Like most NFL teams, the Seattle Seahawks know the difference between wins and losses can be razor thin, turning on the bounce of a fumbled ball, one bad pass—or a running back’s slight hesitation at the line of scrimmage.
Not all factors impacting a game are controllable. But coaches and team management know that to stay competitive, they must leverage every tool at their disposal to get the most out of their players on game day. That’s why, when the Seahawks noticed a newly acquired asset in the backfield wasn’t reaching his impressive potential, they turned to the numbers.
In a league where every advantage counts, the Seahawks are now leading the charge in applying data-driven insights to every corner of the franchise—using advanced Amazon Web Services (AWS) technologies to empower players to make critical split-second decisions, enable talent evaluators to balance their own instincts with quantifiable data, and much more.
“I think, in the past, we’ve always been pretty sure we’re doing the right things,” says Sam Ramsden, the Seahawks’ Director of Player Health and Performance. “But with the processes we now have in place, supported by AWS and machine learning, we feel like we know things. We can be more confident.”
Data-driven performance
The Seahawks’ coaches didn’t wait long to act after noticing their newly recruited running back was struggling to hit the hole.
“We used our GPS data and we gathered some insights,” explains Ramsden. In the past, the player had exhibited an uncanny explosive ability. But in looking at that information from over the course of three weeks, coaches saw that the player had missed an array of opportunities to showcase that speed. Something was holding him back.
Winning in the NFL requires an elite operation that stretches across every corner of a franchise—meaning the Seahawks needed to continually find new ways to weave emerging technology into how they approach the sport. The staff enlisted the help of the AWS machine-learning tool SageMaker—a fully managed AI service that enables data science teams to build, train, and deploy machine learning models that can translate vast sums of data into actionable insights.
With SageMaker, the Seahawks were able to analyze the data collected and find patterns that could help address the issue. With this support, coaches were able to give the running back specific, targeted advice to help him achieve the explosive bursts he was known for—and all without putting more traditional pressures on his performance in a way that might lead to injury. “After about three weeks of that,” explains Ramsden, “he was feeling better and better. And the coaches were gaining more and more confidence in him.”
“He played in a game on the East Coast,” Ramsden says, “and ripped one out for 55 to 60 yards to the end zone.”
The gift of time
To prepare for the 2021 draft, the Seahawks put to use another AWS AI service: Amazon Rekognition, which offers customers pre-trained or customizable computer vision tools to analyze videos and images for insights.
The draft is a famously time-intensive process for any team; there’s even been Hollywood movies made about just the draft. Even the best scout cannot possibly manage to scan through—let alone thoroughly review—all the video that college and international football has to offer. Still, such footage is critically important. It’s where talent is often hiding—and talent leads to wins.
“All the things we’ve been able to discover using Rekognition otherwise take a lot of time for the human eye to capture,” says Suttles. “And being able to take in more data gives us a leg up—that little bit of competitive advantage that we’re looking for.”
Knowing this, the Seahawks used AWS’s tech to reimagine their methods for recruiting. First, they built a data lake in Amazon S3 combining NFL Next Gen Stats—data collected by the NFL, using AWS tools, that track trends in player performance—with real-time player health and performance data to evaluate the team’s strengths and weaknesses. Next, they used Rekognition, now calibrated to their needs, to weed through those videos and track players and their formations on the field, involvement in various plays, spacing, and more. They combined this analysis with their comprehensive team profile to identify players that could be a good fit for the team.
A cycle of ongoing improvement
These days, there is a very fine line between the NFL’s top teams—meaning even minor adjustments to team, strategy, and operation can go a long way. With AWS, the Seahawks have found a rare, competitive edge—leveraging data, analytics, and machine learning to foster a cycle of ongoing improvement for their team as well as their overall brand.
By using SageMaker, for example, the team’s management has discovered a way to optimize their starting lineup, identifying how their players might match up to those of their competitors as a way to take advantage of their opponent’s tendencies and weaknesses. During the pandemic, the Seahawks also deployed AWS’s machine learning services to monitor safety protocol for audience-attended games—analyzing things like distancing behavior, mask usage, and crowd density.
Perhaps most telling of all, the players, too, have now adopted these tools as an integral part of their game. “What I’ve seen over the last few years is, going from us having to get the players to do baseline measurements and to wear this device or capture this piece of info,” says Suttles. “To now, where the players are coming to us. They take an active interest in the reporting and what we’re capturing and are asking more questions.”
This story was produced by WIRED Brand Lab for AWS.
