TL;DR — Datenix aggregates GMV estimates from EcCompass (primary), then cross-validates every top-GMV shop against SimilarWeb monthly visits to filter mis-attributed brand-level revenue. Result: verified 36,383 · suspicious 498 · excluded 3,051.
Third-party GMV estimators often mis-attribute brand-wide revenue to national sub-sites. Example: gamestop.md (Moldova) shows $2.2B in raw data — but the site has zero SimilarWeb traffic. That's a mistake.
We define a shop as legit when: GMV / (12 × sw_visits) ∈ [$10, $500] per annual visit — the typical range across real e-commerce (Chanel $8/visit, GameStop.com $11, Underarmour.com $31, Skechers $34).
| Label | Condition | Action |
|---|---|---|
| verified | GMV/annual-visit ∈ [$10, $500] | keep raw GMV |
| no_sw_data | GMV < $100M, no SW coverage | keep raw (small shops OK) |
| unverified_no_sw | GMV > $100M, no SW data | 0.3x discount for ranking |
| suspicious_ratio | GMV/visit > $500 | gmv_verified = 0 (excluded) |
| excluded_subdomain | workshop./staging./cdn./b2b. etc | excluded entirely |
Beyond traffic, we also cross-check product count (from Shopify /products.json): shops with fewer than 5 products claiming >$10M/year get capped at $10M. Shops with fewer than 20 products claiming >$100M get capped at $100M.
Every /api/shops/domain/{d} response includes both the raw gmv_2025 and the sanity-adjusted gmv_verified plus a gmv_sanity label. Consumers should use gmv_verified for ranking and reserve gmv_2025 for audit.
| # | Domain | GMV 2025 | Monthly visits |
|---|---|---|---|
| 1 | walgreens.com | $95.97B | 61M |
| 2 | verizon.com | $83.39B | 43M |
| 3 | tesla.com | $55.23B | 27M |
| 4 | lowes.com | $52.15B | 150M |
| 5 | t-mobile.com | $37.71B | 127M |
| 6 | tjmaxx.tjx.com | $34.53B | 15M |
| 7 | hp.com | $32.96B | 47M |
| 8 | bestbuy.com | $31.26B | 111M |
| 9 | nike.com | $31.17B | 113M |
| 10 | dollargeneral.com | $25.22B | 14M |