Inside ninety days, a $12M skincare brand went from 0/200 to 78/200.

Anchor case study — placeholder figures pending real client approval per spec §H5.

Situation

A $12M-revenue DTC skincare brand. Profitable, modest growth, two-quarter Meta-returns flat, branded search beginning to slip. The CMO had read three AI-search newsletters and pinned them to a Notion doc no one opened. The CFO was beginning to ask why CAC was up 47% year over year.

What we ran

The standard four-stage system — Audit → Instrument → Optimize → Compound. The audit landed on Tuesday. The phone rang Friday.

Stage 01 — Audit

200-prompt sweep across ChatGPT, Perplexity, Gemini, Claude. The brand was cited zero times. Three competitors were cited 22, 18, and 12 times respectively. The audit deck included the prompts the brand was conspicuously absent from, the SKUs that were vulnerable, and the SKUs where they had natural momentum no one was harvesting.

Stage 02 — Instrument

Server-side detection of AI traffic deployed in week two. Citation tracking pipeline live in week three. By week four, the team could answer “what did AI search drive last week” with a number.

Stage 03 — Optimize

JSON-LD schema rewrite shipped first. Then ACP enrolment for ChatGPT Instant Checkout. Then a PDP-by-PDP rewrite focused on the 12 SKUs the audit flagged as both vulnerable and high-margin. Citation seeding on three review platforms. Crawler permissions audit. Six things, sixty days, no blog posts.

Stage 04 — Compound

The brand crossed competitor C in week 6, competitor B in week 8, competitor A in week 11. Month four: $94K in AI-attributed revenue, +22% AOV vs. paid social.

Results

• 0 → 78 / 200 citations in the audit refresh
• $0 → $94K AI-attributed revenue in month four
• +22% AOV from AI-referred orders vs. paid social
• AI search became the brand’s second-best acquisition channel by retainer’s end

The full case study — engagement structure, week-by-week deliverables, financial model — is available on request to qualified prospects. The audit reproduces the diagnostic on your category in twenty minutes.