I Audited Summer Fridays Against AI Shopping Agents. Sephora's Best-Seller Is Almost Invisible.

Executive Summary

• Brand: Summer Fridays. Influencer-founded skincare brand launched 2018 by Marianna Hewitt and Lauren Gores Ireland. 2M+ products sold by 2022. Stocked at Sephora and Amazon. ~1.1M Instagram followers. Best structured data of any brand audited to date (Product JSON-LD with AggregateRating).
• AI visibility score: 2/15 queries surfaced the brand. Gemini found the Jet Lag Mask on the face mask query (2nd position). Copilot surfaced it on a lip balm query via an eBay shopping tile. ChatGPT returned zero across all 5 queries.
• The pattern: The best technical infrastructure in the audit set — JSON-LD with AggregateRating, Okendo reviews, strong editorial presence, Amazon storefront — and it still barely surfaces. The bottleneck is the body_html descriptions: 1-3 sentences per product, with all the rich content trapped behind JavaScript.
• Key finding: Social media fame does not transfer to AI visibility. 1.1 million Instagram followers, viral TikTok products, influencer founders, and Sephora retail dominance generated zero ChatGPT recommendations. AI shopping agents and social media are different channels driven by entirely different signals.
• Root cause: The body_html descriptions are paper thin. A $68 vitamin C serum is described in one sentence. Clinical study results, ingredient breakdowns, and usage instructions exist on the rendered pages but are loaded via JavaScript, invisible to crawlers and AI agents. The structured data (JSON-LD with AggregateRating) is the best in the audit set but only covers basics — name, price, rating. No ingredient or attribute data in the schema.
• Fix complexity: Low. The technical foundation is already the best in the audit set. The fix is content: move the existing rich page content into the body_html field and extend the JSON-LD with product attributes. The Trustpilot profile (1.8/5, unclaimed) needs attention.

The brand

Summer Fridays was founded in 2018 by Marianna Hewitt and Lauren Gores Ireland in Los Angeles. The Jet Lag Mask became Sephora's best-selling skincare product within two weeks of launch. Over 2 million products sold by 2022. The Lip Butter Balm became a TikTok viral sensation. Multiple Allure Best of Beauty Awards.

Around 1.1 million Instagram followers. Stocked at Sephora and Amazon with an official brand store. Clean, vegan, cruelty-free. Subscribe-and-save model. Recycling programme. The brand expanded from skincare into lip care and hybrid makeup.

I selected Summer Fridays because it has the strongest technical data infrastructure of any brand I have tested. Product JSON-LD with AggregateRating is present on product pages — something most Shopify brands, and most brands in this audit set, completely lack. I wanted to test whether good structured data plus massive brand recognition plus editorial dominance would finally produce strong AI visibility.

Almost. But not quite.

The test

I ran 5 queries across ChatGPT, Gemini, and Copilot. 15 tests total.

The queries:

1. "What's the best face mask for tired, dehydrated skin?"
2. "Can you recommend a tinted lip balm with SPF?"
3. "What's the best overnight mask for dry skin under £50?"
4. "I need a lightweight moisturiser for daily use. Suggestions?"
5. "What's a good vitamin C serum from a clean beauty brand?"

The results

Summer Fridays surfaced on 2 out of 15 tests. But one of those requires an asterisk.

Query 1 (face mask for dehydrated skin): 1 out of 3. Gemini placed the Jet Lag Mask in second position behind Laneige. ChatGPT recommended Dr Jart+, Garnier, No. 7, Hello Sunday, Simple, and L'Oreal. Copilot recommended Sisley, 111SKIN, Bubble, Maskology, Anua, and Eucerin. The product most literally matching this query — a mask designed for tired, dehydrated skin with 2,697 reviews — was found by one platform out of three.

Query 2 (tinted lip balm with SPF): 1 out of 3, with a caveat. Copilot surfaced Summer Fridays in first position, but in a shopping tile from eBay rather than as an organic AI recommendation. This is a product feed result, not the AI choosing to recommend the brand. The Lip Butter Balm (3,825 reviews, 4.7 stars) was found via marketplace data, not via the brand's own site. ChatGPT recommended Kopari, Hello Sunday, Lanolips, Barry M, and NIVEA. Gemini recommended Fresh Sugar, Cay Skin, Sun Bum, and NARS.

Query 3 (overnight mask under £50): 0 out of 3. Laneige appeared on both Gemini and Copilot. The Jet Lag Mask, which is explicitly positioned as an overnight mask and costs $49, was invisible. ChatGPT recommended Dream Drench, Avon, 7days, Revuele, Boots, and Biovene — brands that have almost no editorial presence compared to Summer Fridays.

Query 4 (lightweight moisturiser): 0 out of 3. CeraVe, La Roche-Posay, Neutrogena, and The Ordinary dominated Gemini and Copilot. The pharmacy-tier consensus seen in the Dr Sam's audit holds for generic moisturiser queries. ChatGPT recommended E45, Garnier, Clarins, Simple, and Dove.

Query 5 (vitamin C serum, clean beauty): 0 out of 3. The CC Me Serum with Vitamin C + Niacinamide ($68) was invisible. ChatGPT recommended Antipodes, Green People, RAAIE, and Synthsis Organics. Gemini recommended Ursa Major, Typology, Mad Hippie, True Botanicals, and Cocokind. Copilot recommended Beauty of Joseon and Neutriherbs. The same fragmentation pattern seen in the Herbivore audit — each platform draws from different sources for clean beauty queries.

Why this is happening

Summer Fridays has the best technical infrastructure in the audit set. And still barely surfaces. Three things explain the gap.

1. The body_html is paper thin. The CC Me Serum costs $68 and has a one-sentence Shopify description: "This brightening vitamin C serum helps to improve skin tone and texture, while reducing the appearance of dark spots and discoloration over time." That is it. The rendered product page has clinical study results ("100% saw improvement in skin texture"), ingredient breakdowns, usage instructions, and skin type guidance. All loaded via JavaScript. All invisible to crawlers and AI agents.

The Jet Lag Mask is slightly better — two sentences covering five use cases. The Lip Butter Balm gets three sentences including a purchase limit. But for products at this price point with this level of editorial coverage, the descriptions are critically thin.

2. Social media fame does not transfer. This is the category-level finding. 1.1 million Instagram followers. Founded by influencers. Viral TikTok products. None of this appeared in AI shopping agent recommendations. Not once on ChatGPT. Not once on Copilot.

AI shopping agents do not consume Instagram followers, TikTok views, or influencer authority. They consume structured data, editorial roundup citations, and machine-readable product information. The channels that built Summer Fridays are invisible to the channels that will recommend products in the future.

3. The Trustpilot is a liability. 1.8 out of 5 with 30 reviews. Profile unclaimed. 70% one-star ratings about shipping and customer service. The brand's review strength lives on Sephora and Okendo (where the Lip Butter Balm has 3,825 reviews at 4.7 stars). But an unclaimed 1.8-star Trustpilot is a negative trust signal that AI agents can surface. This mirrors the Taylor Stitch pattern (2.0/5, 23 reviews) from the fashion audits.

The eBay finding

The Copilot result on Query 2 deserves separate attention. Summer Fridays appeared in a shopping tile — but the listing was from eBay, not from summerfridays.com. Copilot's shopping tiles are driven by merchant feeds (Bing Shopping). When a brand's own product data is thin, marketplace listings can fill the gap.

This means:

• The customer clicking that result goes to eBay, not to Summer Fridays
• The sale, the review, and the data all benefit the marketplace, not the brand
• The brand's DTC channel is being cannibalised by marketplace data that is richer than its own

This is what "data ownership" means in AI commerce. If your own site does not provide rich, structured product data, someone else's listing of your product will.

The Laneige benchmark

Laneige appeared on 2 out of 3 mask-related queries (Q1 and Q3) across Gemini and Copilot. Their Water Sleeping Mask and Lip Sleeping Mask have become the AI default for hydrating masks.

Summer Fridays' Jet Lag Mask is a direct competitor. Comparable reviews, comparable editorial coverage, comparable price point. But Laneige wins consistently. This makes Laneige a useful benchmark — what are they doing in their product data layer that Summer Fridays is not?

What Summer Fridays could do, in priority order

Phase 1 (1-2 days):

• Move the existing rich page content into the body_html field. Clinical study results, ingredient breakdowns, usage instructions, and skin type guidance already exist on the rendered pages. They need to be in the Shopify description field where crawlers and AI agents can read them. This is the single highest-impact fix.
• Claim the Trustpilot profile. Respond to complaints. A 1.8-star unclaimed profile is worse than no profile at all.

Phase 2 (3-5 days):

• Extend the JSON-LD with product attributes. The existing Product + AggregateRating schema is a strong foundation. Adding ingredient highlights, skin type, vegan/cruelty-free status, and clinical claims would make Summer Fridays one of the most machine-readable skincare brands.
• Add product differentiation context. Three moisturisers (Jet Lag Mask, Cloud Dew Gel Cream, Rich Cushion Cream) with no guidance on how they differ. Each description should explain who should choose it.

Phase 3 (2-3 weeks):

• Build category-owning content targeting the queries AI agents answer. "Best face masks for dehydrated skin" and "best overnight masks" are queries Summer Fridays should own. First-party content targeting these patterns, with clinical data and product comparisons, gives AI agents authoritative material to draw from.
• Ensure Amazon listings match enriched DTC descriptions. The eBay shopping tile finding shows that marketplace data can cannibalise DTC visibility. If Amazon listings are richer than summerfridays.com product pages, Amazon wins the AI recommendation.

Close

Summer Fridays has more going for it than any brand I have tested. JSON-LD with AggregateRating. Multiple Allure awards. Sephora's best-selling skincare product. 3,825 reviews on the Lip Butter Balm alone. 1.1 million Instagram followers.

It surfaced on 2 out of 15 AI tests. One of those was from eBay.

Laneige appeared more often. CeraVe appeared more often. Brands with a fraction of Summer Fridays' recognition appeared more often.

The gap is not fame, reviews, editorial coverage, or even structured data. The gap is one thing: the Shopify description field. One to three sentences per product, while everything else — the clinical studies, the ingredient breakdowns, the usage instructions — lives behind JavaScript that machines cannot read.

A million followers built this brand. A few paragraphs of product description would make it visible to the agents that will recommend products to those followers.