Source
urlhttps://read.readwise.io/read/01kvxra7fjbh0t75yx52wjzpdd
rawraw/i-spent-2000-on-paid-ads-for-my-saas.md

TL;DR: Perfect Interview.ai spent roughly $2.6K on Meta ads and got decent top-of-funnel metrics, but only ~0.47 purchase ROAS. The real lesson is not “Meta ads do not work” — it is that paid-ads-scaling only works when the product has enough LTV and low enough churn to absorb the creative-testing burn.

Biggest lessons

  • Paid ads are a testing tax before they are a money printer. The team treated Meta as a creative-testing lab: test multiple ad creatives, isolate winners, then scale them through CBO campaigns and broader audience pockets.
  • Creative and product economics beat dashboard hacks. The co-founder concludes modern Meta broad targeting is good enough that campaign strategy mostly makes marginal improvements; the product, offer, and creative decide the outcome.
  • The numbers were close enough to learn, not close enough to scale. They saw ~200K impressions, ~$11 CPM, ~2% CTR, ~58¢ CPC, ~$39 average purchase value, and ~0.47 purchase ROAS — useful signal, but still negative.
  • Subscription LTV matters more than first purchase ROAS. In theory, a $30 checkout could become profitable after renewals; in practice, a job-search product has success-triggered churn because the best user outcome is getting a job and leaving.
  • Bad paid economics pushed strategy upstream. Instead of continuing to buy traffic, the team paused paid ads and shifted toward a B2B white-label career-coach angle where monthly contracts could support higher CAC.

Why it matters

  • This is a grounded counterweight to the “just scale ads” story in paid-ads-scaling. Paid can work, but not before LTV, churn, and funnel fit are proven.
  • It connects app-market-research to mobile-app-monetization: a good ad creative is not enough if the product cannot retain or price users high enough.
  • It explains why idea-validation should include monetization shape, not just whether users click.

What links here