Moment Matcher for Sports

Inspiration

The opportunity for sports adversing is massive: billions in ad spend seeking better returns, trying to define a specific moment to resonate their ads with audience. Sports provide a unique advantage in defining the moment. The established rules and clear objectives of a game create universally understood events that serve as definitive markers for high-impact moments.

What it does

The product watches the main video and identifies moments of peak emotion, like excitement or happiness, to find the most impactful time to show an ad. The system analyzes your available ads and ranks them based on how well they will connect with your specific target audience, matching the best ad with the perfect emotional moment.

How we built it

Our approach was to build the contextual ad-matching system as a dual-path pipeline that analyzes ads and video content in parallel before converging at a final integration stage. We began with a Persona Analysis by implementing a method to score each ad from the perspectives defined in the Persona Config. Next, we added a Similarity Analysis algorithm to evaluate for content relevance and audience alignment. To synthesize these results, we developed a Composite Scoring function that combines all the metrics to produce a final Ranked Ad List. With placing the ads, we integrated an external Video Analysis API to perform an Emotion Timeline Analysis, which allowed us to extract timestamps and generate emotion scores for the content. To pinpoint the most valuable insertion points, we leveraged the Gemini API to build our Ad Opportunity Detection feature. This process identifies peak emotional moments and generates placement suggestions based on that emotional context.

Challenges we ran into

Our first challenge was turning abstract emotions like "excitement" into simple data our system could reliably detect. Our biggest challenge was building the "brain"—our matching algorithm—to know when to match an emotion and when to contrast it to get the best result.

Accomplishments that we're proud of

We successfully built an automated system that matches the right ad to a specific audience segment. Our Persona Analysis engine analyzes and ranks ads based on how well they align with predefined audience profiles. We achieved a fully integrated solution that connects the right ad with the right moment. Our final output is a timestamped ad placement plan that pairs the best ad for a segment with the moment of peak emotional engagement, explaining exactly why they match.

What we learned

We learned that the "perfect moment" for an ad is not just a single event, but a peak in the video's emotional arc. Identifying the best ad opportunities required us to build a system that analyzes the entire Emotion Timeline, rather than just detecting simple actions.