How dependent are we on top sellers?
№ 038 · Supply concentration
Definition
MetricShare of GMV from the top X% of sellers (concentration curve)
Unit% of GMV per seller percentile
Pareto curve of seller share of GMV. Top 1% of sellers above 40% of GMV is concerning (those sellers can switch platforms). Build by sorting sellers by GMV and plotting cumulative share. Example: top 1% of sellers = 47% of GMV (high dependency).
- GMV
- Gross merchandise value. Total value of goods sold through the platform before returns, discounts and take rate.
Benchmarks
| Bottom 30% | Median | Top 30% |
|---|---|---|
| head-heavy and tightening: top 1% above approx 60% of GMV (TikTok Shop level) or concentration rising period over period, creating dependence and pricing-power loss to top sellers | pronounced power law is the norm at scale: on Amazon US, 1.6% of active sellers produce 50% of 3P GMV and 111 sellers produce 10%; eBay reported 80% of GMV from sellers above $10K per year | healthy power law with a deep tail: top 1% of sellers at roughly 20-40% of GMV, top 20% near 80%, and the head's share stable or falling as the tail grows |
large public marketplaces (Amazon, eBay, Etsy, TikTok Shop) 2019-2026, GMV modeled from seller feedback and units; direction of the trend matters more than the level, and smaller marketplaces are naturally more concentrated · Levels are scale-dependent: a young marketplace with 200 sellers will look far more concentrated than Amazon and that is not automatically unhealthy. The risk read is the trend (tightening head) plus single-seller dependence above approx 10% of GMV. Amazon data is modeled from feedback reviews, not disclosed GMV.
Category split omitted: Concentration curves by product category are not published by two independent Tier 1-2 sources; available data is per-platform.
When it looks bad
The Lorenz-style curve bows further from the diagonal each quarter, and the top-10-sellers share line trends up while the count of sellers above a minimum GMV threshold stalls.
Top 10 sellers move from 24% to 37% of GMV in three quarters; one seller reaches 12% of GMV and then demands a commission cut with a credible threat to leave.
What to do about it
- Set board-level dependence thresholds (no single seller above approx 10% of GMV, top 10 below approx 35%) and trigger mitigation when breached: dual-sourcing the category, cloning the seller's assortment via recruitment, or negotiating term commitments in exchange for fee tiers.
- Grow the middle class of sellers deliberately: allocate a fixed share of discovery surface to sellers outside the top decile with quality guardrails, and measure the count of sellers above a meaningful GMV floor as a first-class KPI.
- Contract the head: for sellers above the threshold, trade lower fees for 12-month volume or exclusivity commitments so concentration risk is at least term-secured.
- Watch category-level concentration separately; a diversified blended curve can hide a category where one seller is 60% of supply and a stockout or defection kills the category.
Sources
- Marketplace Pulse US Amazon: 111 sellers = 10% of 3P GMV, 1,020 = 25%, approx 8,000 (1.6% of actives) = 50%; concentration nearly doubled since 2023; approx 20% of sellers generate approx 90% of revenue marketplacepulse.com ↗
- Marketplace Pulse top 1% of sellers took 21% of reviews on Etsy, 42% on Amazon, 60% on eBay; eBay CFO: 80% of GMV from B2C sellers above $10K per year marketplacepulse.com ↗
- Marketplace Pulse TikTok Shop US: top 1% of sellers drive approx 60% of tracked GMV, top 0.1% over a quarter; more top-heavy than Amazon marketplacepulse.com ↗
- Andreessen Horowitz measure % of GMV from top X sellers; fragmented supply is more defensible, concentrated supply can walk and take transactions with it a16z.com ↗
- Andreessen Horowitz (Olivia Moore) supply-side GMV expansion of retained sellers naturally concentrates GMV; pair concentration with retention to read it correctly a16z.com ↗