How is your website ranking on ChatGPT?
Google Discover now lets people follow publishers and preview a dedicated source space. Use this practical playbook to unify entities, standardize cross‑posting, and convert Discover impressions into lasting followers.
Vicky
Sep 27, 2025
@id, name, logo, and sameAs links to X, Instagram, YouTube, LinkedIn, and Wikipedia when available. Mirror that in WebSite and each Author Person entity.\n- Keep handles, bios, avatars, and brand colors consistent so the source space preview looks cohesive.\n- Add VideoObject for self‑hosted clips and align YouTube titles and keywords with article targets so Shorts and stories co‑index around the same entity.\n- Maintain Open Graph and Twitter Card parity on every article to control preview titles, images, and descriptions.\n\nFor deeper entity alignment tactics across assistants, see Model-aware AEO for Copilot.\n\n### 2) Standardize cross‑post canonicals and UTMs\n- Choose one source of truth per idea. The long‑form article is canonical; derivatives reference it.\n- On site, set rel=canonical to the primary article. For recaps or galleries, use rel=canonical to the source or rel=alternate when both should exist.\n- In social posts that link back, standardize UTMs so Discover‑surfaced social clicks are attributable:\n - utm_source=discover\n - utm_medium=social post\n - utm_campaign=publisher space\n - utm_content={platform}-{post id}\n- Add the same UTM to Shorts descriptions and pinned comments; on Instagram, use link stickers or bio links tied to the post.\n\n### 3) Build an LLM micropost factory that matches Discover snippets\n- Output style targets\n - Title style: concrete and entity‑first. Example: Pixel 9 camera test shows big gains in low light.\n - Snippet style: 1 to 2 short sentences, 160 to 220 characters, outcome first, names key entity, and promises one verifiable takeaway.\n - Image rule: high‑contrast 16:9 thumbnail with the subject centered. Avoid heavy text overlays.\n- Prompt pattern per article\n - Summarize the central claim in one crisp sentence a reader can validate in a quick scan.\n - Generate three variants tuned to X, Instagram captions, and Shorts descriptions. Each includes the canonical link with standardized UTMs and one relevant hashtag. No clickbait and no hedging.\n- QA guardrails\n - Ground every micropost in facts from the canonical article and its cited sources.\n - Add a compliance checklist for claims, numbers, and names before scheduling.\n\n### 4) Optimize your publisher space preview in Discover\n- Treat the source space as a storefront. Curate the last 6 to 9 posts around one clear beat with tight visuals and consistent naming.\n- Align Wikipedia and Knowledge Panel data with your Organization schema so the short description is accurate and trust‑building.\n- Add a Follow on Discover callout on site that deep links to your source space when available. Place it in article footers, About pages, and newsletters.\n\n### 5) Instrument Discover impressions to follower growth\n- Primary KPIs\n - Discover impressions and CTR by page in the Search Console Discover report.\n - Clicks from Discover‑surfaced social posts using standardized UTMs.\n - Repeat exposure proxy: percent of Discover sessions returning within 14 days on the same device cohort. Correlate with follow prompting periods.\n- Proxy for follow conversion\n - You cannot directly measure follow taps today. Build a weekly model relating changes in Discover CTR and repeat exposure to windows when you run Follow CTAs and publish LLM micropost sets. Annotate the model with campaign dates and the September 17 rollout. Use step‑change detection to estimate lift.\n- Experiments\n - A/B pages: add a Follow on Discover CTA to half of high‑traffic articles for 14 days while holding cadence and topics constant. Compare repeat exposure and Discover CTR deltas.\n - Timing test: cluster releases so the canonical article and three microposts publish within 90 minutes. Compare to staggered posting on similar topics the next week.\n\n### 6) Content system and governance\n- Ship each major story as a set: canonical article, one Shorts clip, one X post, and one Instagram post. Share the same hero image family and keywords.\n- Maintain a style guide that bans vague curiosity headlines and requires one named entity and one outcome in every title.\n- Keep a human‑in‑the‑loop review for LLM outputs. Require source citations in the editorial worksheet and verify microposts match those sources.\n\n## Quick start checklist\n- Add Organization and Author JSON‑LD with sameAs to X, Instagram, and YouTube. Verify logo and brand fields.\n- Pick three cornerstone beats and prune off‑topic social posts for 30 days to concentrate your preview space.\n- Stand up a micropost generator prompt and a two‑person QA workflow. Ship in batches tied to your publishing calendar.\n- Standardize UTMs and update social link templates and YouTube defaults.\n- Add Follow CTAs on top articles and newsletters. Start a weekly Discover‑to‑follow proxy report with annotations.\n\n## What to watch next\nCoverage indicates the follow feature and cross‑platform posts are rolling out, with users previewing a dedicated source space before following. Expect iterative changes and more platform types over time. Keep your entity graph clean and your templates flexible so your space stays coherent as formats expand.