Identify the business role before commissioning a feed
A directory and a direct service provider should not start with the same Google integration brief. A directory helps people compare multiple businesses. A direct provider sells its own service. Establish that distinction before buying a feed integration or asking a developer to add markup.
On 18 September, Google updated its aggregator and supplier documentation to include local-business queries. These features concern users in the European Economic Area. A Portuguese address alone does not establish eligibility for every search feature or query.
Use the feature's own participation requirements
Google describes the aggregator unit as a feature for eligible vertical search services, including directories. The documentation sets out an expression-of-interest process, relevant content, data provision, and quality requirements. It links local-business queries to the Local Point of Interest integration documentation.
For the supplier unit, Google identifies direct providers and says no additional data beyond accessible web content is required, although feeds may enhance results. The supplier unit appears alongside an aggregator unit. These are specific feature rules, not a universal instruction to build a feed for every local business.
Give the data a stable source
For a directory considering participation, inventory the records it actually controls. Each listing needs a stable identifier and a documented source. Decide how the team distinguishes a business, a branch, and a service offered at that branch. A name change should not create an unrelated duplicate if it is still the same entity.
The following is an internal quality checklist, not a substitute for Google's current feed specification. Use the official schema for required fields and formats. Keep your own ownership, approval, and correction records alongside it so the integration remains maintainable.
| Internal check | Failure to prevent |
|---|---|
| Stable business and branch identifiers | Duplicates or accidental merging of separate locations |
| Source and verification date | Old information presented as recently checked |
| Landing-page agreement | The feed and public listing describing different offers |
| Closure and correction status | Removed businesses returning in the next export |
Separate formatting checks from business truth
A validator can detect missing fields and invalid types. It cannot establish that a business is still open or that a photograph belongs to the correct branch. Build separate review steps for data shape and factual accuracy. Quarantine uncertain records rather than filling gaps with generated guesses.
AI can help flag suspicious duplicates or inconsistent descriptions. A proposed match should include its evidence and confidence, then go to a reviewer when the consequences matter. Do not let an automatic similarity score merge two businesses simply because their names look alike.
Design the correction loop before the first submission
In a fictional directory of repair services, an owner reports that a branch closed. The team updates its source record, removes the public listing where appropriate, records the effective date, and ensures the next export reflects the change. A stale spreadsheet should not reintroduce that branch during a later batch.
Use a versioned export and retain submission results. Distinguish a rejected file, a rejected record, and a valid submission with no observed search appearance. Those outcomes require different actions. Monitor meaningful failures and route them to the person who can correct the source.
Scope the smallest worthwhile integration
Start with the eligibility map, then decide whether the directory has the data quality and maintenance capacity to proceed. An integration that nobody owns after launch creates a recurring problem.
JQ's workflow integration service can map the source data, validation, and human correction steps before implementation. The useful deliverable is a dependable information process, with search appearance treated as an outcome to observe.
Link Map
Sources and editorial review
Sources reviewed on 21 September 2026. The practical workflows and illustrative examples are JQ AI SYSTEMS analysis unless explicitly attributed.
- Google Search Central: Aggregator unit (Updated 18 September 2026).
- Google Search Central: Supplier unit (Updated 18 September 2026).
- Google Actions Center: Local VSS integration overview (Standing documentation, checked 21 September 2026).