Measuring creator and UGC attribution

TL;DR

• Early days: read the spike. With no other marketing, a creator posts, you see installs and revenue that day, and you divide revenue by views to get RPM.¹
• At scale (Cal AI had 50–70 creators posting on one day): weigh secondary signals. A 1M-view video with 4× the comments, likes and saves of another likely drove the spike. Combine signals into a renewal score.²³
• Use incrementality-style checks: does the spike match the creator audience’s geography and demographics?³
• Vanity domains (flightcheatcode.com, buyabusiness.com) that redirect to the app separate your organic campaign from paid and other agencies, and they’re easier to say than the brand name.⁴⁵
• Per-creator promo codes / affiliate links: one Cal AI creator video sometimes drove hundreds to thousands of code uses.⁶⁷
• Judge cohorts on churn months later, not sign-ups: SMB Market’s high-view campaign showed 45% churn after 2–3 months.⁸

Details

Spike reading (Jake Castillo). As daily volume grows, spikes get subtler, but you can still learn to read even small ones. He calls the best app marketers “experts in data analysis” because organic attribution is murky.¹² Coconote (Zack Hargett) did the same: when a video is going viral for a couple of days, measure the down-funnel impact on those days.⁹

Vanity URL mechanics (Dawson Gibbs). About 90–95% of viewers are on their phone in TikTok/Instagram, so the domain sends them straight to the App Store listing (a QR code on desktop). Use it when the brand name isn’t memorable; skip it for household names. In late 2026 his agency was still testing this on a few apps.⁴⁵

Codes and links. Affiliate tools (Insert Affiliate) use the strongest automatic attribution available and fall back to a short code typed in the app. That covers viewers watching YouTube on a TV who can’t click.⁷ Apple offer codes are a quick alternative with limits — see Affiliate programs for subscription apps — commissions, tracking, recruiting.

Paid boosting gives cleaner data. Spark Ads with an app-install anchor give a direct cost per install, and SideShift ranks videos by comment sentiment and the CTA’s quality.¹⁰

Self-reported attribution. An onboarding “How did you hear about us?” question also helps here; see Hard paywall only for paid traffic (gating by "How did you hear about us?").

Disagreements / caveats

• All of these are approximations. Several viral posts on one day, paid running in parallel, and App Store search halo (people search the brand later) blur everything. Treat renewal decisions as judgement backed by data.
• Promo codes undercount: many buyers never enter one (see the affiliate “spillover” in Affiliate programs for subscription apps — commissions, tracking, recruiting).

See also

• Pricing influencer deals — RPM cap, flat rates and minimum view clauses — what to do with the RPM number

Footnotes

1. How Cal AI turned creators into a growth machine — Jake Castillo, Cal AI @ 36:15 — Jake Castillo ↩ ↩²

2. How Cal AI turned creators into a growth machine — Jake Castillo, Cal AI @ 37:51 — Jake Castillo ↩ ↩²

3. How Cal AI turned creators into a growth machine — Jake Castillo, Cal AI @ 39:07 — Jake Castillo ↩ ↩²

4. The 1 Billion Views Content Strategy Every App Should Steal in 2026 @ 11:57 — Dawson Gibbs ↩ ↩²

5. The 1 Billion Views Content Strategy Every App Should Steal in 2026 @ 14:37 — Dawson Gibbs ↩ ↩²

6. Meet The Marketing Genius Behind Cal AI @ 22:05 — Alfie Dickens ↩

7. The affiliate marketing playbook for apps @ 04:00 — Michael Butler ↩ ↩²

8. The 1 Billion Views Content Strategy Every App Should Steal in 2026 @ 17:11 — Dawson Gibbs ↩

9. $6.7M ARR, No Paid Ads, and an Exit to Quizlet in 2 Years – Brett Bauman & Zack Hargett, Coconote @ 28:04 — Zack Hargett ↩

10. I Studied 1,000,000 UGC Ads, This Is What Will Grow Your App @ 31:13 — Drew Levin ↩