How agencies find local businesses whose ratings lag their neighbors

Reputation and local marketing agencies find their best prospects by comparing each business's rating with the businesses of the same kind around it. A hotel at 3.6 is in trouble on a street where the others average 4.4, and doing fine where they average 3.5. So the useful list is not "every business under 4 stars" but businesses rated clearly below their local average, with enough reviews for the rating to be real, run by an owner who can say yes, and ideally missing something an agency fixes, such as a website. Found that way, every name on the list comes with a reason to call.

This is for agencies selling reviews management, local search, websites and social media to restaurants, hotels, clinics, salons and trades. It covers how to define a weak rating for an area, which weak ratings are worth pursuing, a worked example, and an AI employee that delivers new prospects every week, with first messages that wait for your approval.

Why a lagging rating is a buying signal

Owners know their rating. When it sits below the competition, they feel it in bookings and walk-ins, and many have tried to fix it alone. An agency that arrives with a clear comparison (you are at 3.7, the eight places within a kilometer average 4.3) is describing a problem the owner already has, with numbers they can check. That is a much easier opening than a general pitch about marketing.

Defining weak for the area, not in general

Ratings differ a lot between categories and cities, so a single threshold misleads. Two markets from xMap's places database, within 10 km of each city center, November 2026, show the range:

3.8Hotels, Toronto: average rating

47Toronto hotels rated 4.5 or more, of the 167 with a rating

3.6Supermarkets, London: average rating

38London supermarkets rated 4.5 or more, of the 423 rated among the 437 nearest the center

A Toronto hotel at 3.6 is only slightly under its city's average; a dentist at 3.6 in most cities would be far below it. The method that holds up everywhere is to compare each business with its own neighbors: Nearby places, set to give each incoming place a count, returns the average rating of the places of a type within the distance you choose (nearby_avg_rating), so the gap can be worked out place by place with a Calculated field.

Which weak ratings are worth a call

• A real gap. At least 0.5 below the local average is a fair starting line; adjust once you see your results.
• Enough reviews to matter. A place with three ratings is noise. Too few reviews can itself be the problem you solve, but pitch it differently.
• Independent, not a chain branch. Branch managers rarely buy agency services; owners do. Keep only independents.
• Something missing. No website or phone on the listing is a second, easy win to offer.
• Still open and trading. Leave out places marked closed.

A worked example: restaurants in Nairobi

Say an agency in Nairobi sells reviews programs and simple websites to restaurants. Its flow searches three neighborhoods it can serve in person and finds, made-up figures here, 140 restaurants. For each, it works out the average of restaurants within 1 km. Twenty-two sit at least 0.5 below that average; eleven of those have more than 30 ratings; eight are independent; three of the eight have no website on their listing. Those three get a call this week, the other five get an email. The agency starts with eight warm prospects instead of 140 cold ones.

The prospect finder as an AI employee

1. Schedule: every Tuesday morning.
2. Area of interest: the neighborhoods or cities you serve.
3. Nearby places: the category you sell to, independents only, open places only.
4. A second Nearby places, Each incoming place with counts, 1 km: the same category around each, which adds the local average rating.
5. Calculated field: rating_gap = nearby_avg_rating - rating.
6. Filter: rating_gap of 0.5 or more.
7. Only new: only businesses not sent before.
8. xMap LLM: a short, polite first message per business that states the comparison and offers one specific help, using only its own figures.
9. Outreach with Approve each first switched on, or simply Email the list to the sales lead.

What lands in the inbox each week is a handful of new names, each with its rating, the local average, the gap, its phone and website status, and a draft message to approve, edit or drop. For clients you already serve, a When something changes trigger can watch for a rating falling below a line you set, so you call them before they call you.

Opening the conversation without insulting anyone

Lead with the neighborhood, not the flaw: what the area looks like, where they sit, and what similar businesses did. Offer something useful in the first message, such as the three most common themes in recent reviews you read yourself, or a note that their listing has no website. Never threaten, never promise a rating, and never offer fake reviews; those are against the rules of the platforms that host them and against the law in many countries.

Mistakes that waste the list

• One national threshold. Compare with the neighbors, not with a number.
• Chasing chain branches. The decision is usually made at head office.
• Ignoring the review count. A low rating on very few reviews is a different pitch from a low rating on hundreds.
• Sending the same business twice. Use Only new, and keep a list of who declined.
• Cold messages nobody approved. Keep approval on, at least until the team trusts the drafts.

Templates to start from: businesses with low ratings and no website, more reviews for my clients, rivals near me whose rating drops and rival marketing agencies in my city. More for agencies on the marketing and agencies page.