How Hunter Isaacson built viral consumer apps with 150M users

Executive overview

Most viral apps die after the first wave. NGL survived by embedding itself into a platform's social graph — not by fighting for attention independently.

The core strategy: graft a new product onto an existing distribution network, reduce onboarding to seconds, and design a viral loop so tight that even churned users return.

Building on someone else's graph is the unfair advantage most founders ignore.

NGL's origin and viral trigger

• Started Halloween 2021, the same week Instagram opened story links to all users
• Built the app in one month; six months passed with almost no users
• Tested influencers at $100 a time, trying to seed the product across different networks
• First real signal came from a low-income country — but they lacked Android, so growth stalled
• Built Android quickly; one TikTok went viral and activated global growth within a week
• The viral TikTok worked because it was controversial and emotional — a fake "gotcha" message screenshot

Why the viral loop held

• NGL's link lives in bios and stories, so even churned users encounter it again through friends
• High redownload rate sustained long-term growth without continuous acquisition spend
• Onboarding was 15–20 seconds: drop user straight onto the core screen, no friction
• North Star metric: how many users post the link and get at least two replies
• CTA "get your own messages" outperformed "download this app" by a wide margin
• Kept the product almost unchanged after going viral — no extra tabs, no bloat

Revenue model

• Multi eight-figure annual revenue from a primarily free product
• Weekly subscriptions priced $1–$7 depending on country GDP per capita
• Only a small percentage of users pay — volume makes it work
• Paid features: message sender insights, profile view data, additional link games

The platform-arbitrage framework

• Find an existing social behaviour on one network; switch the distribution mechanism
• NGL replicated anonymous messaging (YOLO, Yik Yak, Ask FM) but used Instagram's graph — the largest available
• Same framework works across any major network: Snapchat, X, Facebook, WhatsApp
• Smaller or newer graphs can also work — the viral loop quality matters more than platform size
• Apps that failed after initial virality often added complexity; NGL stayed simple

Other apps and lessons

• Wink (70M downloads): let people swap Snapchat handles via Tinder-style swiping — solved Snapchat's weak friend-suggestion graph
• Zoom University: double-dating app for college students; 40K waitlist with a 7:1 female-to-male ratio by leaning into the pandemic meme
• Bags: crypto trading app; solves token discovery via influencer feeds, group chats with native trading, and Apple Pay on-ramp; earns 1% fee on transaction volume
• Crypto PvP gaming will work eventually — the failure so far is poor product focus, not the concept
• Every failed app taught a design or distribution lesson; the test-flight graveyard is part of the process

Scaling from $10K MRR

• Split budget into small experiments across UGC, influencer posts, and paid ads
• Test a dozen creatives; find what format works before scaling spend
• Solve product-led growth in parallel: what percentage of new users invite a friend?
• CPI is effectively lower when a share mechanic compounds acquisition
• Always include an invite screen — a friend's recommendation is an endorsement
• Once a format and a share loop both work, scale both together