GameChanger

Meet Our Client

GameChanger is elevating youth athletics, creating sports tech that covers youth sporting events just like the pros. Its mobile platforms bring sports moments together for coaches, parents, volunteers, fans, and athletes. Everything from scores and stats to game streaming and highlight clips can be found in one place.

At the heart of these experiences are artificial intelligence (AI) and machine learning (ML). Volunteers keep score, recruiters study player data, athletes build portfolios, family members cheer, and coaches review film. Meanwhile, AI and ML automate processes and build comprehensive datasets to explain the story behind the action.

Their Challenge

While AI and ML are crucial to GameChanger’s business model, there was a time when the company captured live-stream assets without mining the valuable data—from body positioning to play time—hidden within that footage.

“A large part of our business is about capturing information surrounding what happens during the game,” says Leonard Grazian, Engineering Manager, Computer Vision at GameChanger. “We needed to get more out of our assets, relieve as much of the burden as we could from users, and use computers to pick up the slack.”

To improve quality and delivery of its offerings, GameChanger needed to label its video data quickly and accurately.

Starting with volleyball, the company invested in third-party labeling tools and asked volunteers to comb through the vast amounts of game data needed to drive ML models. GameChanger quickly realized, however, that this approach bred poor quality and was impossible to maintain: It was costly, burdensome, and ineffective.

“We definitely didn’t want to go through that process for every single project, and that’s when we started looking at vendors and found CloudFactory,” Grazian explains.

Our Solution

Because GameChanger’s video assets include images of children, it takes security and privacy very seriously. When evaluating potential partners, data protection was a priority. Vendors were put through a rigorous security review and asked to provide security credentials.

To evaluate technical capabilities, GameChanger also asked each labeling provider to participate in a test to determine:

• How easy is it to provide them with the data?
• Are mistakes being made by analysts?
• Is the data being returned in the requested format?

“CloudFactory stood out from the crowd for many reasons,” explains Grazian. “Their labels were very accurate, their return was very prompt, their data formatting was correct, and they surfaced a number of edge cases and asked for disambiguation, which was huge.”

CloudFactory’s Accelerated Annotation solution combines an AI-driven platform with expert data annotators who proactively identify edge cases and anomalies and provide thoughtful recommendations on how to resolve them. At the start of every project, GameChanger decides what needs to be built—models for shot detection or pitch counting, for example—and brings that idea to CloudFactory, along with a data sample.

Through integration, video data is never downloaded—it’s simply referenced, which helps GameChanger maintain tight control over the information without worrying about data leaks.

The Results

With CloudFactory’s help, GameChanger can scale its AI projects efficiently, bringing new products to market quicker, with fewer inaccuracies, while also improving the quality of existing offerings.

“CloudFactory moves so much faster than our internal team could, and their throughput is so much higher,” explains Grazian. “For half the cost, they can label 10x the amount of data in a given period of time.”

Based on input from CloudFactory, Grazian says GameChanger has even transformed its own strategies for data preparation and labeling so it will be better prepared for future AI projects.

“We have definitely increased our time to market for our new app, and that’s largely due to CloudFactory’s contribution,” says Grazian. “We’re getting higher-quality datasets at faster speeds with much better ROI.”