Business listings + off-site authority

Search engines and AI answer engines do not take your word for your own authority. A meaningful part of it is read off other pages: a business directory that lists you, a partner site that links to you, a person who cites you in their own bio — a page you do not control, saying you exist and are real. Nothing on your own site can substitute for that, however well written, though it is one signal among several, not the only one.

This skill claims the handful of external listings that matter most for a small site, and checks that the profiles your schema already points to actually resolve. It is the action step for the gap website-content-guide's E-E-A-T section names: knowing Authoritativeness is thin is only useful if it leads somewhere. This is where it leads.

What this skill is not. It does not do cold outreach to strangers — no guest posts pitched to sites with no relationship, no backlinks bought or begged from people who have never heard of the entity. That is relationship work and editorial judgment, not a checklist, and a skill that pretended otherwise would overclaim. It claims listings you are entitled to claim, verifies links you have already earned, and — in section 4 — prompts you to ask people who already have a real connection to the entity (a partner, a sponsor, a speaker) for the link that naturally follows from that connection. Nothing here pays for a link or approaches anyone with no prior relationship.

Run this only if the site represents a real, named, claimable entity — an event series, a studio, a local business, a sole proprietor or consultant trading under their own name, or a named community or organization. Skip it for a personal blog with no business or practice behind it, or a resource hub with no identity of its own to list. If unsure, ask the owner: "does this have a name someone could search for on Google Maps or a directory?" A no means skip.

Human-in-the-loop: creating accounts on Google/Microsoft/directory platforms, and any identity verification they require (a postcard, a phone call, a video call), are actions only the owner can do. The agent drafts every field value it has real data for — the name (§1 step 3), category, description, hours, and the sameAs URLs. Address and phone number are not in config.ts; get those directly from the owner rather than guessing, and the owner pastes and clicks throughout.

Console UI labels are localized — translate the quoted labels if the owner's account language is not English (German: Unternehmensprofil erstellen, Standort hinzufügen, Kategorie, Verifizierung).

1. Google Business Profile — the biggest single lever

This is the one that feeds Google Maps, the local pack, and the knowledge panel that can appear beside search results for the entity's own name.

0. 🧑 Check eligibility before starting. Google Business Profile requires a genuine customer-facing interaction — either a real address customers visit, or staff who travel to perform a service at a customer's location. "Named entity" is not itself sufficient: a purely online community, or an event series with no physical venue, may not qualify for any category, and Google can reject or suspend a profile that does not meet this bar. If the entity is genuinely location-based (a studio, an office, a recurring venue), continue; if it is online-only, skip to section 3 and rely on sameAs and directories instead.
1. 🧑 Search for the business first — on Google Maps or in Business Profile's own "Add your business" flow, search the exact name and address before creating anything new. A listing may already exist (created by a customer review, a data aggregator, or a past employee) — claim or request access to that one rather than creating a duplicate, which risks a suspended profile and split reviews.
2. 🧑 If nothing exists to claim, go to https://business.google.com and sign in with the account that should own the listing — an account the organization itself controls (not one person's personal login shared around; Google Business Profile supports adding named managers/owners individually, which is the right way to give a team access).
3. 🧑 Add your business → enter the real, customer-facing name — the same one in SITE.name, which the starter's Organization schema carries as alternateName (schema's name field holds COMPANY.legalName, the legal entity name, which is deliberately not always the same string — use the public-facing one here, not the legal one, since that is what customers actually search for).
4. 🤖 Draft the category, description, website URL, and hours from the site's own content — do not invent anything not already stated on the site.
   • Category: pick the closest official Google category that matches how the entity actually operates. "Service area business" applies only when staff genuinely travel to serve customers at their location, with the city/region listed instead of a street address — it is not a default for "has no office."
   • Description: reuse the positioning statement from POSITIONING.md if the site has one, trimmed to Google's limit. Do not write new marketing copy here.
   • Website: the production domain, not a preview URL.
   • Address and phone: not in config.ts — ask the owner directly rather than guessing or leaving them blank.
5. 🧑 Complete verification. Google offers postcard, phone, email, or video verification depending on the category and account history — the options shown are Google's choice, not something to work around.
6. 🧑 Once verified, publish the profile.

2. Bing Places for Business — the Bing-side equivalent

Smaller reach than Google, same mechanism: a claimed, verified listing in the index Copilot's search grounding draws on, not a guaranteed direct feed into any specific Copilot answer.

1. 🧑 Go to https://www.bingplaces.com and sign in with a Microsoft account.
2. 🧑 Check "Import from Google" first — if the Google Business Profile above is already live, this can pull the listing across and skip re-entering everything.
3. 🤖 If importing is not available, draft the same fields as §1 step 4 (category, description, website, hours) from the same source — the site's own content, not invented copy. Address and phone still come directly from the owner.
4. 🧑 Complete whatever verification Bing requires and publish.

3. Verify the sameAs profiles actually resolve

website-content-guide's E-E-A-T section calls for sameAs profiles that actually resolve — Organization/Person schema pointing at LinkedIn, GitHub, a public register, and so on. The schema plumbing itself lives in website-seo-geo and config.ts (SAME_AS); this skill is the check that what it points at is real.

1. 🤖 Read the current SAME_AS array (or equivalent) out of config.ts / Organization schema.
2. 🤖 Fetch each URL — reuse outgoing-link-audit's fetch mechanics (its script form is at skills/new-website/templates/astro/scripts/check_external_links.sh, copied into every scaffolded site) but apply this section's own, stricter status policy below, not that script's: deleting a real authority signal is a worse mistake than flagging a broken outbound link, so this is deliberately more conservative than the script's own classification — do not treat a non-200 as dead on the first try: a 404/410 is unambiguous — the server itself says the resource is gone — so that alone is genuinely dead; a 5xx is the target's own server failing, not proof the profile is gone, so re-checking once is not enough evidence to remove it either — treat 5xx as unverified until a human confirms it, logged out, in a real browser, same as a bot-block; 403/429/999/ timeout is usually a bot-block (LinkedIn and similar platforms routinely return these to automated fetches) and also counts as unverified; a URL that redirects to a different domain (and returns 200 there) is neither dead nor healthy — it may point at a renamed or rebranded profile and needs a human look, the same as outgoing-link-audit's own REBRAND/MOVED case.
3. 🧑 For anything confirmed dead (real 404/410 only): claim the missing profile, fix the URL, or remove the entry. For anything unverified (5xx, bot-block, or redirect-to-different-domain): check it yourself, logged out, in a real browser before touching schema — do not remove a sameAs entry on an automated fetch failure alone, that would delete a healthy profile the toolkit's own fetch just couldn't reach.
4. 🤖 Add the Google Business Profile and Bing Places listings (sections 1 and 2 above, not this section's own steps 1-2) to SAME_AS once they are live and verified — use the public listing URL (the Google Maps place page, the public Bing Places listing) that a visitor or crawler can actually open, never the private dashboard/management URL used to edit the listing.

4. Category-relevant directories — a short list, not an open crawl

Pick from the site's actual category. This is deliberately a short, curated list, not "submit everywhere" — low-quality directory link farms have net-negative value and some actively hurt trust signals. If nothing in the category has a directory worth being listed in, say so and move on — "no suitable directory found" is a valid outcome, not a gap to keep chasing.

• Event / community site: Meetup, Eventbrite, lu.ma, and any local tech-community hub or aggregator that already covers the region.
• Local business or studio: the 2–3 directories that are actually authoritative for that trade in that region (ask the owner which ones their peers use — this is local knowledge, not something to guess).
• Any site: if a partner, sponsor, or speaker already has their own site, ask whether they would link to the event/organization page they took part in — this is the highest-value, lowest-effort link available, because it is editorially justified and already half-earned. website-testimonials is the sibling skill for the testimonial itself; this is the follow-up step of asking for the link that goes with it.

🧑 Submitting to each directory is a human action (most require an account and a human-verified claim); 🤖 draft the listing text from the site's own copy for each one so the owner is pasting, not writing, at every step.

5. Confirm it's working

• Search the entity's own name on Google — a Business Profile panel or Maps pin can appear once Google finishes processing the verified profile; there is no guaranteed timeline, so treat this as a check to repeat, not a one-time deadline.
• Search on Bing — same check for the Bing Places listing.
• Re-run the sameAs check in step 3; every URL is either confirmed live (200) or manually verified in a logged-out browser (for anything the automated fetch left unverified — a bot-block, a 5xx, or a redirect to a different domain), and points at the entity's real profile.
• This toolkit has no dedicated score for Authoritativeness. The honest way to see whether it moved is qualitative: ask an AI assistant to evaluate the site's E-E-A-T again after a few weeks and compare its answer to before — an informal check, repeated, not a packaged metric. This skill claims listings and verifies links; it does not itself measure the outcome.

Scope notes

• This is claiming and verifying what you are entitled to, plus asking people with an existing connection for the link that follows from it (§4) — not cold link building or outreach. Guest posts, press pitches, and cold backlink requests are relationship work and stay a human decision, always.
• internal-link-audit and outgoing-link-audit cover the site's own link graph (links within the site, and links the site makes outward). Neither touches links pointing at the site from elsewhere — that gap is what this skill fills.
• The schema plumbing (sameAs, Organization/Person) is website-seo-geo's contract; website-content-guide decides what signals to ship. This skill is the step that goes and makes those signals true, rather than aspirational.