--- title: "Publisher Audience Interest Graph: One Profile, Every Format" description: Your reader, viewer and listener are one person. An anonymous interest graph builds one first-party profile across article, video and audio. No PII, no login. url: https://preview.artificialpoets.com/use-cases/cross-media-graph/ site: Artificial Poets type: page date: 2026-08-09T19:46:31+00:00 modified: 2026-08-18T03:08:54+00:00 --- # Your Reader, Viewer and Listener Are One Person. **The same engine, on new surfaces** "Why does my podcast app have no idea what this person reads?" Because each format got its own tooling and none of it sees the person: site analytics, video watch history and podcast subscriber list are three records with zero shared signals. The Platform builds one anonymous interest profile per person and orchestrates every surface you own. Not an identity graph, so no PII and no login; an interest graph, built from what someone reads and watches this month. Our measured results are on web, and this is the same engine on a new surface. [See it on my numbers](/request-a-demo/) Horse & Rider · Ride TV · Equine Network · Equus Magazine · The Score · Equine Network Lockup ## The problem, in your numbers Recirculation inside a single site is a discipline. Recirculation across the portfolio — the one referral network you own outright — is unworked ground. - **−43%** Publisher search traffic, forecast over the next three years (Reuters Institute survey of news executives) - **−2/3** Google referral traffic lost by one major portfolio operator by Q1 2026 (AdExchanger) - **38→41%** Internal recirculation share of pageviews — the largest controllable lever left, still worked one site at a time (Chartbeat, via Digital Content Next) - Readers who touch a second title on the multi-title network we measure (our panel) ## Three strangers, or one person On the multi-title network we measure, fewer than 0.2% of readers touch a second title. That is not a failure statistic — it is the size of the asset nothing is currently working. One profile orchestrating every surface is the specific thing Netflix has (~80% of viewing from recommendations, its published figure) and media portfolios don't. - **Today: 3 profiles · 0 shared signals** — Site, video app and podcast app each hold the same person with their own profile — the same person, three times. - **With the graph: one anonymous interest profile** — No PII, no login. Not identity, which lives in a customer file and changes over years, but interest, which moves week to week and is what actually decides the next article, episode or show. ## The evidence, at its honest tier - **Our measured results are on publisher web surfaces.** — The cross-media graph runs the same mechanism — vectorization, fingerprinting, intent matching, orchestration — with no measured deployment on this surface yet. That is the honest tier, and we print it. - **What the engine did on web** — multi-page sessions +46% on enabled titles (10.3% → 15.1%) while comparison titles fell −16% (12.0% → 10.1%) over the same nine months - **Largest single title** — +19.1% in the first quarter after enablement, +4.0% at steady state — both published; plan on the second - **A first cross-media deployment starts the way the web deployments did** — four-week baseline, frozen metrics, unchanged sister properties as the comparison group. The claim gets earned before it gets made The web deployment behind the mechanism, not a cross-media result: [two titles grew multi-page sessions 46% while the comparison group fell 16%](/customers/equine-network-rollout/). **Introducing** ## Artificial Poets Platform One engine behind every solution on this site. It learns your archive and your readers, then acts inside your CMS, your templates and your ad stack. - It learns your archive: **Every story you have published, current again** — The engine understands each piece by what it is about, not when it ran or where it was filed. A feature from 2019 competes for the next slot on merit with one from this morning. - It reads the visit: **What a reader wants, without asking** — Interest builds from what someone actually does in the session. No login, no third-party cookies, nothing leaving your domain. Useful on the second pageview, not the tenth visit. - It chooses: **The right next read, not the popular one** — Someone comparing products and someone following a running story want different things. A most-read list gives both the same five links and serves neither. - It serves: **There before the reader leaves** — The feed, the recommendations, the search answer and the signup ask all run on the same engine, in your templates and your ad stack. Any slot that arrives with them is yours to sell. - It proves: **A lift you can defend, or we say so** — Every deployment runs beside titles that did not get it, plus a serving pause. That is how a result becomes a number you can take to a board instead of a vendor claim. ## FAQ ### If nobody logs in, what actually connects an article read, a video view and a podcast play to one person? On surfaces you own and instrument, a first-party anonymous profile built from behaviour. Where a format is distributed by someone else, particularly podcast feeds, there is no reliable join, and identity resolution by IP is far weaker than vendors imply. We would rather name that boundary than sell across it. ### Doesn't this require unifying our data layer first? No. The profile is built from on-page behaviour by the same JavaScript integration, per property — not by merging your CMSes, CDPs or subscriber databases. Titles that share nothing but an owner can share a graph. ### What share of our audience would ever accumulate enough cross-format behaviour for this to fire? A minority, and that is worth sizing before committing. Publishers running authentication typically see around a fifth of their audience identified, and cross-format behaviour is rarer still. The measured figure we can offer is how thin the connection is today: fewer than 0.2% of readers on the network we measure touch a second title. ### Who owns the graph, and where does the data live? The profiles are anonymous interest fingerprints — no PII, no login required, no pooling across customers, GDPR/CCPA-compatible. Even the six UK magazine groups that pooled audience data for a joint marketplace led with publisher control of that data (Press Gazette). Same answer here: your properties, your graph. ### Someone reads three pieces on a topic. What says they want the documentary rather than a fourth article? Nothing yet, and that is the honest state of this surface. Format preference is durable in the research, not just topic preference, so cross-format recommendation can be a worse recommendation. That is exactly the thing a first deployment has to measure rather than assume. ### Won't cross-title clicks just reshuffle sessions we'd have gotten anyway? At under 0.2% cross-title readership, the baseline is close to zero — there is very little to reshuffle. But that is an argument, not evidence, so a deployment is designed to produce the evidence: enabled properties measured against unchanged sister properties, the same comparison design that put +46% next to −16% on web. If cross-property discovery only moves sessions around, the measurement will say so. ### Is this one identity across formats, or topic tags with a nicer name? The fair answer is that it is one anonymous behavioural profile where we can observe behaviour, and subject-level affinity where we cannot. Calling that a single identity graph across every format would overstate it. Our measured results are on publisher web surfaces, and the cross-format extension carries no measured claim yet. ## Related - [One Audience. Every Property.](https://preview.artificialpoets.com/solutions/media-networks/) — "We own a dozen properties — why does our audience behave like it belongs to strangers?" - [Get Them to the Second Episode](https://preview.artificialpoets.com/use-cases/video-recommendations/) — "Subscribers churn before a second episode — how do I get them watching in the first session?" - [Every Episode Findable](https://preview.artificialpoets.com/use-cases/podcast-discovery/) — "We make great shows nobody finds — how does a listener discover the next episode, or the next show?" - [You Can’t Get the Traffic Back. You Can Get the Session Back.](https://preview.artificialpoets.com/resources/traffic-back/) — The economics of session depth, and the causal evidence behind it: +53% pageviews per user, 95% CI [+48%, +58%], with the modelling assumptions written down. ## See how this works on your titles Thirty minutes on your analytics. We tell you what share of your sessions stop at the first page, and what this would realistically move first for a setup like yours. You leave with the annotated read. No deck. [See it on my numbers](/request-a-demo/) How these numbers are made: [the measurement method](/solutions/measurement/). Figures from a multi-title publisher network measured continuously on publisher web surfaces, August 2025 to May 2026. Last updated August 15, 2026. ## Questions this page answers ### What is a cross-media audience graph? It is one anonymous interest profile per person that spans the formats you publish: articles, video and audio. Today those are three separate records with no shared signals. On the multi-title network we measure, fewer than 0.2% of readers ever touch a second title, which is the measured size of the unconnected asset. ### How does one profile span formats without a login? The profile is anonymous and first-party, built from behaviour on properties you own, with no PII and no third-party cookies. Stated plainly, our measured results are on publisher web surfaces; carrying the same profile across video and audio is the same engine on a new surface, and we print that distinction. ### How much does it cost? There is no public price list; pricing follows portfolio size and the formats involved. Properties come online one at a time rather than all at once, so cost tracks the rollout, and the ones not yet enabled serve as the comparison group. ### Who is it for? Media operators publishing in more than one format, where the reader, the viewer and the listener are demonstrably the same audience but every system treats them as strangers. It is also the page for the operator whose valuation depends on owned audience being real rather than claimed. ### What are the alternatives? A CDP unifies records you already collect, which for anonymous audiences is most of nobody. Identity graphs resolve logged-in users, a minority of publisher traffic. Building it in-house is what the largest portfolios did with engineering organizations you do not have. This builds its own graph from on-site behaviour. ### How do I get started? Start on the surface that is measured. Web properties deploy first, on the usual four-week baseline and named enablement day, and the graph extends to video and audio from there, with the same measurement design applied to each new surface before any claim is made about it.