Industry 02
CPI · Trial to paid · D30 retention
For subscription apps buying installs on TikTok, Meta and Google. We pay for the users who start a trial and stay, not the cheapest installs, and we fix the attribution that tells you which is which.
Installs that never open twice — Campaigns optimized to installs find people who download and forget, and CPI looks great until you check day 7.
Blind iOS attribution — SKAN postbacks arrive late and coarse, and a weak conversion-value schema hides which campaigns start trials.
Trial to paid is a black box — Ad platforms stop at the install, so budget can’t tell a trial that converts from one that cancels on day two.
Creative burns in days — TikTok audiences tire of an app concept within a week or two, and CPI drifts up with it.
Remap SKAN — We rebuild the conversion-value schema around trial start and early engagement, so iOS postbacks say something useful.
Optimize to trials — Campaigns move from install events to trial-start events through your MMP, even if volume dips at first.
Split iOS and Android — Separate campaigns, budgets and targets, because signal quality and trial rates differ by platform.
Brief for the first three seconds — Weekly creative built on one everyday use case per ad, tested on TikTok and Reels.
Read cohorts, not days — We judge channels on trial to paid and D30 retention by install cohort, not on yesterday’s CPI.
Work
Proof
from
this
industry.
Illustrative demo figures for fictional clients. Results vary with margin, market, offer and spend.
The Growth Audit shows which of these moves pays back first for you.
