Topic: Take rate comparison e-com vs gaming; AI investment priorities
Key points:
Advertising credits are a "tiny fraction" of new customer value, immaterial to take rate; unified auction means constant take rate across all categories.
GPU purchases are planned a year in advance; infrastructure investment spikes run through P&L, not capitalized.
EU advertisers can work with APP today, but inventory for web/shop advertisers in EU is not open; EU is "low teens percentage" of business.
Q4 guidance includes no incremental assumption for onboarding new advertisers via referral program.
Mgmt stance: Neutral — take rate is structurally identical across categories; disciplined pay-as-you-go AI investment; EU expansion deprioritized vs. platform general release.
Q7 — Robert Sanderson
Topic: International availability and expansion; advertiser onboarding friction
Key points:
Web/shop ads available everywhere except EU traffic and China; app advertisers are global.
Current Western shops succeed in English-speaking markets (Canada, Australia, New Zealand); Japan is APP's second biggest market but requires local presentation.
Localization via LLMs is "fairly solved"; human behavior is universal, so model translates math globally.
Mgmt stance: Bullish — international expansion has low lift due to universal math models; success will accelerate as platform broadens.
Q8 — Martin Yang
Topic: Cohort performance in Q3; PSU issuance for engineers
Key points:
Return on ad spend for customers has improved over the past year; tooling upgraded from old dashboard to Axon Ads Manager.
Ad creative nuance (35-second average vs. social's 7 seconds) took months to educate customers; compounding effects in knowledge, usage, and tooling.
PSU pool issued in October for a group of engineers, also usable for future recruiting.
Mgmt stance: Bullish — positive trends from compounding improvements; product is 1 year old in a large market, with long runway to replicate gaming success.