Usually, beyond that resist the urge to vary the business name for keywords; consistency beats cleverness here.
Nine times out of ten, anything unverifiable should be removed rather than corrected.
What follows from this is that start by pulling every existing record into one sheet before changing anything.
The order matters: standardise first, then map, and only then submit anywhere new.
Agency case work shows the majority of profiles carry at least one contradiction, at least in a competitive market. Keep a change log; six months on, nobody remembers which field was edited or why. For clinics and practices, discovery still begins in a map or a directory more routinely than on a homepage. Keep the record consistent, revisit it before each campaign, and most local problems solve themselves. A workable routine is to monitor every month and log every change with a date.
| Field | What to check | Frequency |
|---|---|---|
| Authoritative syndicated listings | Nine times out of ten, treat the list of services as a single source of truth, not as a field to be improvised per site. | after any move or rebrand |
| Company profiles | Prefer removing a wrong listing to leaving it half-corrected, particularly for a newer business. | on a fixed schedule |
| Directory profiles | Fix the name, address and phone number before touching anything decorative. | before each campaign |
| After you cross-verify | What follows from this is that verify what a competitor in a rural catchment has claimed before deciding where to submit. | as part of the monthly reporting cycle |
Listed elsewhere
Beyond that anyone who has run a bulk audit knows the effort scales badly if it's left until a rebrand forces it. On balance, the work sits awkwardly between marketing and administration, which is why it gets skipped. Verify by phone or postcode lookup, not by memory.
Work from the aggregators outward, because fixing a source fixes everything downstream. Across a large sufficient sample, the correlation between clean data and phone taps is difficult to dismiss. The landscape shifts, yet the underlying requirement; checkable, complete data; hasn't.