent more video views and provided 90 percent more capacity to upload new videos from partners and content providers, compared to the previous online video player. This has given Minute Media the capacity to grow and add more partners. And for partners, it means a more efficient process for providing new content with higher video quality to end users.
Building a Smarter Semantic Engine
With its platform firmly in control, Minute Media set out to deliver on a common request from partners: a way to automatically discover and link to related videos.
The available options provided results that were “too random,” explains Weiss, and often promoted videos unrelated to the original one.
“The challenge was to provide related content dynamically and in real time, with the best performance possible,” he says. “With online video, every millisecond matters—it can mean a delay in serving content and ads to users.”
Based on the video player’s success, Minute Media again chose a combined approach to developing a semantic engine: building certain components in-house while relying on AWS for others, and working with DoiT to identify the best choices for the project.
For the semantic engine, Minute Media developed a matching algorithm while using machine-learning capabilities in Amazon Comprehend to identify and categorize patterns in the video content. Together, AWS Lambda and Amazon S3 balanced loads at peak times to minimize lags in video delivery.
The semantic engine proved 85 percent accurate, compared to around 70 percent for competing products. According to Weiss, it gave Minute Media an advantage over rivals and contributed to winning more than 20 new partners.
The development model of building solutions in concert with AWS works well for Minute Media. “We’re a grown-up startup,” said Weiss. “And it’s something that startups like, to build their own solutions.” At the same time, Minute Media appreciated DoiT’s help in deciding which AWS services would be best for its needs.
“When DoiT advises an AWS service, I know it will work well, and it will be stable. If I want to enrich it with a piece we’ve developed, we always get support from DoiT and AWS to ensure the parts fit together.”
Scaling to Meet Shifting Demand During Lockdown
In the changeable world of video delivery, “It all comes back to scaling,” says Weiss.
The ability to scale bandwidth and compute power became essential to Minute Media during the pandemic. As large sporting events were canceled, Minute Media’s sports-focused publishing partners saw demand for video drop. At the same time, traffic on user-generated content sites increased, as more people turned to connecting online when in-person meetups were restricted.
The AWS pay-for-what-you-use model allows Minute Media to manage costs as its workloads shift and gives it the flexibility to continue to onboard new partners looking to cut costs in an uncertain economic environment.
“We pass on the savings from running such an efficient infrastructure to our partners,” said Weiss. “In the end, everyone is happy.”