Costco Wholesale Corporation — Through the Warren Buffett Lens
Thesis
Costco is not a retailer. It is a membership subscription business disguised as a warehouse club. The company earns the bulk of its profit from annual membership fees, not from selling goods. This is a critical distinction. When you buy a share of Costco, you are buying a stream of recurring cash flows that grow as members renew and as new warehouses open. The merchandise is sold at a minimal markup—essentially at cost—which builds trust and drives traffic. The moat is wide: low prices that competitors cannot match due to Costco's scale, efficient supply chain, and the simple fact that its profit model does not depend on merchandise margins. Members pay for the privilege of shopping, which creates a powerful alignment—Costco's incentive is to keep prices low and quality high to retain members. Over the long term, the business compounds capital at high rates because each new warehouse generates predictable returns and the membership base is sticky. The company has no debt to speak of, and management thinks like owners. This is exactly the kind of business I can own for decades without losing sleep.
Key Value Drivers
- Membership fee growth (renewal rates and new sign-ups)
- Same-store sales growth (traffic and average ticket)
- New warehouse openings and their ramp-up economics
- Incremental return on invested capital on new warehouses
Key Risks
- E-commerce disruption from Amazon or others could erode traffic
- Wage inflation or supply chain disruptions squeeze thin merchandise margins
- Membership saturation in core US market limits growth
- Management succession after Craig Jelinek's eventual retirement
Key Metrics to Monitor
- Membership renewal rate (US/Canada >92%, global >90%)
- Return on tangible equity (sustainably >20%)
- Free cash flow yield (normalized, after maintenance capex, >4%)
- Warehouse-level operating margin (steady ~3.5-4%)
- Inventory turns (>11x annually, reflecting efficiency)
Want to see how different lenses come to different conclusions?
Run your own thesis →