OperationsHow to reply to negative reviews across a franchise
A reply standard franchisees can follow: what a good response includes, worked examples, and which complaints should go to head office.
A step-by-step method for multi-site brands to track Google reviews branch by branch, the measures to watch, and where spreadsheets stop working.

To track Google reviews across every branch, pull each site's new reviews into one sheet every week, tag them against a fixed list of themes, and compare each branch with the others rather than with your brand average. Google's own tools and a spreadsheet will carry this to around ten sites. Beyond that, reading and tagging becomes a job in itself.
This guide is for marketing and operations leads at café, bakery and restaurant groups, whether the sites are company-owned or franchised.
Customers judge the branch in front of them, not your brand as a whole, so a healthy brand average can hide the sites that are costing you visits. When someone searches for a café nearby, Google shows them that one site's rating and reviews.
The Competition and Markets Authority estimates that as much as £23 billion of UK consumer spending is potentially influenced by online reviews each year (CMA, January 2025). In BrightLocal's 2026 survey of 1,002 US adults, 97% said they read reviews for local businesses and 74% looked for reviews from the last three months (BrightLocal, 2026). The same survey found 68% would only use a business rated four stars or more. It is a US sample, so read it as a guide to behaviour rather than a UK measure.
Two things follow for a multi-site brand:
Google lets you manage many locations from one account and reply to each site's reviews. It doesn't compare branches for you or tell you what the reviews are about.
What it does:
What it doesn't give you as standard:
Google is also not the only place your customers write. BrightLocal's 2026 survey found Google's share as a review source fell from 83% to 71% in a year, and people used six review sites on average (BrightLocal, 2026). For food and drink brands, add Tripadvisor and the delivery apps if they matter to your sales.
Use one sheet with one row per review, a fixed list of themes and a weekly routine. Six steps cover it.
Here are the columns to start with, with one example row:
date,branch,stars,theme,review_text,replied,reply_date
2026-10-06,MAN-01,2,wait time,"Twenty minutes for two flat whites.",yes,2026-10-07
Track the same six measures for every branch over a rolling 90 days, then rank branches against each other. Ninety days matches the window most review readers look at: 74% in BrightLocal's 2026 US survey sought reviews from the last three months.
| Measure | How to work it out | What it tells you |
|---|---|---|
| 90-day average rating | Average stars of reviews from the last 90 days | How the branch looks to people reading now |
| Gap to group median | Branch's 90-day rating minus the median across all branches | Which sites most need attention |
| Share of 1–2 star reviews | 1–2 star reviews ÷ all reviews, last 90 days | Problems an average smooths over |
| Top complaint theme | Most common theme among 1–3 star reviews | What to fix first |
| Reply rate | Reviews with a reply ÷ all reviews | Whether the branch is keeping up |
| Median reply time | Median days between review and reply | Whether replies arrive while they still matter |
As a working rule, treat a branch with fewer than 20 reviews in 90 days as directional only. One bad week can move its average a long way.
Replying is worth measuring because it can move the rating itself. Researchers Davide Proserpio and Georgios Zervas found that hotels on Tripadvisor received 12% more reviews after they started replying, and their ratings rose by 0.12 stars on average (WARC, 2018; Marketing Science, 2017). The study covered hotels, not restaurants, so read it as a guide rather than a promise.
The spreadsheet method usually stops working somewhere past ten sites, because reading and tagging grows with every branch you add. The arithmetic is simple. At 15 new reviews per branch per week and 30 seconds to read and tag each one, ten sites take about 75 minutes a week. Twenty sites take 2.5 hours, before anyone writes a reply.
Three other things tend to go wrong as you grow:
This is the work Akili was built to take on. It reads reviews and social comments for every branch, sorts them by theme, scores each site on the same measures and ranks what to fix first. It's well suited to UK café, bakery and restaurant groups. If you'd like to see it on your own branches, book a demo.
You can manage every location from one Google account using Business Profile Manager, and reply to each site's reviews there. To compare branches over time, you'll need to bring the reviews into a spreadsheet or use a review tool.
Check for new reviews daily so replies go out quickly, and look at the branch-by-branch picture weekly. In BrightLocal's 2026 US survey, 32% of consumers expected a reply by the next day.
Either can work, as long as one named person owns replies for each site and everyone follows the same guidelines. One workable split is branch managers replying to routine reviews, with serious complaints passed to head office. Our guide to replying to negative reviews across a franchise sets out who should reply to what.
There's no fixed number. As a working rule, treat fewer than 20 reviews in 90 days as directional, because one bad week can move the average a long way.
Yes, through the Google Business Profile API, which returns each review's rating, text, date and reply for each location. Without a developer, the practical route is to copy new reviews into your sheet during a weekly check.
OperationsA reply standard franchisees can follow: what a good response includes, worked examples, and which complaints should go to head office.
Customer insightHow one weak branch pulls down a multi-site brand's rating, why the group average hides it, and a simple method to find the branch that's costing you most.
We will show you how customers see every location, and which fixes will grow revenue and cut churn.