Case studies

What the work looks like.

Four engagements that represent the patterns we see most. The mechanics are what matter — in almost every case the problem was something already broken rather than something missing.

A note on these. These are representative composites drawn from common engagement patterns, with client names and identifying details removed. We don't publish named client results without permission, and we'd rather show you the mechanics honestly than attach a logo to a number you can't verify. Ask us on a call and we'll walk through specifics.
Multi-locationRestoration12 locations

Locations competing against each other

Situation

A restoration group across twelve markets in the Southeast. Strong revenue in three markets, almost nothing from the other nine, despite identical marketing spend and equivalent crews. The assumption was that the nine were in weaker markets.

What we found

  • Nineteen profiles for twelve locations — seven duplicates created by aggregators over several years, three of them ranking above the real ones
  • Every location configured with the same primary category and overlapping service areas, so in two metros the branches were suppressing each other
  • Four locations still carrying a phone number from a system replaced two years earlier
  • NAP mismatches across the major aggregators — suite numbers present in some records, absent in others

What we did

Resolved the duplicates first — merges where possible, removals where not. Separated service areas so each branch owned a defined catchment with no overlap. Differentiated secondary categories to reflect what each location actually specialised in. Corrected the phone numbers at source rather than listing by listing, then rebuilt citation consistency across the core aggregators.

What changed

The nine underperforming locations began appearing in discovery searches in their own markets within roughly two months, and call volume followed. The two cannibalising branches both improved once they stopped competing for the same searches. No new content, no new spend — the demand was already there and was being split between duplicate profiles.

Single locationDental practiceSuspension

A suspension that stopped the phone overnight

Situation

An established practice lost its profile with no warning. New patient enquiries dropped to almost nothing within days. Two reinstatement requests had already been submitted and rejected before we were involved.

What we found

  • The business name on the profile included a service phrase and a city that weren't part of the registered name
  • A second profile existed at the same address for a practitioner who had left three years earlier
  • The earlier reinstatement requests had been submitted without changing anything, which is why both failed

What we did

Corrected the business name to match the registration exactly. Resolved the departed practitioner's profile. Assembled documentation establishing the practice operates at the address — registration, utility records, signage photographs, professional licensing. Then submitted a single reinstatement describing the business and stating specifically what had been corrected.

What changed

Reinstated, and the profile recovered its previous visibility over the following weeks. The important part isn't the reinstatement — it's that the first two attempts failed because nothing had been fixed before requesting. Repeated submissions don't help and can make recovery harder.

FranchiseFood service40+ locations

Listing drift across a franchise system

Situation

A franchise system where head office maintained brand standards centrally while individual franchisees managed their own profiles. Over several years the profiles had diverged — naming conventions, hours formats, category selection, and in some cases the brand name itself.

What we found

  • Eleven variations of the brand name across forty-odd locations
  • Holiday hours set correctly at a minority of locations, meaning the rest showed as open when closed — a reliable source of one-star reviews
  • Primary categories differing between locations offering identical service
  • No process for detecting changes, so drift was invisible until someone noticed

What we did

Established one data standard and applied it in bulk rather than location by location. Set a controlled change process so franchisees could request updates without editing directly. Put monitoring in place to alert on any profile edit, by anyone. Built per-location reporting so each franchisee could see their own numbers alongside the head-office roll-up.

What changed

Consistency stopped being a recurring cleanup project. The hours fix alone removed a steady stream of negative reviews that had nothing to do with the food. Governance was the deliverable — the ranking improvements followed from consistency rather than being pursued directly.

AEO / GEOHome servicesAudit

The assistants were recommending competitors

Situation

A home services company ranking well in conventional local search wanted to know whether AI assistants were affecting their enquiry volume. Nobody internally had checked.

What we found

We asked the major assistants the questions customers ask — who handles emergency work in the area, who's available at night, who to call for a specific problem — and recorded what came back.

  • The company was named in a minority of responses despite ranking in the top three conventionally
  • Two competitors appeared consistently, both with substantially more review text describing specific services
  • Several responses cited an outdated service area, traceable to directory records nobody had updated
  • The website stated services in marketing language rather than plainly, making the facts hard to extract

What we did

Rewrote service pages to state plainly what's offered, where, and when — extractable statements rather than positioning. Implemented structured data properly across services and locations. Corrected the outdated directory records at source. Adjusted the review process to encourage customers to describe the specific work done, since that text is what the assistants draw on.

What changed

Appearance in assistant responses improved over the following months, though this is an area where measurement is still developing and we're honest about that. The more useful outcome was that the audit surfaced outdated records affecting conventional search too — which nobody had noticed because the rankings still looked fine.