Writing

Notes on measurement

• 'People don't remember how they heard about you' is a testable claim Jun 10, 2026 The standard objection to survey-based attribution gets repeated as fact. It's actually a hypothesis — and the spend-share test is how you check it.
• Apple said nothing about attribution at WWDC. That's the news. Jun 10, 2026 For the first time since AdAttributionKit's introduction, WWDC came and went with zero measurement updates. Both readings of the silence point to the same conclusion.
• 'How did you hear about us' options: the list that works for mobile apps Jun 9, 2026 A copy-paste option list for mobile HDYHAU surveys, and the reasoning behind every line — what to include, what to merge, and what to leave out.
• In-app surveys: best practices, and the one survey most apps never run Jun 9, 2026 NPS and CSAT measure satisfaction. The attribution survey measures acquisition — different rules for timing, audience, and where the answers land.
• Mobile app attribution: every method, and what each one can't see Jun 9, 2026 Last-click, SKAN, probabilistic, MMM, incrementality, and self-reported attribution — what each method actually measures, and the blind spot each one carries.
• How to ask 'how did you hear about us' without ruining the answer Jun 5, 2026 The question is trivial to ask and easy to ruin — the failure modes are mechanical, and so are the fixes.
• Self-reported attribution is deterministic. SKAN isn't. Jun 5, 2026 Post-ATT, the under-used third leg of measurement is asking the user — deterministic at the install level for the channels no postback will ever cover.
• Three reports, one number Jun 5, 2026 The morning ritual of reconciling three attribution numbers that never agree, and why the survey is the only one that sees the channels with no click.