Guest : Nikita Bier, Built viral apps like TBH (sold to Facebook for $30M) and Gas (sold to Discord).
Key Takeaways
- Look for latent demand – Find where people are trying to meet a need inefficiently and build a better solution around it.
- Teens are the best target for viral apps – Their habits are not fixed, they invite others to use apps, and they see each other daily, fostering viral spread.
- TBH and Gas succeeded through small, tactical wins – Positive feedback, live chat support, and targeted school launches were critical.
- Crisis management – Proactively address misinformation (e.g., the human trafficking hoax) by controlling the narrative and securing public retractions.
- Test one thing at a time – Focus on validating each core aspect of a product in stages, rather than all at once.
Podcast Notes
Early Ventures: Politify & Outline
Politify: App that showed how different political policies affect users’ finances; 4M users. Pivotal idea, but pivoted from government-focused apps to consumer apps.
Outline: Focused on building financial tools for government contracts, but Nikita realized he preferred building viral consumer apps.
Transition to Consumer Apps
After Politify and Outline, shifted to consumer apps. Started Midnight Labs, experimenting with different apps, leading to the creation of TBH.
TBH: Birth and Success
Concept: Anonymous, positive feedback polls based on a Snapchat trend.
Launched in one Georgia school – Within 24 hours, 40% of students downloaded it. Geofenced the app to manage growth.
Viral mechanic: People want to know what others think about them positively, without receiving bullying or harassment.
Building for Teens vs. Adults
Teens are more open to trying new apps and invite others to use them. As people age, their willingness to invite friends declines.
For every additional year after 13, app invitation rates drop by 20%. By age 22, user acquisition becomes much harder.
Leveraging Live Chat
24/7 live chat support inside the app to get direct user feedback and solve user issues quickly.
Helps in quickly identifying pain points and improving the user experience.
Lessons from TBH
Viral growth is unpredictable and chaotic – Prioritize based on immediate issues (e.g., servers crashing).
Key metrics: Track hourly active users rather than daily to gauge real-time engagement.
Geofencing: Limited the app’s rollout to prevent system overload while keeping growth manageable.
Selling TBH to Facebook
9 weeks after launch – Sold to Facebook for $30M. Nikita's small team of 4 surprised Facebook with how efficiently they built and scaled the app.
Big-tech Product Management
Product managers in big tech are more focused on coordination and approvals rather than hands-on product design.
Nikita believes “product management is not real” in large tech companies like Facebook, where designers often hold more influence over the final product.
Leaving Facebook and Starting Gas
Gas was a new iteration of TBH, focusing again on positive, anonymous polling but built to work within updated iOS restrictions.
Tested in the same school as TBH five years later, but with new growth strategies due to changes in user behavior and regulations.
Human Trafficking Hoax and Crisis Management
Gas faced a viral hoax accusing it of involvement in human trafficking.
Action taken: Nikita and his team contacted journalists, schools, and TikTok to correct misinformation. The strategy worked, bringing down account deletion rates from 3% to 0.1%.
Selling Gas to Discord
Gas was eventually sold to Discord. Success came from iterating on TBH’s concepts and adapting to a new regulatory environment for contact permissions.
Building Durable Consumer Apps
Consumer app retention is highly unpredictable. Most viral apps fade like “summer songs.”
Nikita admits that building a long-lasting app is more of a Black Swan event. Growth can be engineered, but durability is harder to predict.
Advice for Startup Founders
Focus on sequentially testing and validating each core feature or metric before moving on.
Eliminate variables to ensure accurate feedback during user testing.
