Baseball team digitizes media, uses AI to uncover metadata

The San Francisco Giants' digitization and metadata tagging project helps the baseball organization revisit old content and deliver history to engage fans.

Many sports and entertainment organizations have thousands of old media assets in analog formats collecting dust in their vaults -- often with no idea what they might find on them. But digitizing and tagging this media with appropriate metadata can solve that problem.

In 2018, the San Francisco Giants baseball team got the green light to digitize approximately 16,000 digital media assets accumulated over the team's history to coincide with its 60th anniversary, including U-matic 3/4-inch tape, film reels and audio cassettes.

Digitizing the media was step one; using AI to uncover metadata and tag the files was step two.

"Without the metadata [on these files], we would have no way of understanding what is being digitized," said Paul Hodges, vice president of content entertainment with the San Francisco Giants.

The digitization process

The Giants are working with Iron Mountain, an information management company based in Boston to digitize their files. Once the files come back in a digital format, the master files are sent to a digital asset management system the Giants house on premises -- Square Box Systems' CatDV.

Copies of those files are also sent to the Veritone cloud, where it lives in the vendor's digital asset management system. The Giants' original intention was to move all of the media with tagged metadata back to the on-premises system at Oracle Park, the team's home stadium. But the pandemic has forced video editors to work from home and rely on Veritone's Digital Media Hub to fulfill day-to-day media tasks.

San Francisco Giants' Paul HodgesPaul Hodges

The Veritone system automatically runs those assets through transcription and facial recognition, using a face-training library created from rosters and other data sources. Veritone's built-in AI tools uncover metadata -- players' names, dates and time correlations of when they show up in each asset -- and automatically tag the files with that information.

"We did the math and found that we would have had to hire 13 to 16 interns for a year to log all this metadata," Hodges said. "So instead, we took that money and pivoted into AI."

Uncovering metadata and tagging content used to be a very manual activity, but much of that can now be automated by AI, said Alan Pelz-Sharpe, founder and principal analyst at Deep Analysis.

Deep Analysis' Alan Pelz-SharpeAlan Pelz-Sharpe

There are a number of vendors that have AI and machine learning services, including Google Cloud Vision AI, Microsoft Azure Cognitive Services and Amazon Rekognition. There are also other smaller, more specialized companies, including Brandwatch, Hive, GumGum and Clarifai.

The Giants have been working on the digitization process since June 2018, but the project is currently on hold in the wake of the COVID-19 crisis. Like many organizations affected by the pandemic, priorities and resource allocation have shifted, affecting the archive digitization timeline. The sports organization still has another six or seven months of work left to do.

Benefits of digitizing media

The digitization process is enabling the Giants to uncover content that hasn't been seen in a long time, such as a 20-minute video on Willie Mays giving batting practice tips.

"A huge part of our business is just recalling history and reminding fans what has happened on this day," Hodges said.

The lockdown has forced many organizations to mine the vaults. While it was not worth the effort in the past to digitize some of these assets, it is now.
Alan Pelz-SharpeFounder and principal analyst, Deep Analysis

Having digitized files at its disposal was especially helpful for the sports organization when COVID-19 caused live sporting events to go on hiatus. The Giants were able to recall old games and deliver them to fans via virtual streaming events, Hodges said. The team has delivered this newly digitized content to fans via YouTube, Facebook, Twitter and on TV.

The Giants' newly tagged digital files were also helpful when a TV network called asking for footage of a Hall of Famer who died. Within five minutes, the sports team was able to find the requested footage and deliver it to the station.

"The lockdown has forced many organizations to mine the vaults," Pelz-Sharpe said. "While it was not worth the effort in the past to digitize some of these assets, it is now."

Digital media enables customer engagement

Sports is built on content, said Liz Miller, vice president and principal analyst at Constellation Research. As digital engagement has ramped up, smart teams have developed roadmaps to lean into taking the "product" of the games and iterate moments to delight fans from a multitude of sources.

Constellation Research's Liz MillerLiz Miller

"That has always been the road[map], but COVID-19 put it on warp drive," Miller said.

As fans sit at home, locked down and looking for entertainment, digital content and engagement is in massive demand, Miller said. But the trick for the Giants will be how this new digitized content is packaged and delivered to fans.

Sports organizations should blend present day and classic content to keep fans engaged and talking, Miller said. For example, current players could provide commentary on classic games.

"Smart teams are finding ways to thrill their fans, keeping engagement going for what is essentially the longest offseason ever," Miller said. "COVID-19 took live events off the table and brought business as usual to a screeching halt, but it did not stop the momentum and need for fan engagement."

Next Steps

SharePoint Syntex automatically uncovers documents metadata

How automated content tagging improves findability

Seeking truth in crisis times shows importance of metadata

Baseball's Rangers seek analytics edge with Tableau

Dig Deeper on Content management software and services

Business Analytics
Data Management