The youth sports market has shown impressive growth, reaching a global value of $37.5 billion in 2022, with projections indicating an increase of 9.2% annually until 2030. This boom is reflected in the active participation of around 60 million young athletes worldwide. SkillShow, known for its video production in this segment, conducts over 300 events annually, generating content for more than 20,000 young athletes.
However, the growing demand has presented significant challenges for the company. Although it has an internal team of only seven full-time employees, SkillShow has had to rely on more than 1,100 contractors each year. This dependency has raised operational costs and extended delivery times to up to three weeks per event, making it difficult to respond to market needs.
The management of approximately 230 terabytes of video data each year has brought additional complications, including long upload times, high storage costs, and sometimes inefficient data management systems. In light of this situation, SkillShow was forced to suspend new events in early 2023, limiting its growth potential in a rapidly recovering market, especially post-pandemic, where video content for recruitment has become essential.
To address these challenges, SkillShow partnered with Amazon Web Services (AWS) to develop an automated video processing line. Initially, the team explored various alternatives for player identification but faced difficulties related to facial recognition, mainly due to the inconsistent quality of videos and lighting issues. Ultimately, they decided to opt for a solution that used audio recordings to identify and automatically generate clips based on the narrators’ commentary, which proved to be more effective and cost-efficient.
With the implementation of AWS services such as Amazon S3 for storage, AWS Lambda for serverless processing, and Amazon Transcribe for speech-to-text conversion, SkillShow was able to create a scalable system that addresses its operational issues. This innovation reduced the video production time from three weeks to just 24 hours per event, streamlining delivery for both athletes and scouts.
Recently, at an event in Chicago, the new system processed 69 clips with 93% accuracy in identifying and naming videos. At another event, the Northwest Indoor, approximately 270 clips were managed with over 90% accuracy, demonstrating the scalability and efficiency of the system.
With the possibility of further refining its processes, SkillShow is well-positioned to meet the growing demand for youth sports content while maintaining consistent quality across all its events. This case highlights the transformative potential that AWS’s machine learning services can offer organizations by optimizing workflows that require intensive resources, allowing them to focus on delivering better value to their customers.
Referrer: MiMub in Spanish