---
title: TikTok as a UA channel for apps — broad targeting, web-event campaigns and seasonal budgets (2022 view)
slug: tiktok-ads-for-apps-web-campaigns-2022
summary: Reflectly's 2022 TikTok playbook (Ryan Thorpe) — broad targeting, optimising a landing-page event instead of app installs (downloads at roughly half the price of CPI campaigns, by his estimate), skipping an MMP on a mostly-organic portfolio, concentrating spend in September/January, and Spark Ads with micro creators instead of big influencers. Dated — read with the 2025–26 caveats.
domain: mobile-apps
tags: [user-acquisition, tiktok-ads, web-campaigns, spark-ads, seasonality, influencers]
related: [creative-format-bucket-testing, boosting-organic-winners-spark-and-partnership-ads, web2app-ad-optimization-events, skan-aem-capi-for-app-campaigns, content-creators-vs-influencers, apple-ads-scale-ceiling]
sources: [yt:RE2x5f4oUlM, yt:kXXFtl0ocU0]
updated: 2026-10-24
---

# TikTok as a UA channel for apps (2022 view)

> **Dated source:** recorded August 2022, early post-ATT, pre-Andromeda. Treat the tactics as ideas to retest, not current platform rules.

## TL;DR
- Ryan Thorpe (Reflectly, ~25 wellness/productivity apps) used **broad targeting**. TikTok's algorithm learns the audience from the creative and from who converts.[^1]
- He ran **web campaigns optimised for an on-page web event** (a landing page with an App Store button) instead of app-install campaigns. He estimated downloads at **~half the price** of pure CPI campaigns at scale.[^2][^3]
- Reasons he gave: less competition than in install campaigns, the pixel still learns iOS users, and people who tapped through to the store "are just going to download".[^4]
- Tracking was "maybe 75–80% accurate" without an MMP. Most paying traffic was organic, so paying an MMP to track it made no sense for a 25-app portfolio.[^5][^4]
- **Pulse budgets seasonally.** Pull back in summer and push in September and January, when health/productivity intent and ROAS peak.[^6]
- **Micro creators + Spark Ads over big influencers.** Boost a video from a 400–2,000-follower creator who looks like your customer. Thorpe thinks most big influencers are overpriced for ~95% of apps.[^7]

## Details
**Web-event maths (his illustration).** Send 10 people to the web page for ~$1–2. If half download, 5 downloads cost well under $1 each, vs ~$2 per download on a CPI campaign. These were off-the-cuff numbers, and his headline estimate was "maybe half the price", not a measured result.[^2][^3]

**Why TikTok in 2022.** As Meta weakened post-ATT, spend moved to Apple Ads and TikTok. TikTok acted "like Instagram 10 years ago": shorter attention but higher engagement. The same video can run as organic and paid, which raises the odds of going viral either way.[^8]

**Portfolio angle.** Thorpe argued it can be cheaper long-term to **acquire apps with existing users** and cross-promote than to spend $1M on tracked UA when you might track ~20% of it.[^9]

**Seasonality as a strategy.** With imperfect tracking, concentrate budget in the periods where you know ROAS is strong rather than spending evenly all year. Host Jonathan Fishman added that some advertisers grabbed market share when CPMs fell early in COVID.[^6][^10]

## Disagreements / caveats
- **2025–26 view:** Samet Durgun ranks TikTok last for subscription apps (after Meta, Google, Apple Ads). Successes skew Gen Z, and you mostly see the winners; "90 plus percent will be failures".[^11]
- Optimising a top-of-funnel web event conflicts with today's advice to optimise as deep as volume allows ([[web2app-ad-optimization-events]], [[meta-qualified-trial-optimization-event]]).
- Running without an MMP is defensible only when paid is a minor channel, as here.
- Thorpe's "celebrity likeness" voice tactic drew a cease-and-desist (see [[creative-format-bucket-testing]]).

## See also
- [[creative-format-bucket-testing]] — Thorpe's creative process
- [[boosting-organic-winners-spark-and-partnership-ads]] — Spark Ads in 2026
- [[people/ryan-thorpe]]

[^1]: yt:RE2x5f4oUlM @ 06:42 — Ryan Thorpe
[^2]: yt:RE2x5f4oUlM @ 08:24 — Ryan Thorpe
[^3]: yt:RE2x5f4oUlM @ 09:10 — Ryan Thorpe
[^4]: yt:RE2x5f4oUlM @ 09:58 — Ryan Thorpe
[^5]: yt:RE2x5f4oUlM @ 07:35 — Ryan Thorpe
[^6]: yt:RE2x5f4oUlM @ 11:36 — Ryan Thorpe
[^7]: yt:RE2x5f4oUlM @ 19:41 — Ryan Thorpe
[^8]: yt:RE2x5f4oUlM @ 04:04 — Ryan Thorpe
[^9]: yt:RE2x5f4oUlM @ 02:30 — Ryan Thorpe
[^10]: yt:RE2x5f4oUlM @ 12:24 — Jonathan Fishman
[^11]: yt:kXXFtl0ocU0 @ 04:49 — Samet Durgun