Thesis

Domestic violence is a persistent and devastating issue, and technology has the potential to provide innovative intervention methods. I had tons of passion to solve a problem in that space from a technological perspective. One particularly insidious problem within this space is covert voyeurism: perpetrators monitoring data, apps and camera usage to control their victims. This behavior is not only a red flag in relationships, but can also be an indicator of existing or future abuse. The team hypothesized that creating a mobile application using facial recognition and biometric signatures to detect voyeurism could empower victims and improve their safety.

Validation

We pursued this idea with the support of an accelerator, which provided us with significant resources and guidance. We spent three months developing the concept during the accelerator program and continued working on it for additional three months afterward.

Initially, our approach focused on detecting stressful situations using smartwatch heart rate monitoring combined with other physiological and technical indicators. However, we quickly pivoted away from that idea since we encountered a significant technological challenge: if even the largest tech companies struggle to differentiate between genuine distress and routine activities like jogging, we knew we wouldn’t be able to solve it alone.

To refine our facial-recognition-and-biometric-signals solution, we conducted over 100 interviews with women and relevant organizations to understand existing solutions and their limitations. These interviews revealed that current detection tools were either too expensive, unreliable, or required technical expertise that victims often lacked. Additionally, many victims suspected they were being monitored but had no way to confirm it, leading to psychological distress and inaction.

Why it didn’t progress 

1. Lack of Viable Market: Despite extensive research, we struggled to identify a sustainable market for this solution. The problem extends far beyond a single application – addressing it effectively requires a broader focus on awareness and education, at least at this stage.

2. Technological Barriers: Operating System Limitations: Developing a product that relies on an operating system posed a major hurdle. Even minor OS updates required significant attention and adaptation on our end, making it a technically demanding and resource-intensive endeavor.

3. A Difficult Third-Party Market: We realized that a B2G (business-to-government) model would be necessary to scale, but that was a critical roadblock. Relying on government funding in impact-driven areas is risky – administration changes can abruptly cut financial support, making long-term sustainability uncertain.

Insights

1. A Solution in Search of a Problem: Ultimately, we realized that while the issue is serious, it is not perceived as urgent enough for people or organizations to pay for a solution. The business model simply does not exist in this space. When a founder spends too much time trying to force business traction, it can be a red flag that they are more focused on fitting a solution to a market rather than truly solving a problem.

2. Market Adoption Barriers: While victims and advocacy organizations recognized the value of our idea, scaling it into a viable consumer product was challenging due to distribution and awareness hurdles. Perhaps cybersecurity startups could play a more effective role in addressing this issue.

3. Unpredictable User Behavior; Our target users did not exhibit consistent consumer behavior patterns. To define what constitutes an anomaly (a woman in danger), we first needed to establish what “normal” behavior looks like – a complex challenge from a behavioral data perspective.

Resources

https://www.statista.com/statistics/1224597/snooping-on-the-phone-of-a-romantic-partner-us-by-gender

---