A local SEO agency works on the searches where the person searching is standing somewhere, and where being nearby and being known count as much as having a good website. Most of what decides those searches sits on surfaces you do not own, which is what makes the discipline unusual.

Two estates get managed at once. The first is the listing layer, meaning your Google Business Profile plus the maps, directories and data sources that describe your business without ever asking you. You influence that layer and you do not control it. The second is your own site, where pages about places and services have to earn ordinary positions and then give somebody a reason to phone. A vendor working only on the first produces a tidy profile and no growth. A vendor working only on the second keeps wondering why the three listings above the links take all the calls.

What we own. Which searches are worth winning inside your catchment, meaning the area you can realistically serve. The profile and listing work behind them. The pages on your site that carry local intent. The review process. The earned coverage that builds standing in your city. And reporting tied to calls and booked work.

Why clients pick us for it. Search, design and build sit with one senior team, so when a location page needs rebuilding nobody files a ticket with a third party and waits. That matters more here than in most search work, because local gains are so often blocked by something structural in the site.

What we refuse. Listings at addresses you do not occupy, incentives offered for reviews, and city pages generated in bulk. Each still works briefly and each carries a penalty you inherit, not us.

Brand Vision has run this work since 2018, with over 500 projects behind it, and the 250+ five-star reviews attached to our own name came from asking properly, which is the same mechanic we build for clients. Outside verification helps too, and ours is a 5.0 on Clutch drawn from over 64 recorded client interviews. The average lift in leads we have recorded across engagements sits at 250%, and that figure needs its ceiling said out loud. It blends programs of unlike size and unlike starting condition, several running well past a year, and it records what has already happened elsewhere instead of promising anything about your market. Look through what we have shipped and the people who run it before you commit a dollar, and the wider search practice is where all of this sits.

Eight, run in a deliberate order instead of sold as a bundle, because doing them out of sequence wastes months. Listings get corrected before pages get built, and pages get built before authority work begins.

1. Google Business Profile optimization and management. Categories, attributes, services, photos, hours, posts and the questions section, set properly once and then maintained. Maintenance is the part vendors skip, and profiles decay without it.
2. Local keyword and map pack strategy. Which searches carry your demand, which of them trigger map results at all, and which towns or neighbourhoods are winnable from where you actually sit. This decides what everything else aims at.
3. Location and service-area pages. Real pages for the places you operate in, carrying substance a competitor could not paste their own name into. One page per place you can defend, and no page for a place you cannot.
4. Citation and NAP consistency. NAP is the shorthand for name, address and phone number. The work is making those three agree across the platforms that feed the rest of the web, then clearing the duplicate listings that split your signals in half.
5. Review generation and response. A process that asks every customer at the right moment, plus replies to all of them. Built to run inside your operation, since a review program that depends on an agency stops the month the agency does.
6. Local link and community authority. Coverage from the organizations, publications, associations and events in your actual market. Slow, hard to copy, and the reason established competitors are difficult to shift. It uses the same off-page discipline as the rest of the practice, aimed at a local target list.
7. Multi-location and franchise management. Profiles, pages, permissions and standards across many sites, plus the competition between your own pages that nobody notices until positions start swapping week to week.
8. Local reporting. Calls, direction requests, enquiries and booked work per location, read against the previous period and against the same month last year.

Two things sit alongside these instead of inside them. Sector rules change the work, and for clinics and practices what you may ask of a patient and what you may publish is set by a regulator and not by a marketer, so that gets established before any review process is designed. And where a read has to come first, before any program gets committed to, a standalone review will tell you whether local is where your demand genuinely lives.

A local search puts three separate answer boxes on one screen, and each one is won with a different mix of work. A national query gives you one set of links to compete for. A query with a place attached, or one typed on a phone in a particular spot, gives you three, and they do not share a scoreboard.

The first is the map pack, the block of business listings with pins that sits near the top. Nothing on your website ranks there. The entry ticket is a verified business listing, and the sort order comes from how well that listing matches what was typed, how far you sit from the searcher, and how well known the business is. A company with a mediocre website and an excellent listing beats a company with the reverse, which offends people and is true.

The second is the set of ordinary blue links underneath, reordered for the searcher's location. Your website competes here on normal terms, with the qualification that a page carrying real local substance outranks a page that merely contains a city name. Directories, aggregators and best-of lists occupy a great deal of this space, and earning a place on those lists is frequently faster than outranking them.

The third is being named when somebody asks an assistant for a recommendation instead of typing a query. That answer gets assembled from listing data, review text and whatever independent sources have written about your category, then compressed down to two or three names. There is no position four to aim at. What can be measured is how frequently your name comes up over a fixed list of prompts and which sources get cited beside you, which is the honest limit of AI search work and why we describe it as share of mentions instead of ranking.

The commercial point is which surface holds your demand. In some categories the map pack takes almost everything and the links underneath take very little, and restaurants are the clearest case, where the decision often happens without a website ever opening. In others the reverse holds. Property searches usually get settled on a page with listings and neighbourhood detail on it, with the map results doing far less of the work. Buying a tactic before knowing which of the three surfaces your calls come from is the most common way a local budget disappears.

Google sorts local results on three things it states publicly, meaning relevance, distance and prominence, and only two of the three are inside anybody's control. That single fact is why honest local SEO looks different from what most vendors sell.

• Relevance. How well your listing and your site match what somebody typed. Category is the heaviest lever, followed by the services and products listed on the profile, the business name as it legitimately appears, and the language on the page the listing points at. Being relevant for emergency plumbing and being relevant for bathroom renovation are two different jobs even when one company does both.
• Distance. How far you sit from the person searching, or from the place named in the query. This is a hard constraint. If a competitor is four blocks from the customer and you are eleven kilometres out, no amount of work lifts you above them for that particular searcher. It also means your ranking is not one number. It changes point by point across a city, which is why we sample it on a grid of locations and report the share of your area where you appear.
• Prominence. How well known the business is, assembled from reviews, links, mentions in local press and directories, and the general weight of your name online. This is the slow input, the one competitors cannot copy quickly, and the one most vendors leave alone because it takes real work. Coverage earned from organizations and publications in your own city does more here than a hundred directory submissions, which is why we run it as earned authority against a local target list.

So the ceiling, stated plainly. Nobody can guarantee the top three, because one of the three inputs is your address next to the customer's. What can move is relevance almost immediately, prominence over months, and the proportion of your service area where you appear at all. In a dense downtown market like Miami that proportion gets won street by street. Across a spread-out region it is a longer job with a lower ceiling, and saying so upfront beats selling a radius we cannot cover.

Two things distort the picture when you compare yourself to a rival. A competitor may be ranking on a business name stuffed with keywords, which breaks the guidelines and can be reported. And in legal services and other high-value categories the incumbents usually hold prominence built over a decade, so the realistic first target is the searches and the neighbourhoods where they are weakest.

Local SEO runs as a monthly program on a flat fee agreed in advance, and movement worth reporting takes four months to a year depending on how contested your catchment is. A vendor promising the top three inside thirty days is either chasing phrases nobody types or is not telling you the truth.

How the time actually goes. The first two to three weeks are audit and access, meaning claiming and verifying every profile, pulling the listing records that already exist about you, checking what the site currently does with local intent, and reading how the businesses winning today got there. Month one to three is corrective. Category and profile work lands, the worst listing errors get fixed, and the first genuine location or service-area pages ship. Movement can arrive quickly in this phase because much of it is repairing something broken. Month three to six is the building phase, where review volume accumulates, local coverage starts landing, and further pages go live. Month six to twelve is where compounding shows, and where prominence work separates you from businesses that only tidied a profile.

What moves the number.

• How many locations you have, since every staffed address is its own profile, review stream and page
• Competition in the category, which varies enormously between a suburban trade and a downtown clinic
• The state of your existing listings, because a business that has moved twice or changed names once carries a real cleanup cost
• How many searches you intend to win, and how far apart the places are
• Whether reviews need a process built from nothing or already arrive steadily
• Whether the site can carry new pages without development work

Two commitments worth stating before you ask. We work to a six-month minimum, because a program abandoned in month four has paid for the build-up and collected none of the payback. And every account stays in your name, meaning the profile, the listing logins, the review platform and the analytics. Nothing gets built in a way that requires us afterward.

Where the right answer is a smaller purchase than the one you enquired about, you will hear that. Sometimes a single audit and roadmap plus one month of cleanup is the whole of it, with your own team running reviews from there. And where you need calls before month four, paid search covers the near term while the slower work builds, which is frequently the right combination instead of a choice between the two.

If you sell nationally, have no premises customers visit, and nobody chooses you based on where you sit, local search matters far less than the rest of the discipline and you should not buy it from us. That answer costs us work every time we give it, and it is still the right one.

Three tests settle it quickly.

Does a place name appear in the searches that matter. Read the queries already producing enquiries. If almost none of them carry a city, a neighbourhood, or the words near me, the map results are not where your demand lives. Two companies in the same industry can fail this test in opposite directions.

Would a customer care where you are. Some buyers genuinely need proximity, because somebody has to arrive or they have to travel. Others never think about it once. For most business-to-business software the vendor's address has no bearing on the decision, and dressing that up as a local opportunity costs a quarter nobody gets back.

Do you have a real address with people at it. Google requires either a location where you meet customers or a genuine service area worked from a real base. Renting a mailbox to manufacture a listing is a suspension risk and we will not set one up. Companies do it, the listings come down, and the removal frequently takes the good listing with it.

What a smaller version looks like for a business that fails the tests and still wants the basics right. One properly built profile at your head office, correct categories, a clean set of records so your details agree everywhere, and then stop. That is a few hours of work once, and not a retainer. Everything after it belongs in the parts of search that carry your actual demand, which for a national store is catalogue and category work and for most others is content and authority.

The case that looks like this service and is not. A brand with warehouses in six cities and no walk-in trade does not need six profiles, because a warehouse nobody visits is not a location in the sense Google means. What it may need is pages for the regions it genuinely serves, which is website work and has nothing to do with listings.

One last note against our own interest. Plenty of businesses fail every test above and still appear in map results for their own brand name. That listing deserves to be accurate anyway, because it is the first thing a buyer sees when checking that you exist, and correcting it is an afternoon.

A Google Business Profile is a ranking asset and a conversion asset at once, and most we inherit are half filled in. The fields that decide the outcome are rarely the obvious ones.

• Primary category, chosen from Google's fixed list. The heaviest relevance signal on the profile, and it governs which searches you are eligible for at all. Picking the category that describes what you wish you sold instead of what you mostly do is the expensive mistake here, and the useful test is which one the businesses already winning have chosen.
• Secondary categories, used sparingly. They extend reach into adjacent searches. Adding eleven of them blurs the picture and drags you into filtered results you convert badly on.
• Attributes. The structured facts that feed filters, covering wheelchair access, parking, outdoor seating, appointment requirements, languages spoken and ownership descriptors. Some you set and some get assembled from customers. An empty attribute set makes you disappear the moment somebody narrows a search, and for a hotel or venue attributes are most of what filtered search reads.
• Services and products, entered individually. Every entry is a matching opportunity, and most profiles have none filled.
• A description written for a human. It carries little ranking weight and it does get read, so it should say what you do, who for, and what is different.
• Photos and video, kept current. Two profiles at the same position do not convert the same. The one with recent, real imagery of the actual place wins the visit, which is where a food and beverage profile is made or lost.
• Hours, including holiday hours. Wrong hours produce a particular kind of one-star review that is very difficult to answer.
• Posts and the questions section. Posts are short updates for offers, events and news. They do little for ranking and they occupy visible space on a branded search, which makes them a conversion tool. The questions section matters more than people expect, because anyone can answer a question about your business and the most upvoted answer is the one everybody sees. Seeding the real questions with correct answers, then watching them, stops a competitor or a confused stranger speaking for you.

If you travel to customers and have no premises they visit, you are a service-area business and the rules change. The address gets hidden, the area is defined by the places you genuinely work in, and a coworking desk or a mailbox will not survive verification. Businesses doing both keep the address and add the area. In a bilingual market such as Montreal the description, the service list and the review replies carry a language decision too, and defaulting to English quietly costs you.

Reviews move rankings and they move the decision, and the two respond to different things. Volume and average rating do most of the ranking work. Recency, specificity and your replies do most of the persuading.

On generation. A steady process beats a campaign every time. Asking at the moment the customer is happiest, through the channel they already use, with a link that opens the review form directly, produces more reviews than any amount of signage or reminding. What is not permitted is choosing who gets asked based on how you expect them to feel, which is called gating and which platform policy prohibits. Paying for reviews or trading a discount for one is prohibited as well, and it is exactly the sort of thing a competitor reports. So we build the asking into the operation, where it survives us leaving.

On volume against recency. Total count and average rating are the durable signals, and a business sitting on four hundred reviews holds an advantage nobody closes in a quarter. Recency is the part you can compete on immediately. Eleven reviews in the last two months reads as a going concern, and a profile whose newest review is three years old reads as closed no matter what the total says. Steady flow also protects the average, since one bad week lands much harder on a stagnant profile. The target is a rate, not a number.

On owner responses. Reply to all of them, briefly, in a recognizably human voice, and reply to the good ones as well. Responses get read by the next customer far more than by the person who wrote the review, and a defensive reply to a two-star review does more damage than the review did. What works is acknowledge, move it to a phone call, and say what changed. Never argue the facts in public even when you are right, because your audience is a stranger deciding whether to call.

On critical reviews. A flawless five average reads as suspicious, and a few negative reviews handled well is credible. Reviews that break platform rules can be flagged and sometimes removed. Reviews that are merely unfair usually cannot, so the durable answer is more recent good ones sitting above them.

On everywhere else. Google is the ranking one and it is not the only one being read. Sector platforms, industry directories and Apple's listings all get consulted. In health and wellness and at law firms the profession's own rules on soliciting, replying to and displaying reviews override anything a marketer would prefer, and that belongs with your regulator or your counsel before a process gets designed.

Thin city pages fail because they answer no question a real person has, and because a search engine can plainly see the same page repeated twelve times with the place name swapped. Duplication is the visible symptom. Emptiness is the actual cause.

Consider what somebody in that city wants to know. Whether you truly work there. How quickly you can get to them. What you have done nearby. What it costs in that market. Who to phone. A template page tells them none of it, so it earns no links and no engagement, and it competes with your other eleven city pages for the same terms. Publishing sixty of them can suppress the two that would have worked.

A location page worth building carries most of the following.

• The address, hours and phone number for that specific place, matching the profile exactly
• The people who work there, named, because a local page with a real person on it converts differently
• Work, projects or cases from that market, with places a resident would recognize
• The services genuinely offered there, since branches rarely carry identical ranges
• Directions, parking, transit and access detail only somebody who has been there could write
• Reviews from customers in that market instead of the company-wide feed
• Wording a competitor could not paste their own name into

A service-area page is a different object and needs treating as one. It exists for a place you travel to and cannot claim to be based in. What makes it legitimate is evidence of actual jobs completed there, response times, coverage boundaries stated honestly, and content about that area's particular conditions, meaning older housing stock, a specific bylaw, a commute that changes the schedule. Across a metro the size of Chicago those differences between neighbourhoods are real and worth writing about. Across a list of suburbs you have never worked in there is nothing to write, which is the clearest signal the page should not exist.

Scale is the whole decision. Build a page for a place only when you can produce genuine substance for it and you intend to compete there. Six real pages beat forty empty ones, and the six can be researched properly inside the same content program that handles the rest of the site. For real estate and development this is close to the entire game, since neighbourhood-level pages are what buyers search and the market data behind them is the reason anyone links to you.

Citations are hygiene, which means getting them right unlocks nothing on its own and getting them wrong holds everything else back. The goal is agreement, one version of your name, address and phone number everywhere a machine can find it.

Five steps, in this order.

1. Inventory what already exists. Search your name, your phone number and every address you have ever used. Most businesses find listings they never created, generated automatically from purchased data, plus one or two duplicates of the profile they do manage.
2. Settle the source of truth. Decide the canonical version of the name, the address format, the phone number and the hours, then make the website and the primary profile agree with each other. Structured data on the site helps machines read those details without guessing, which is one small part of the technical track.
3. Correct the platforms that feed the others. A handful of data aggregators supply business records to a long tail of small directories, so a wrong record upstream regenerates itself weeks after you fix the directory. In North America that means the main aggregator feeds plus the first-party platforms, covering Apple, Bing, Yelp and the national yellow pages listing.
4. Merge or suppress duplicates. Two live listings for one address split your reviews and your signals and can hold both of them down. This is the highest-value item on the list and it is tedious, which is precisely why it goes undone.
5. Monitor, because records drift. Suite numbers get dropped, a call-tracking number replaces the real one because somebody added it without telling anyone, and an aggregator refreshes an old record over your correction.

Two honest notes. Past the platforms that genuinely matter, additional citations do very little, and a package promising three hundred listings is selling volume because volume is cheap to produce. And call tracking can be used safely here, so long as the main number on the profile stays the real one and the tracking number sits in a secondary field.

The event that breaks all of it at once is a move. A new address invalidates every record simultaneously, and the businesses that suffer worst are the ones that update the website and nothing else. If you are relocating in a market like San Francisco, the listing migration deserves planning alongside the lease, and a name change or a merger behaves the same way. Where the picture is genuinely unclear, a full audit tells you what is out there before anyone starts editing.

Multi-location local SEO is a governance problem wearing a marketing costume, and the failure is almost always your own pages and profiles competing with each other. Ten locations is not ten times one location.

One profile per staffed location, and no more. Every address where customers are served, or where work is genuinely based, gets its own verified profile with its own categories, photos and review stream. Departments and individual practitioners sometimes warrant an additional listing and usually do not. Listings created purely for coverage get removed, and the removal is rarely surgical.

One page per profile, each genuinely different. The profile points at the page for that location and nowhere else. A profile pointing at the homepage throws away the strongest signal it has to give.

Cannibalization is the recurring problem. Two pages chasing near-identical terms in adjacent areas split their own signals, and the search engine picks whichever it prefers, which is often neither of the ones you wanted. It shows up as pages trading positions week to week. The fix is deciding which page owns which term, differentiating the content until the distinction is real, and consolidating the pages that were never going to compete.

Structure that scales. A locations index linking every page, breadcrumbs, and consistent templates with a section per site for content nobody at head office could write. At forty locations this becomes the same discipline as holding standards across a large estate, which is why it borrows directly from work at that scale.

A defined split between head office and the location. Brand, templates, categories and data standards centrally. Photos, hours, review replies and local events at the location, because nobody at head office knows the parking changed. Franchise systems need that written down, including who funds what and what a franchisee may publish on their own, since ambiguity there is where the friction always starts.

Honest about the cost. Running forty profiles and forty review streams is headcount, and software does not change that. Tools genuinely help with bulk edits and monitoring, and they do not answer a review in a way anybody believes. We staff this deliberately and hire for it instead of spreading it across people already doing something else.

For groups operating across provinces or states, reporting has to hold at two levels, per location for the manager and rolled up for whoever holds the budget, with the weakest locations visible instead of averaged away. Our own practice runs from Toronto alongside three US offices, so the arithmetic of one brand across separate markets is something we do to ourselves before we do it to a client.

Reporting is against calls, enquiries and booked work per location, never against impressions. A local report leading with how many times a listing was displayed is dodging the only question that matters, which is whether the phone rang more than it did last month.

What arrives every month.

• Calls, messages, direction requests and website clicks from each profile, with the trend and a same-month-last-year comparison, because a January figure means nothing beside December
• Forms, bookings and enquiries by location, tied to conversion tracking that was verified before anything started, since numbers built on a broken tag mislead more than no numbers would
• Map visibility sampled across a grid, meaning your position at points spread through the service area, which shows the share of your territory where you appear instead of one flattering position
• Localized organic positions for the terms we chose to pursue, including the ones we deliberately left alone and the reason
• Review count, average rating and rate of new reviews per location, plus how many got a reply, since those are the leading indicators for everything else
• Appearance in assistant answers across a fixed panel of prompts for your category, reported as share of mentions and cited sources, which is the honest limit of what generated answers let anybody measure
• Listing health, covering new duplicates, suspensions, unauthorized edits and hours that reverted on their own
• What we recommend next, with a reason attached, and last month's recommendations marked done or not done

Where a number can only be directional, the report labels it that way instead of dressing it up as exact. Local attribution has real limits. Somebody who finds you in the map results and phones from the pin never touches your website, and a share of the calls a profile reports are wrong numbers and suppliers. So the profile data gets read alongside your own record of where enquiries came from, and we ask you to keep asking callers how they found you. It is unglamorous and it is the most accurate source available.

Two things about how this runs. Every profile edit, page change and outreach task is logged where you can open it at any hour, so nobody waits for a monthly call to learn what happened. And a bad quarter gets explained by us before you find it, with the change of plan attached.

Where you would like an outside opinion on what your current numbers are hiding, or on how the search program around this should be sequenced, send us what you have and you will get a straight answer, including the case for spending the money somewhere else.