Deferred deep links cut onboarding drop-off 16 points
A marketplace app sent paid users to a generic post-install home screen. New users had to re-find the listing they clicked.

They wired Branch deferred deep links so the post-install session opened straight to the clicked listing, passing UTM + listing ID through the install.

Result on 90k attributed installs:
— activation (first listing view) rose from 61% to 77%
— D7 retention +4.2 points, paid CPA effectively −14%

✓ Same spend, better post-install continuity
✓ Branch links survived the install gap reliably on Android
✗ iOS deferred match weaker without a clipboard/SKAN bridge
✗ One QA miss sent 3% to a 404 deep link for a week

Verdict: deferred deep linking is the cheapest CPA cut you're not doing.
Best for: apps with content-specific landing intent.

ChatGPT 5.6 Luna и Terra подешевели

OpenAI резко снизила цены на GPT-5.6 Luna и Terra спустя три недели после релиза, сделав их заметно дешевле для массового использования. Sol осталась премиальной по прежнему прайсу, но получила Fast mode с ускорением до 2,5 раза. Вывод: компания давит ценой и скоростью, чтобы быстрее нарастить спрос и долю рынка.

➡️ Читайте на сайте: https://aff.top/blog/chatgpt-5-6-luna-i-terra-podesheveli

🧠 Ещё больше инсайтов → в канале AFF.top

Доменная зона .web делегирован в корневую зону DNS

Verisign вывела .web в корневую зону DNS: теперь это полноценный TLD, но открытая регистрация ещё не стартовала. Сначала доступ получат владельцы совпадающих доменов в .com через LRP. Вывод для рынка: хорошие EMD в .com могут стать входным билетом в .web, если успеть занять брендовые имена раньше общего запуска.

➡️ Читайте на сайте: https://aff.top/blog/domennaia-zona-web-delegirovan-v-kornevuiu-zonu-dns

🧠 Ещё больше инсайтов → в канале AFF.top

В Facebook Ads появился раздел «Conversations»

Facebook Ads добавил Conversations с автоответом на комментарии: по ключевым словам можно сразу отправлять сообщение в личку. Для арбитража это новый способ прогрева и передачи ссылки без клоаки: в креативе можно просить оставить комментарий, а заинтересованных уводить с вайта на блэк уже в ДМ. Идея спорная, но её стоит тестировать.

➡️ Читайте на сайте: https://aff.top/blog/v-facebook-ads-poiavilsia-razdel-conversations

🧠 Ещё больше инсайтов → в канале AFF.top

Killing probabilistic attribution on Android dropped 'organic' by 9% — for the better
An app trusted fingerprint (probabilistic) matching as a fallback. On Android, where deterministic GAID match is available, they audited the gap.

Disabling probabilistic for sources that supported GAID re-classified 9% of prior 'organic' installs back to the paid sources that actually drove them.

✓ Cleaner credit — paid channels stopped looking under-performing
✓ Killed fingerprint-driven double counting between two networks
✗ Reduced match on sources that genuinely lack device IDs
✗ Self-reporting networks (Meta/Google) unaffected — different pipe

Verdict: prefer deterministic on Android; reserve fingerprint for true gaps.
Best for: Android buyers with inflated 'organic' buckets.

CTIT fingerprinting exposed a sub-source faking installs: 23% clawed back
A buyer's MMP showed clean aggregate fraud scores, but one sub-publisher's click-to-install-time histogram was bimodal — a normal hump plus a hard spike under 10 seconds.

That sub-10s spike is the signature of install hijacking: the fraudster fires a click the instant a real install begins.

Isolating that sub-publisher and appealing recovered 23% of its installs as rejected.

✓ Sub-source CTIT beats account-level fraud scores for granularity
✓ Hijacking has a distinct, repeatable time signature
✗ Requires raw-data export and your own histogram tooling
✗ Networks dispute click-hijack claims harder than click-spam

Verdict: pull CTIT per sub-publisher; aggregates hide the fraud.
Best for: anyone buying via blind sub-publisher networks.

SKAN vs Aggregated Event Measurement: which iOS signal to optimize on
A buyer ran iOS through Meta and couldn't reconcile Meta's AEM numbers with their MMP's SKAN feed — they differed by 14%.

Testing showed why: AEM uses Meta's own 8-event/72h modeling, SKAN uses Apple's single postback. For the same campaign they're measuring different windows and different attribution logic.

The call: optimize the campaign on AEM (denser signal for Meta's algorithm), report business truth on SKAN-fed MMP.

✓ Two signals, two jobs — stop trying to reconcile them
✓ AEM's modeling fills gaps SKAN redacts
✗ AEM is Meta-only; cross-network truth still needs SKAN
✗ Modeled data is an estimate, not deterministic

Verdict: optimize on AEM inside Meta, report on SKAN across networks.
Best for: iOS Meta buyers drowning in two conflicting dashboards.