Fewer people are playing Madden than they once did, despite the fact that NFL ratings are higher than ever. To launch Madden NFL 15, EA Sports needed the popular franchise to regain its cultural relevance and connect with a new generation of football fans—an audience who shares in the love of the NFL’s rivalries. Those rivalries unfold massively online throughout the NFL season, especially in GIFs and memes, a favorite language of the internet.

Shared GIFs took over the web as NFL fans latched onto GIFERATOR as a way to comment on live games and upcoming matches. GIFERATOR created a new language for NFL rivalry, spreading Madden content across the social web, well outside of the target audience.

GIFERATOR earned praise from traditional media and press for combining Madden and NFL content in a way that was unique and relevant to each user.

Measures of success included over 160 articles and 14MM earned impressions through traditional media. Audience engagement was 3x greater than the brand’s benchmark, and visitors created 373K GIFs to be embedded or shared through social media, text, and email. The site produced 3 visitors for every shared GIF, showing how successful GIFERATOR was in spreading its message through the fan community and beyond.

We created Madden GIFERATOR, a real-time NFL GIF engine that let fans celebrate winning moments and zing their rivals throughout the season. With each brag-worthy NFL play, GIFERATOR dynamically combined relevant Madden game footage, stinging headlines, and dynamic stats to instantly deliver share-worthy content at the height of the online conversation.

The real-time experience was powered by live NFL data and was built to find the biggest and most conversation-worthy plays. We fine-tuned our event triggers by running the GIFERATOR against game data from the entire 2014 NFL season. We found data-driven triggers for biggest scores, biggest turnovers, and most share-worthy plays.

To create a visual language that was uniquely Madden, we exported layered artwork from the video game’s developer edition, allowing us to scale our visual library to a huge pool of players across every NFL team—something that would have been unfeasible to do by hand.

To create the most relevant, dynamic headlines, we built a logic system to match players and play types with an exhaustive library of copy. The headlines were updated in real-time as they were published to include dynamic statistics, essentially creating infinite possibilities for the GIFs, and keeping the content fresh throughout the entire season.

For every NFL game, GIFERATOR generated a live stream of animated GIFs triggered by the action unfolding on TV. Those GIFs were delivered in a second-screen experience, and also published in real-time ads distributed through Google’s global ad network, target ads to the fans already looking for game content.

Each fan would see his team’s side of the story.

, fanning the flames of the rivalry as it unfolded.

By using live NFL data, GIFERATOR ads showed game data that was often slightly ahead of the content on websites where the ads were displayed. Ad units were distributed to coincide with upcoming match-ups and viewers were shown ads relevant to their favorite team at moments of peak interest before games. Affinity for specific NFL teams factored into which versions of each ad were served to consumers, based on individual search history. Ad units brought our audience to the desktop or mobile experience, where they could build their own GIF complete with a caption.

Fan-created GIFs were shared massively during live games, and used to comment on events during the game, moments after those events took place. Shared content was a combination of our pre-captured player animations, creative background images, and either our prepared captions or those written by fans. GIFERATOR scaled easily because it was built to create and display relevant content based on a live data feed during live games, rather than relying on consumer-generated content.