Allocation pattern · B2C SaaS

Finance App: 60-Day Churn 65% on TikTok vs 35% on Meta

A seed-stage finance app came in two months ago. Personal budgeting subscription, $6.99 monthly. $1.6M ARR. Paid running 60 percent on TikTok at $4.20 cost-per-install, 30 percent on Meta at $6.80 CPI, 10 percent on Apple Search Ads. The dashboard story was “TikTok is the winner.”

The cohort math told a different story. Of paying subscribers acquired through TikTok, 65 percent had churned within 60 days. Meta cohort churn at 60 days was 35 percent. Same product, same offer page, same paywall design. Two different audiences inside the channel.

The diagnosis (3 sentences)

TikTok was producing trial starters who fit the demographic but not the psychographic. The creative angle was a viral “save $200 in 30 days” hook that converted curiosity into installs, but converted curiosity into a 60-day churn pattern because curiosity is not the same as commitment. Meta was producing slightly more expensive subscribers who actually stayed.

The allocation move

  • Reallocate 25 percent of TikTok budget to Meta over a 30-day shift. Meta cohort retention twice as strong means the same dollar produces 2x the 90-day revenue.
  • Keep TikTok in the mix at the lower share, but switch its job from acquisition to top-of-funnel creative angle development. TikTok is the cheapest place to test new hooks at scale. Validated hooks then move to Meta where the audience converts and retains.
  • Switch TikTok optimization event from install to 30-day retained subscriber if the platform’s pixel and server-side setup support it. Pulls TikTok’s algorithm toward audiences that retain, not audiences that try.
  • Rebuild the creative angle library for the finance category against a buyer-state matrix. The “save $200” hook addresses one buyer state (acute problem-aware). Add ritual-driver angles (weekly money check-in habit), identity-driver angles (people who manage their money), and value-driver angles (cost of inaction).
  • Set a new hurdle rate: any channel below 50 percent 60-day retention gets paused unless the CAC is at least 50 percent below the program average to compensate. Decision rule pre-committed.

When this applies. When it does not.

Applies: B2C SaaS subscription apps running 2+ paid channels at $5K+ monthly each, with cohort retention infrastructure in place. 60-day retention diverges by 15+ percentage points between channels. The cheaper-CAC channel is the higher-churn channel.

Does not apply: Apps without cohort retention tracking (build that infrastructure first). Sub-$5K monthly per channel where the cohort size is too small for the gap to be real. Apps where both channels retain similarly and the gap is creative-fatigue rather than audience-quality.

A founder line, anonymized

“TikTok looked like our winner. The 60-day cohort math made it look like our most expensive channel. The audit reframed which number to optimize against. We are still on TikTok because it is the cheapest place to test new creative angles. We just stopped scaling subscribers there.”

For the full B2C SaaS paid diagnostic where 60-day cohort retention by channel is Check 3, see B2C SaaS Paid Strategy: Why Cost-Per-Install Lies. Cohort retention by source is the single most predictive number for subscription unit economics.

Book the 10-Day Allocation Audit.