Get inside the Sharks' heads — before you dive into the tank.

Purpose

The goal is to identify the real drivers of Shark Tank pitch success by combining structured financial/demographic data with transcript-level language analysis.

We uncover which factors — valuation, ask size, founder characteristics, communication style, tone, pacing, sentiment, keyword emphasis, and investor question patterns — actually correlate with getting a deal.

We then turn these insights into a usable tool where founders can upload their own pitch transcript and receive AI-driven feedback and a predicted deal likelihood based on historical patterns.

Assumptions

• NLP features reflect meaningful communication signals Sentiment, pacing, question patterns, confidence markers, and keyword emphasis extracted from transcripts are assumed to represent valid predictors of investor response.
• Unobserved factors exist Variables like editing, personality fit, off-camera interactions, and entertainment value are not included and may influence outcomes.
• Investor questions in transcripts are complete We treat the captured Q&A as fully representative, even though TV cuts out entire question threads for time.