Why reviews get harder with every site
The single-site reviews guide assumes one person sees every review, knows every customer and decides every reply. With several sites, three things change. First, reviews arrive at every branch at once, and the people who know what happened are at the branch, not in the office that watches the listings. Second, the replies start to vary: one manager is warm and quick, one never replies, one argues, and a reader comparing branches sees a brand that behaves differently depending on the street. Third, the centre starts comparing sites, and comparison changes behaviour. When a branch is ranked on its review count or its rating, the quickest ways up are the ones the platforms and regulators prohibit.
The centralised versus local control comparison sets out the general answer: replies run locally within guidelines, with central monitoring behind them. The multi-location hub reports review volume, rating and reply time per site. This guide fills in the detail between those two: who does what, what the guideline says, what is escalated, how every branch asks and how the centre reads the result.
Who replies: local voices, central guidelines
The split works best when it is written down by review type rather than left to whoever sees the notification first. A RACI matrix is the usual form; the table below is the part most chains need.
| Review type | Who replies | Who is told |
|---|---|---|
| Positive review | The branch manager or a named deputy | Nobody, unless it names a member of staff worth thanking internally |
| Negative review about service at the branch | The branch, within the guideline | The area or operations lead, in the regular summary |
| Negative review about something the branch cannot fix (pricing, policy, a product, the website) | The centre, or the branch using an approved line | The team that owns the issue |
| Review alleging injury, damage, theft, discrimination or a safety problem | The centre, after checking the facts with the branch | Operations, and legal or insurance where relevant |
| Review mentioning a solicitor, a regulator or the press | The centre only | Legal and communications |
| Review that looks fake or from someone who was never a customer | The centre reports it; the branch replies only if asked | The branch, so it can check its records |
| Any review at a branch with no manager in place | The centre or a neighbouring manager | The area lead |
Two rules make the table work. Every profile is owned by the central account with branch staff added as managers, as the guide to location pages and listings at scale sets out, so access survives staff changes and the centre can always reply. And every branch has a named deputy, so holidays and sick days do not leave a site silent.
A reply guideline that is not a script
A guideline fixes what every reply must and must not do; a script fixes the words. Scripts fail under reviews for the reason the guide to building a library of saved replies gives for support channels: a reply that could have been sent to anyone tells the reader that nobody read the review. Under a public review it is worse, because the next customer sees the same sentence repeated down the page.
The guideline usually fits on a page. It applies the brand's Brand VoiceThe consistent personality a brand expresses in its words across channels, with tone adjusting to the situation and the audience. to reviews, in the way the tutorial on writing customer service tone guidelines applies it to support: how formal, how warm, whether staff sign with their first name. It lists what every reply does, taken from the single-site reviews guide: written for the next reader, specific to what the reviewer said, calm, never revealing anything about the customer and never arguing the details in public. It lists what no reply may do: offer compensation in public, admit liability, promise a refund, or mention another customer. It gives approved wording for the facts that are always the same, such as how to contact the branch privately, and leaves the rest to the person replying.
A short set of good and poor replies, drawn from the estate's own reviews with customer details removed, teaches more than rules. Review the guideline when the brand changes how it talks, and whenever an escalation shows a gap.
What goes to the centre and how fast
Most reviews never need the centre. The ones that do are the ones where a reply can create a legal, safety or press problem, or where the branch cannot fix the cause. The table above names them; the guideline should name them too, so that a manager knows to stop and pass the review on rather than reply.
Speed matters more for these than for any routine review, because the reply is read while the story is still forming. Set a response time in the guideline for each kind of escalation, decided by the business, and make the route simple: one place to send the review, one person who picks it up, and a deputy. The centre checks the facts with the branch, decides whether the customer needs a private conversation first, and replies or approves the branch's reply.
A complaint that is serious enough to escalate is a complaint first and a review second. Handle it through the service recovery and complaints guide, whose escalation route covers safety issues, legal threats and customers in vulnerable circumstances. When reviews about the same incident start arriving at several branches, or the story moves to social media, the situation has outgrown review management; the social media crisis plan holds the holding statementsA brief first public response in a crisis that acknowledges the issue, tells affected people what to do now and says when more will follow. and the decision about who speaks.
Asking at every branch without breaking the rules
The single-site reviews guide sets out the rules as Google Business Profile Help and the regulators state them: ask every customer rather than only the ones expected to be happy, offer nothing in return, never ask staff, family or partners, never route happy customers to the platform and unhappy ones elsewhere, and never collect reviews in bursts or on a shared device in the premises. The Competition and Markets Authority and the FTC both treat fake, incentivised and suppressed reviews as consumer protection matters. This is general direction, not legal advice, and a chain running anything beyond a plain request should take advice.
Across an estate the risk is not that head office breaks those rules; it is that the way head office measures branches pushes them to. A league table ranked on new reviews, a bonus tied to the rating or a target for reviews per week makes the prohibited practices the fastest way to win. A manager under pressure starts asking only the regulars, sets up a tablet by the till, or asks the team to post. The diagnosis for disappearing Google reviews traces what happens next: reviews filtered or removed, and a profile that looks worse than before.
Make the ask the same at every branch and measure the ask, not the result. Put the review link in the same place in every customer journey: on the receipt, in the booking confirmation, in the follow-up message after the appointment, so every customer receives it whatever the branch's mood. Train every branch on the rules once and again when people change. Recognise branches for replying well and for fixing what reviews raise, not for the count. Review numbers still appear in the report; they are a check on whether the ask is working, not a target.
One monitoring view and a report by site
Every branch's reviews should reach one place the centre can see, with alerts to the branch and its deputy when a review arrives. The reviews guide describes the move from notifications to a monitoring tool that pulls several platforms into one inbox; for a chain it is close to essential, because the alternative is someone opening every profile in turn. Business Profile Help documents organising many profiles into groups with shared access, which is the starting point for that view on the search listing.
The report then answers the hub's question for reviews: how is the brand doing, and which sites need help. For each site and each month, show new reviews, the rolling rating, the share of reviews replied to and how long replies took, the number escalated, and the main themes. Choose a small set of KPIsA metric tied directly to an objective. the branch can influence; reply rate and reply time are entirely in its hands, while the rating is only partly.
| Report column | What it shows | How to read it |
|---|---|---|
| New reviews | Whether the ask is reaching customers | Against the site's own history and its customer numbers, never as a target |
| Rolling rating | What a reader sees first | As a range over several months; a few reviews move it a lot |
| Reply rate and reply time | Whether the branch is answering | Directly; this is the branch's own work |
| Escalations | Reviews that carried risk | By cause, to see what keeps recurring |
| Themes | What customers keep mentioning | Across sites, as the next section describes |
Read every figure against the site's own past. Review counts at one branch are small, and the guide to reading small numbers explains why a rating built on a handful of new reviews jumps around and why a range says more than a single figure. Compare a site with sites of a similar type, as the hub advises, rather than with the chain's best.
Learning across sites
The single-site guide reads a quarter's reviews together to find fixes. A chain can do something a single business cannot: read the same theme across many sites and tell a branch problem from a brand problem. Code each review's main theme with a short list everyone uses, such as waiting time, staff manner, quality, price, cleanliness, booking and parking, then count themes by site each quarter.
A theme that appears at one branch is that branch's to fix. A theme that appears across many is the brand's: a policy, a product, a process or staffing levels at a certain kind of site. Reviews are an unsolicited source of customer feedback, and the voice of the customer guide explains how to put them beside surveys and close the loop, both with the customer and inside the business.
What the themes show also feeds marketing. Reviews are Social ProofEvidence that other people have chosen, used or approved something, used to reduce a buyer's doubt., and the strongest ones at a branch, shown with permission or through the platform's own display, belong on that branch's location page. The same themes tell each branch what its replies should acknowledge and what its customers value most.
A car wash chain brings its branches into one routine (illustrative)
Ulvexmar Car Wash moves replies to its branches
Ulvexmar Car Wash, an invented chain of 14 hand car washes, had replied to every review from head office, where one marketing assistant worked through all 14 profiles on Fridays. Replies arrived about a week late and read the same at every site, and a branch league table ranked managers on new reviews each month, with a prize for the top three.
The chain moved routine replies to the branches. Each manager and a named deputy were added to their own profile under the central account, given a one-page guideline with approved wording for contact details and a set of good and poor replies taken from the estate's own reviews. Reviews alleging damage to a car, or any safety complaint, go to the operations lead, who checks the branch's camera log and replies or approves the reply.
The league table was dropped after an internal check found two branches asking for reviews only from regular customers. In its place, the review link went onto every receipt and the report showed reply rate and reply time by site.
| Measure | Before (quarter) | After (next quarter) |
|---|---|---|
| Reviews replied to | 61% | 94% |
| Typical reply time | 6 days | 1 day |
| Reviews escalated to operations | Not tracked | 11 |
| New reviews across the chain | 412 | 389 |
New reviews dipped slightly once the regulars-only asking stopped. A quarterly theme count across the 14 sites then showed that waiting time, not washing quality, appeared in most low ratings at the four busiest sites, and in very few elsewhere. Those four added a second bay team on weekend mornings.
Common mistakes
- Head office replying to everything. Replies arrive late and read the same; let branches reply within a guideline and keep the centre for the reviews that carry risk.
- A script instead of a guideline. The same sentence under every review tells readers nobody read them.
- Ranking branches on review counts. It rewards the practices the platforms and regulators prohibit.
- No escalation list. A manager replies to an injury claim the same way as to a complaint about the coffee.
- Branch staff owning profiles. When they leave, access leaves with them; the central account owns every profile.
- No deputy. A site goes silent whenever its manager is away.
- Reading one month's rating as a verdict. Branch counts are small; read ranges and each site's own history.
- Reading reviews one site at a time. A theme across many sites is a brand problem no single branch can fix.
