ArtCraft ~90k visits: how much of 31% referral is self-referral?

On public search estimates for marketing root getartcraft.com, organic is only ~8.5k (+102%) with ~338 keywords—brand signal exists, the plate is still small.

On whole-site mix for app.getartcraft.com (August 2026), visits are ~91.4k (+16.5%): Direct ~48%, Referrals ~31%, Organic Search ~8%, Paid Search ~0. Inside Referrals, ~95% comes from getartcraft.com → app.getartcraft.com—own marketing site, not a flood of external publishers.

“One-third referral” must be unpacked

Writing “~31% referral” without splitting self-referral misreads an open-source desktop tool as an off-site media machine. US organic is only ~625 (−32%), ~90% into the marketing home—not an SEO page matrix.

Product: open-source desktop engine

Positioning is an open desktop app for AI video/images with ownership framing (GitHub storytold/artcraft). Desktop share ~84% matches download/install workflows. Founder Brandon Thomas publicly contrasts director-intent tooling with rent-a-model web wrappers.

Keyword surface

• Brand variants → marketing home: artcraft ai / getartcraft.
• Spelling noise → home (brand forming, noisy).
• Model long-tail → thin product pages (e.g. seedance)—narrow SEO surface.

Monetization

Outbound shows stripe.com heavily—payment rail exists, not a P&L. Paid Search ≈0 + Direct near half means commercialization depends on return use after download, not bought SEO.

For builders

• A few hundred US organic visits ≠ growth failure if Direct and own-site referral carry the plate.
• Differentiated narrative does not auto-create SEO traffic.
• Split app domain (visits) vs marketing root (search landings) before judging acquisition.

The test is whether artists open the desktop app again. Quiet Lion.

Self-referral is a feature, not a bug—if labeled

When ~95% of Referral is getartcraft.com → app.getartcraft.com, the “one-third referral” headline is mostly an owned funnel. Mislabeling it as press coverage invents a media engine the product does not have. Desktop ~84% matches download/install reality.

Open-source narrative vs SEO plate

Ownership and GitHub positioning can win artists without winning category keywords. Root-domain organic ~8.5k rising fast still cannot replace return use on the app domain. Judge success by second opens of the desktop app, not by US organic sitting at ~625.

Payment rail without a P&L fantasy

Heavy outbound to Stripe shows checkout is wired. Paid Search ≈0 means growth is not bought search. The hard product problem remains post-install retention for precision workflows.

Quiet Lion splits marketing-root search landings from app-domain visit mix before calling any open-source tool an SEO story.