The core idea
Checkout providers such as checkout.stripe.com, buy.stripe.com, Gumroad, Lemon Squeezy, Polar, Creem, and Whop receive referral traffic from merchants. Similarweb exposes the domains sending that traffic. We use those incoming referrals as a reverse lookup:
- Query incoming referrals for each checkout domain.
- Treat each referring website as a merchant candidate.
- Multiply the checkout domain's total visits by each referrer's share.
- Aggregate the same merchant across all checkout providers.
- Keep only sites above the minimum estimated-visit threshold.
What this is not
This is not a payment count. A user can open a checkout page multiple times, abandon before payment, or arrive from a bookmark. Conversely, some payment flows occur on merchant-owned subdomains and are invisible to this method. The metric is therefore called estimated checkout visits.
Why not Stripe API?
Stripe's official API can only read Checkout Sessions for accounts you control. It cannot count sessions across all merchants on the internet. Any site claiming to show global Stripe checkout counts from the Stripe API alone is misrepresenting the data source.
Category definitions
- SEO: Independent products whose largest non-paid acquisition channel is non-brand organic search.
- GEO / AEO: Products getting measurable discovery from AI assistants and generative search.
- Social: Products where organic social platforms drive a large share of new-user acquisition.
- Viral loops: Products with visible sharing, referral, collaboration, or multiplayer mechanics.
- Other: Products whose primary growth source is direct, email, referral, paid, or not yet clear.
Review process
Raw candidates are not automatically treated as approved products. Before publishing, they are checked for marketplaces, large companies, enterprise B2B products, and sites whose traffic is dominated by paid acquisition. This manual review keeps the leaderboard focused on independent products.
Update schedule
Similarweb referral data is monthly. The collector runs weekly but skips a month that has already been collected, preventing unnecessary API credit use. Email reports are weekly; the underlying dataset changes monthly.