Key points
- BDG's GEO offering, called ECHO internally, pairs Profound analytics with AI agents trained on its own titles' content; three or more brands have bought it.
- In an Aug. 11 to Sept. 30 campaign, Innisfree's visibility across 200 tracked skincare prompts rose 30%, per a BDG case study using Profound data.
- A Scary Mommy explainer lifted Pampers into the top spot for fragrance-free diaper questions; it had held second place before.
- Attribution is still weak: Profound's Josh Blyska says AI influence does not yet flow cleanly into transactions.
Why it matters for marketers: Publishers are now selling measurable AI-citation work as a paid product, giving brands a concrete model, with caveats, for budgeting around answer-engine visibility.
Bustle Digital Group has built a generative engine optimization product for advertisers, known internally as ECHO. It uses Profound's analytics and AI agents trained on BDG content to produce briefs and guidelines, which the publisher's own creative team then follows when writing. Graham Howes, BDG's svp of distribution and AI intelligence, said rising client inquiries prompted the effort. At least three brands have paid for it, and BDG says it is the only publisher working with Profound on an advertiser offering.
The process starts with an RFP and a baseline research phase of two to four weeks, studying where a brand stands and which sites answer engines currently cite. Advertisers see the strategy but do not shape the creative. Howes said heavy-handed advertorial is unlikely to get cited. Finished pieces are promoted on Reddit through paid media, since Reddit remains a major citation source for answer engines.
BDG's case study reports Innisfree's visibility grew 30% across 200 skincare prompts, while Bustle's citation rank climbed from 709th to 12th. Pampers reached first place for fragrance-free diaper questions after a Scary Mommy explainer. Howes declined to share pricing or revenue.
For marketers, the lesson is that AI visibility can be tracked at the prompt level, but the link to sales is unproven. Blyska said the influence layer does not yet connect cleanly to the transaction layer. Go Fish Digital's Ashish Jacob added that prompts, platforms and sampling methods affect results, so numbers from different tools may not match.
Watch for renewals and longer, always-on programs. Howes said some 2027 renewals are coming, and that this content fades over time, especially in seasonal categories like beauty and fashion. Ask any publisher partner how they measure and what they will report.
Sources
This brief is our own summary. Read the full reporting at the source.
