Skip to content
Free 20-prompt read
MenuClose
AI Search + Email · D2C skincare brand · $3.7M run-rate · DTC · 12 months

Monthly revenue doubles once the assistant starts naming the brand

Position one on Google for its hero term and still absent from every assistant conversation. Fixing that, then catching the traffic with email, models monthly revenue from $310K to $680K.

Results modelled over 12 months for a D2C skincare brand.
MeasureBeforeAfter
Monthly revenue$310K$680K
Added cash flow per month—+$370K
Prompts citing the brand, of 2404%61%
Email share of revenue11%29%
Blended acquisition cost$61$37

The problem

Paid traffic was getting more expensive every quarter, and customers were arriving already sold — saying “ChatGPT suggested you” — but only 4% of the time. Asked directly, the assistants named the brand in 9 of 240 tracked buyer prompts. Two legacy competitors appeared in nearly all of them, not because their sites were better, but because dermatology blogs, forum threads and comparison roundups named them constantly and named this brand almost never. The email list of 61,000 was getting one broadcast a week and contributing 11% of revenue.

What the model found

Claude is run over 14 months of support tickets, reviews and live-chat transcripts to cluster what customers actually asked — producing the 240-prompt set and, separately, revealing that three quarters of repeat-purchase language clustered around one question the email program never answered. That finding drives the regimen-builder sequence.

The plays

Cluster the customer data with a model

14 months of tickets, reviews and chat logs grouped by an LLM into 240 buyer questions, each weighted by the revenue of the customers who asked it. Prioritisation follows revenue, not search volume.

Rewrite 34 pages to be extractable

Direct answer in the first 80 words, a defined-terms block, a strength-comparison table with real percentages, and an explicit “who this is not for” section. Retrievers lift whole chunks, so the chunks have to stand alone.

Fix the entity graph

Product schema with real ingredient concentrations, Organization schema with consistent founding claims, and matching category language on Wikidata, Crunchbase and every retailer listing.

Earn corroboration where models look

Placement in 21 comparison roundups and review platforms, plus 6 practitioner-authored guides — mentions that put the brand name next to the category on pages ranking well outside the top ten.

Build the email engine on top

AI-referred visitors convert far better than paid but bounce on first visit. A regimen-builder welcome flow, replenishment timing per SKU and behaviour-based segmentation take email from 11% to 29% of revenue — built on self-hosted Listmonk with the segmentation done in SQL against the order table, because the timing model needs per-customer order intervals a hosted tool will not expose.

The lever

The comparison tables. Assistants answering “which is better for X” need something to compare, and a page that lays out the trade-off honestly — including where a competitor wins — gets retrieved far more than a page that only claims to be best. Second lever, quieter: the replenishment timing only worked because the order history sat in a Postgres table we could query directly — that model is not expressible in a hosted email tool’s segment builder.

How the model was calibrated

Calibrated on three published findings. Semrush puts ~90% of ChatGPT search citations on pages ranking 21 or lower and AI-search conversion at 4.4× ordinary organic. Its separate AI Overviews study finds AI Overview links overlap the organic top 10 only 20–26% of the time — the same gap measured a second way, on a different engine. Litmus puts retail and e-commerce email ROI at 45:1, which is why the last play is worth more than it looks.

Sources

  • Semrush — AI Search & SEO Traffic Study — ~90% of ChatGPT search citations point to pages ranking 21+; AI search visitors convert at 4.4× ordinary organic; 50% of ChatGPT links go to business and service sites.
  • Semrush — AI Overviews Study — 200,000 keywords. AI Overview links overlap the organic top 10 only 20–26% of the time; an average AI Overview carries 11 links and includes the #1 organic result just 46% of the time on desktop, 34% on mobile.
  • Litmus — State of Email / ROI of Email Marketing — Return per $1 spent by sector: retail, e-commerce and consumer goods 45:1; agencies 42:1; software and technology 36:1.

Want the same read on your category?

Send a brief and we will run twenty prompts before we talk.

Modelled: revenue doubles once AI names the brand | syed.ai