# Case Studies — Reference Operators

## Case Study 001: Matthew Gallagher / Medvi
Source: NYT, April 2, 2026
Relevance: DIRECT — same model (AI-built ecom/health business, minimal team)

### The Numbers
| Metric | Value |
|--------|-------|
| Time to launch | 2 months |
| Starting capital | $20,000 |
| Month 1 customers | 300 |
| Month 2 customers | 1,300 |
| Year 1 revenue | $401M |
| Year 2 projected | $1.8B |
| Net profit margin | 16.2% |
| Employees | 2 (brothers) |

### How He Did It
- Niche: GLP-1 weight loss drugs (telehealth middleman)
- Outsourced compliance/doctors/pharmacy to CareValidate + OpenLoop
- AI tools: ChatGPT, Claude, Grok, Midjourney, Runway, ElevenLabs
- Built AI agents for system integration
- AI customer service (with human escalation fallback)
- Media agencies for ad buying only
- No VC funding — bootstrapped and profitable
- Reinvested profits into adjacent products (men's health, meal delivery, women's health)

### Key Lessons for Shay
1. Middleman model — don't build the product, build the distribution
2. Speed > perfection — launched with AI-generated slop, fixed later
3. Hiring slows you down — Watch Gang had 60 people and failed, Medvi has 2 and won
4. Adjacent product expansion — one validated channel, multiple products
5. 16.2% net margin with AI ops vs 5.5% for Hims (60 employees)
6. First month: $20K spend, 300 customers — validation threshold hit immediately
7. Broke his own site on a hike because no one else could fix it — document everything

### The Pattern (apply to Shay's system)
Gallagher's model: Find a regulated/complex category → outsource the hard parts → own the customer acquisition with AI → expand product range once channel proven

Oliver: Reference this case study when evaluating new product ideas. Ask: "Can we outsource the hard part like Gallagher did?"

## Case Study Template (for future studies)
### [Name] / [Company]
Source:
Relevance:
Numbers: (revenue, margin, team size, timeline)
How they did it:
Key lessons for Shay:
The pattern: