Multi-brand companies: keeping entities separate

Parent and product brands need distinct names, URLs, and schema. How to avoid one blob that confuses every assistant.

Holding companies love umbrella messaging. Answer engines do not. If "Acme Global," "Acme Payroll," and "Acme Shop" collapse into one blurry Organization in schema, titles, and about copy, assistants will pick a name at random — or skip you for a sharper competitor.

Multi-brand portfolios need separation on purpose: distinct names, URLs, and schema, with parent links that clarify ownership without merging identities.

What goes wrong when brands blob together

This is the multi-brand version of one brand, one entity. Parent and child can be related. They should not be identical.

Decision table: one site or many?

How to separate entities cleanly

1. Inventory legal and go-to-market names. List parent, product brands, and retired names. Pick the public string for each.
2. Assign primary URLs. Each brand that should be citeable gets a primary HTTPS home (site or section) and a matching schema url. Enforce one canonical host per brand site.
3. Write distinct about copy. One paragraph that only makes sense for that brand — not a shared "we are innovators" blob. Machines need about pages they can read.
4. Emit the right types. Parent: Organization. Product line sold as software: often Organization plus Product or SoftwareApplication as appropriate. Founder personal brands: Person — see personal brand vs company schema.
5. Link, do not merge. Use parentOrganization, subOrganization, or brand to relate records. Keep separate @id values.
6. Split sameAs sets. Child brand sameAs should point at that brand's LinkedIn, Crunchbase, Wikidata — not only the parent's. Guide: sameAs profile links.
7. Align titles and H1s. Child site title tags should lead with the child name, not only the parent conglomerate.

Minimal parent/child sketch

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "@id": "https://www.acme-payroll.example/#organization",
  "name": "Acme Payroll",
  "url": "https://www.acme-payroll.example/",
  "parentOrganization": {
    "@type": "Organization",
    "name": "Acme Global",
    "url": "https://www.acme-global.example/"
  }
}

Do not paste the parent's full sameAs array onto the child and call it done.

Step checklist for a portfolio cleanup

1. Crawl titles, H1s, and Organization names across properties.
2. Flag collisions (same name or @id on different brands).
3. Fix the highest-traffic child brand first.
4. Update directories and review profiles so name strings match the child where the child is listed (directories that still matter).
5. Confirm AI crawlers can fetch each child homepage — allowlist if the WAF interferes.
6. Re-scan each child site after fixes; treat scores as website checklists, not citation forecasts.

How to verify separation stuck

After you ship schema and title changes, do a cold check that does not depend on your CMS preview:

1. Fetch each child homepage HTML and confirm the Organization name string is the child, not the parent.
2. Open each child LinkedIn and Crunchbase profile and confirm the display name and website field match that child.
3. Ask a few category prompts in an assistant that shows citations (say you run 10 prompts that should name the child product). Note which brand string appears — parent, child, or neither.
4. If the parent still wins every time, look for leftover parent Organization injection via a shared tag manager, not for a new "AEO trick."

A BrandKnown scan on each property (~60-second website checklist) will catch access and identity mismatches per site. It will not tell you the portfolio is "winning" AI search.

Content and PR habits that keep the blob alive

Press releases that only mention the parent, guest posts that backlink to the conglomerate homepage, and review sites claimed under the wrong product name all drag entities back together. When Northstar Analytics (fiction) launches "Northstar Pulse," the Pulse site needs Pulse-named titles, Pulse schema, and Pulse review profiles — with a clear parent link — or assistants will keep saying "Northstar Analytics" for Pulse-specific questions.

The same trap shows up in sales decks and partner directories: if every listing uses the legal parent name while buyers search for the product brand, you trained the public graph against yourself. Fix the public strings first; legal entity names can live in the footer and filings.

Mistakes

• One global tag manager injecting parent Organization schema on every child domain
• Brand guidelines that force the parent name into every H1
• Merging Wikidata items because someone shares a legal entity
• Using the same Crunchbase page for three product brands
• Assuming a shared design system means a shared entity
• Blocking bots on the child docs subdomain while allowing them on the parent marketing site

Honest ceiling

Clear separation helps assistants resolve which name to say. It does not buy you citations. Similarweb Gen AI Landscape 2025 (US desktop, Sep 2025) reported ChatGPT referrals averaging about 15 minutes on site vs about 8 for Google, with roughly 12 pages/session vs 9 — and about 7% conversion on transactional sites vs about 5% for Google. Useful if you earn the referral; irrelevant if the model names the wrong sister brand.

BrandKnown scans one site at a time against website checks (Agency plan: $49/mo, 10 saved brands, 100 scans/month). Use that to keep each brand's access and identity checklist honest — not to claim a portfolio-wide ranking forecast.