From zero to $1M ARR in 10 months: how Sunflower built a sobriety app

Executive overview

Most app founders chase scale through dozens of micro-influencers and complex deal structures. Koby Conrad took the opposite approach: hire the single biggest influencer in your niche as a full employee, give them equity and a mission, and let compounding buy-in do the work.

Sunflower, an AI-powered sobriety app, hit $1M annualised revenue in 10 months by combining three things: an embedded influencer who generates 50–100k monthly link clicks, an AI companion that has handled 4 million messages in a year, and a paywall experiment that lifted revenue 48% with a single change.

The core insight: the quality of your marketing relationships — not the quantity of your channels — determines how fast you scale.

Influencer strategy: hire, don't sponsor

• Find the single dominant voice in your niche; ignore everyone else until you have them.
• Offer salary, equity, and a genuine mission — not a per-post fee.
• An embedded influencer understands the product roadmap, iterates on what drives clicks, and is far less likely to defect to a competitor.
• Being someone's largest sponsor — whether an influencer or conference — triggers disproportionate care and preferential treatment.
• Bootstrapped alternative: position the creator as the hero and co-build an app for their existing audience (closer to a 50/50 partnership than a sponsorship deal).
• VC capital lets you approach top-tier creators before you have revenue; without it, build your own audience first.

The AI companion: Sam

• Sunflower's AI sponsor, Sam, started as a basic ChatGPT wrapper; it is now a full agent with memory, follow-up scheduling, safety escalation, and cross-app context.
• The most surprising early finding: a 60-year-old grandmother with 30 years of alcoholism became a power user because the AI was the first entity in a decade to make her feel "validated and worthy of love."
• Sam has handled over 4 million messages — more than a human therapist could process in a lifetime.
• Safety: a VP of therapy reviews flagged conversations; biometric locks and data erasure protect users with abusive partners.
• The companion alone is not the product. Without the surrounding ecosystem (timer, community, courses, teletherapy), engagement and retention collapse.
• Competitive moat: OpenAI and Anthropic cannot fine-tune for behavioural health outcomes because they lack the outcome data. Sunflower can track relapse, days sober, and hospital readmissions and use that to optimise the whole system.

Product: what actually drives retention

• The highest-value feature is not the AI — it is the sobriety timer. Over 1.2 billion people worldwide have a job-to-be-done around tracking time sober.
• Build the smallest possible version first; ugly and live beats polished and unreleased.
• AI now removes the opportunity cost of exploring multiple product directions simultaneously — build everything, measure, trim.
• The teletherapy layer under development will let Sam flag likely relapses to human therapists and assign follow-up homework.

Paywall optimisation

• Starting point: free app with no monetisation (inspired by Duolingo's long runway). AI token costs forced a rethink.
• Key iteration sequence:
  1. Add an annual subscription option alongside monthly.
  2. Make the annual plan the default selection.
  3. Switch to a toggle where monthly is opt-in, annual is default.
  4. Offer weekly/monthly/annual so users self-select.
• Price sensitivity matters: lowering price for Sunflower's cost-sensitive addiction audience increased revenue.
• Adding a trial reminder screen produced a +48% revenue lift, a +46% trial conversion rate increase, and a +70% increase in annual subscription revenue.
• The reminder UI pairs a friendly brand character (a cartoon bee) with the message "we'll remind you before your trial ends" — reducing anxiety before hitting the paywall.

Fundraising and positioning

• First pitch deck failed: it led with the founder's past marketing track record and asked investors to believe in the playbook.
• Successful deck: led with the future market, proved product-market fit with revenue (not usage), included letters of intent from institutional pilots, and showed fast ARR growth.
• YC, Flybridge (led seed), a16z Speedrun, and founders of Revenuecat, Beehive, and SaaStr are on the cap table. Total raised: $6.5M.
• Raising off users without revenue is extremely difficult in the current environment.

The future of app monetisation

• Software creation costs are approaching zero; subscription revenue will be under sustained pressure from commoditisation.
• Physical goods tied to digital milestones (e.g. sobriety milestone pins) may outperform subscription revenue in mission-driven apps.
• Therapy and clinical services — not app subscriptions — represent the bulk of spending in the addiction market.
• Long-term vision: a free Duolingo-style tier at hundreds of millions of users becomes leverage for legislative advocacy on addiction policy.
• The bulk of addiction sufferers (90%) never access any professional care; a multilingual AI companion is the only scalable path to reaching them.