Case Study: How a Clean Beauty Brand Tripled AI Overview Citations and Grew Organic Revenue 218% — OnyxRank

**Industry:** Clean Beauty / Skincare (DTC Ecommerce)

**Plan:** [OnyxRank](https://onyxrank.com) Pro

**Timeline:** 5 months

**Key Result:** Organic revenue up **218%**; AI Overview and ChatGPT citations for ingredient and routine queries up **3.1x**; 61 product and ingredient pages moved from page 3+ to page 1

---

The Challenge

Wildroot Skin Co. (name anonymized at the client's request) is a clean beauty brand selling fermented-ingredient skincare — serums, moisturizers, and spot treatments built around ingredients like bakuchiol, niacinamide, and centella asiatica. They'd built a loyal following on Instagram and through influencer seeding, but paid social was getting more expensive every quarter, and their organic search presence hadn't kept pace with their catalog.

The core problem: **their best pages were invisible where people actually search for skincare answers now.** Someone asking ChatGPT "what's better for acne scars, niacinamide or vitamin C" or getting a Google AI Overview on "is bakuchiol safe for sensitive skin" never saw Wildroot mentioned — even though the brand had genuinely useful ingredient-education content buried on their site.

A pre-engagement audit turned up the real reasons:

• **No structured ingredient data.** Product pages listed ingredients as unstructured paragraph text with no schema markup, so search engines and AI crawlers had no clean way to extract "what's in this and what does it do."
• **Thin, duplicate-feeling PDPs.** 140+ product pages shared near-identical boilerplate descriptions across shades and sizes, which diluted relevance for any single query.
• **Zero E-E-A-T signals for efficacy claims.** No cited studies, no dermatologist review process, no author bios — a real liability in a category where AI systems are increasingly cautious about unverified health and safety claims.
• **Comparison content gap.** High-intent queries like "bakuchiol vs retinol" and "best serum for melasma" had no dedicated pages at all, so Wildroot ceded that traffic entirely to retailers and beauty media sites.
• **Core Web Vitals failures on mobile**, where roughly 80% of their traffic originated — largely from unoptimized product photography and a bloated app-stack of Shopify plugins.

The Approach

OnyxRank ran a five-month program built around three pillars: technical foundation, EEAT-driven content, and GEO optimization for AI-cited answers.

**1. Technical cleanup first.** Before touching content, we resolved the Core Web Vitals issues — compressing and lazy-loading product imagery, auditing the plugin stack to cut four unused apps, and consolidating redundant scripts. Mobile LCP dropped from 4.1s to 1.9s within three weeks.

**2. Ingredient schema and structured data.** Every product page got structured Product and Ingredient markup, mapping each active ingredient to its function, concentration where disclosed, and skin-type suitability. This gave both Google and AI crawlers a clean, extractable data layer instead of marketing prose.

**3. A dermatology-reviewed content layer.** OnyxRank built out 30+ ingredient-education and comparison pages — "Bakuchiol vs. Retinol," "Does Niacinamide Help With Acne Scars," "Is Centella Asiatica Safe During Pregnancy" — each cited to peer-reviewed studies and passed through a licensed dermatologist for review, with reviewer credentials displayed on-page. This is the single highest-leverage EEAT move for any brand making skin health claims.

**4. Rewriting the 140 near-duplicate PDPs** into genuinely unique copy anchored to that specific product's formulation and use case, rather than shared template language — closing the internal cannibalization that was suppressing rankings across the whole catalog.

**5. GEO-specific formatting.** Content was restructured with direct-answer summaries at the top of each page, clear ingredient tables, and explicit citations — the format AI Overviews and chat assistants preferentially pull from and link to.

Reporting was structured around two dashboards the client checked weekly: organic revenue and AI citation tracking, so the team could see GEO investment translating into pipeline, not just vanity impressions.

The Results

Over five months:

• **Organic revenue up 218%**, cutting paid social's share of total revenue from 71% to 46%
• **AI Overview and ChatGPT citations up 3.1x** for tracked ingredient and comparison queries, with Wildroot now appearing in answers for 34 of their 50 priority terms (up from 11)
• **61 product and ingredient pages** moved from page 3+ into the top 10, including three comparison pages now ranking #1-3 for their target terms
• **Mobile Core Web Vitals** passing on 96% of templates, up from 22%
• **Average order value from organic traffic rose 14%**, since visitors arriving via ingredient-education content were pre-sold on the "why" before landing on a product page

Key Takeaways

• **Structured data isn't optional for ingredient-heavy ecommerce.** If AI systems can't parse what's actually in your product, they can't cite you — no matter how good your copy reads to a human.
• **EEAT is a conversion lever, not just a ranking one.** Dermatologist-reviewed content with visible credentials builds the trust that turns an information-seeking visitor into a buyer, especially for health-adjacent claims.
• **Duplicate PDP copy quietly caps your entire catalog's ceiling.** Fixing 140 pages of near-identical boilerplate unlocked ranking headroom that no amount of link building would have touched.
• **GEO and traditional SEO reinforce each other.** The same direct-answer, well-cited content that wins AI Overview citations also converts better on-page — it's one investment, not two competing budgets.

---

Selling a product where ingredients and trust matter — skincare, supplements, or anything else people research before buying? [Get your free audit](/free-audit) to see exactly where your technical foundation and EEAT signals are leaving AI citations and organic revenue on the table, or [see our plans](/pricing) to find the right program for your catalog size.