Contents
The entity problemWhat drift looks likeTwo names per brandThe auditThe fixRelated service
SEO Hong Kong →The entity problem specific to established Hong Kong businesses
Many established Hong Kong businesses have accumulated conflicting entity signals over their years of operation. Domains have changed, often between .com.hk and .com. Premises have moved, with the old addresses persisting in directory listings. A series of agency partners have each built links and citations to slightly different variations of the brand name, all in good faith and all slightly different.
The result is that search engines and AI models hold inconsistent signals about what the business is, where it is and what it is called. A newer competitor with a clean, consistent entity footprint from a single domain will often outrank an established business whose underlying authority is stronger but whose signals are incoherent.
What twenty years of drift looks like
Written out for an illustrative established retailer, the conflicting records machines are trying to reconcile look like this, and every row is a real pattern from audits.
| Where the signal lives | What it says |
|---|---|
| The current website | Brand name in English, .com domain, current Central address |
| Google Business Profile | Same name with "HK" appended, the previous shop's address in the photos |
| OpenRice and local directories | The Chinese name, the .com.hk domain from 2015, an old phone number |
| Legacy press coverage | An earlier trading name, links to the retired domain |
| A past agency's citations | The English name spelled with an ampersand the brand dropped years ago |
No single row is damaging, and together they describe five almost-businesses. A machine resolving the entity has to guess which facts belong together, guessing costs confidence, and confidence is what rankings and AI recommendations are made of. The newer competitor wins because its one short history contains no contradictions.
The Hong Kong twist: two names per brand
Hong Kong adds a dimension most entity guides never mention, because an established brand here usually has two names, English and Chinese, and the machines meet both. The Chinese name lives in OpenRice listings, local press and customer reviews while the English name runs the website, and unless something connects them explicitly, the signals split across what reads to a machine as two weakly documented businesses instead of one well-documented one.
The connection is made deliberately: the Chinese name stated as alternateName in the Organization schema, both names present on the About page in running text, and the major listings carrying both consistently. For brands whose Chinese-language AI visibility matters, this is the single most valuable entity fix, since a Chinese-language model that cannot connect the Chinese name to the English domain cannot recommend the store behind it.
The audit process
Entity alignment starts with a full citation audit, covering every mention of the brand name, address, phone number and URL across directory listings, social profiles, press coverage and third-party sites, in both languages. Inconsistencies are catalogued and prioritised by the authority of the source carrying them.
The corrections that matter most are typically in Google Business Profile, where Google's own guidelines require the name and address to match the real-world business consistently, in the major local directories (OpenRice, Time Out Hong Kong, Yahoo Hong Kong), and in any press coverage still using an outdated domain or address.
The second layer is structured data on the brand's own domain. Organisation schema with a complete sameAs array, pointing at the brand's confirmed profiles and listings, gives search engines a single authoritative reference point that the conflicting older signals resolve against.
The fix, in order
- Decide the canonical facts once: legal name, trading name in both languages, one domain, current address and phone, and write them down as the register every correction refers to.
- Fix Google Business Profile first, because it is the entity source Google trusts most and the fastest to correct.
- Work down the directory list by authority, correcting names, addresses and URLs, and retiring listings for premises the business left years ago.
- Publish the complete Organization schema with sameAs and alternateName, so every remaining stray signal has an authoritative record to resolve against.
- Request corrections on high-authority press that carries the old domain or address, accepting that some will never update and letting the weight of corrected signals outvote them.
Measure the result where entities show themselves: the knowledge panel presenting the right facts, branded search resolving cleanly to one business, and AI assistants describing the brand accurately in both languages. Those readings improve over weeks as the corrected signals get recrawled, and the register from step one keeps every future agency and listing on the same facts, which is what stops the drift from starting again.
Sources & references
- Google Business Profile Help, Guidelines for representing your business on Google(support.google.com)
- Schema.org, Organization(schema.org)
- Google Search Central, Organization structured data(developers.google.com)
