Skincare Supplement Brand (8-Figure Scale)

Setup

A direct-to-consumer skincare supplement brand with a validated VSL funnel came to SkyHouse with a working offer and a spend ceiling no one could break through. The product converted. The funnel worked. The creative wasn't finding the angle that could carry real volume.

Summary

In January 2026, the brand was spending $456K/month at 0.92x ROAS — operating at a loss on meaningful volume. SkyHouse was brought in to find the creative angle that could carry 3x the spend while improving, not degrading, ROAS. Within 60 days, one winning angle tripled spend to $1.31M/month and pushed ROAS to 1.53x. The engagement scaled to $2.36M/month by August 2026.

Challenge

• Below-Breakeven at Scale: $456K/month spent in January at 0.92x ROAS — significant losses at a volume too large to ignore.
• Creative Ceiling: Control campaigns running since late 2025 had exhausted their audience learning. CPMs rising; click-through efficiency falling.
• No Proven Breakout in Queue: Multiple angles tested without any producing the decisive signal needed to justify aggressive scale spend.
• Speed-to-Scale Gap: The process for moving a validated test into a scale campaign was slower than the market window.

Approach

Three-phase velocity model:

• ABO Test: New angle launched at controlled spend with Ad Budget Optimization — 3 headlines × 3 body copy variants, one creative per ad set.
• Gate evaluation: 3–5 days of data. Kill if the signal is noise; advance if CPA < $93 target with meaningful volume.
• Advantage+ Scale: Cleared tests move immediately to CBO Advantage+ at scale spend. No intermediate hold period.

Five net-new angle tests between Jan 9 and Feb 20. The breakout: Ryan's "Top 5 Face Video" (Feb 20) — a native-feel video ad that launched directly into Advantage+ CBO, skipping the test phase and compressing the learning phase significantly. By March 1, this single campaign was spending $29,717/day.

Lander architecture was deliberately held constant while creative angles were varied — a clean single-variable test design. The "Compare / La Mer" angle tested luxury comparison positioning alongside the Face Video test. Two buyers (Calvin and Ryan) ran separate campaign threads on the same offer simultaneously — parallel testing, not sequential.

What We Learned

Creative angle is the primary lever when the funnel is already working. But angle velocity only matters if you have a structure that reads signal cleanly — one ad per ad set, named creative concepts, version-coded buyers. The Feb 20 Face Video didn't win because we got lucky. It won because the account was built to tell us exactly which variable moved the needle, and we acted on that signal within days.