Managed vs pure marketplaces

A pure platform matches and gets out of the way (Craigslist-ish, early eBay). A managed marketplace inserts ops: vetting, scheduling, quality control, even employing or heavily standardizing supply. Some "Uber for X" businesses take W-2-like control — which can trigger regulatory classification fights.

Many companies start managed to guarantee quality, then productize ops into software to climb toward platform margins. That journey is the business.

Unit economics on two sides

SaaS CAC pays for a logo. Marketplace CAC is often two-sided: you may pay to acquire supply and demand, plus ongoing incentives.

Contribution ≈ (take rate × GMV from cohort) − incentives − payment fees − support − fraud losses

A city that looks profitable on gross take can be negative after driver bonuses and refunds. Track contribution after incentives by geo and category. Frequency matters: weekly transactors can afford higher CAC than once-a-year transactors, unless AOV is huge (real estate vs rides).

Regulatory and labor classification (contractors vs employees) can rewrite the COGS structure overnight in mobility and delivery. Treat legal constraints as part of the model, not a footnote.

Gurley's factors, used as diligence

Bill Gurley's classic essay argues that not all markets are equal. Practitioners still use the themes:

1. Large, fragmented supply and demand — no single player can dictate terms.
2. High transaction frequency or high AOV.
3. Inherent pain / broken status quo.
4. Price transparency and discovery value.
5. Transaction / payment flow through the platform.
6. Network effects.
7. A path into adjacent verticals.
8. Defensibility — trust, brand, data, switching costs.
9. Frequency vs ticket size that still supports CAC payback.
10. Regulatory and ops burden you can actually bear.

Use the list as a checklist, not a scorecard that guarantees success. Uber-like ops intensity is a feature of some markets and a death sentence in others if you lack capital.

Common mistakes

• Staying managed forever and wondering why margins look like services.
• Judging a city on gross take after incentives, refunds and fraud.
• Treating Gurley's list as a pass/fail exam.

Check yourself

1. Why start managed if it hurts margin? To guarantee quality and fake supply during the cold start — then automate toward platform margins.
2. Gross take looks fine in a city. What else do you subtract? Incentives, payment fees, support and fraud. Contribution after incentives is the real number.

Key takeaways

• Pure platforms have software-like margins and a harder cold start.
• Managed players buy quality and supply with ops, then must automate.
• Marketplace CAC is two-sided. Track contribution after incentives, locally.
• Gurley's factors help judge whether a market wants a marketplace.

Further learning

All Markets Are Not Created Equal (Bill Gurley)Since Benchmark’s investment in Ebay 15 years ago, we have been fascinated by online marketplaces. Entrepreneurs accurately recognize that the connective tissue of the Internet provides an opportunity to link the play…Above the Crowd

Back to the map: marketplace economics.