I have audited a lot of citation profiles. Not in a heroic, look-at-me way; it is just what happens when you spend a decade telling small business owners why their map pack ranking dropped after a phone system change. The pattern repeats. Someone moved suites. Someone bought a vanity number. Someone let a marketing intern submit to 200 directories in 2019 and nobody has touched it since. The result is the same: a citation profile that contradicts itself, and a Google Business Profile that quietly slides out of the local pack.
This article introduces a framework I have been using for client work, which I call CANVAS. It is not magic. It is a way of forcing yourself to think about citations as a structured dataset rather than a checklist of submissions. If you read the whole thing you should be able to apply it on Monday morning.
Why most citation strategies underperform
The honest reason most local SEO citation work fails is that it gets treated as a procurement task. Buy 100 listings. Submit to a service. Tick the box. The problem with treating citations like office supplies is that search engines do not read your submission queue; they read the actual data on the actual web, including the bits you forgot about.
The volume-over-accuracy trap
Citation volume was a useful proxy in 2012. It is not now. According to StackMatix’s analysis of what still matters in local SEO citations, most businesses benefit from 40 to 60 citations across major aggregators, industry directories, and local sources. After that, you hit diminishing returns fast. I have seen profiles with 180 citations rank worse than competitors with 35, because the 180 included six different versions of the phone number.
The trap is psychological. Volume feels like progress. You can show a client a spreadsheet of submitted directories and it looks like work. Whereas auditing the existing mess, finding the abandoned 2016 Yelp listing with the old suite number, and getting it corrected, looks like nothing on a status report. But it is the work that moves the needle.
Why “list everywhere” advice fails
Generic submission lists treat all directories as equivalent nodes. They are not. A citation on a directory with 800 indexed pages and zero organic traffic is not a citation; it is a liability waiting to drift out of sync with your real data. The Jasmine Directory guide on Jasmine Business Directory recommends a simple sanity check: run a site: query on Google for any directory you do not recognise. If fewer than 1,000 pages are indexed, the directory has no organic presence and your citation there carries no weight.
I would add a second check: search for an existing customer of that directory by exact name. If the directory’s own profile pages do not rank for the businesses listed on them, the citation is invisible to both users and the search engines that might use it as a corroborating signal.
Hidden costs of inconsistent data
Inconsistency is not a passive state. It actively costs you. Algorithms have to decide whether “Smith and Sons”, “Smith & Sons Ltd”, and “Smith & Son’s” describe the same entity. When they cannot decide cleanly, they hedge, which means your business gets less weight in the entity graph than a competitor whose data agrees with itself.
Did you know? Out of the top 6 foundational ranking factors for local SEO identified in Moz’s industry-standard Local Search Ranking Factors report, 3 are citation-related, and 6 of the top 13 competitive difference makers involve citation quantity, quality, and consistency. Source: UFO Rocks summary of the Moz report.
The CANVAS citation framework defined
CANVAS stands for Core, Authority, Niche, Verification, Audit, Sync. Six layers, applied in that order, treated as a continuous loop rather than a one-time project. The name is a mnemonic, not a brand. I am not selling you a course.
classDiagram
class CanonicalRecord {
+String businessName
+String streetAddress
+String suite
+String city
+String state
+String zip
+String phone
+String website
+defineCanonical()
+validateField()
}
class CoreLayer {
+GBP
+BingPlaces
+AppleBusinessConnect
+Facebook
+setSourceOfTruth()
}
class AuthorityLayer {
+DataAxle
+Foursquare
+Localeze
+Yelp
+pushAggregatorData()
}
class NicheLayer {
+String category
+verifyLicence()
+claimListing()
}
CanonicalRecord <|-- CoreLayer
CanonicalRecord <|-- AuthorityLayer
CanonicalRecord <|-- NicheLayer
Core, authority, niche, verification, audit, sync layers
Here is what each layer does.
Core: the platforms that feed everything else. Google Business Profile, Bing Places, Apple Business Connect, Facebook. If these are wrong, nothing downstream can save you.
Authority: the data aggregators and broad horizontal directories that other services consume. Data Axle (formerly InfoGroup), Foursquare, Localeze, Yelp. These are where bad data multiplies if you do not control them.
Niche: industry-specific platforms that matter for your category. Healthgrades for a doctor, Avvo for a lawyer, TripAdvisor for a restaurant. These often outrank horizontal directories for category-specific queries.
Verification: the mechanics by which each platform confirms you are who you say you are. Postcard, phone call, video, document upload. Each platform has its own and they change.
Audit: the recurring discovery process to find listings you did not create, listings you forgot, and listings that platforms overwrote with stale data from somewhere else.
Sync: the propagation step. Pushing canonical data outward through aggregators and managed services, then confirming that the push actually landed.
How CANVAS differs from tier-based models
Most citation frameworks I have seen use a tier-based model: tier 1 directories, tier 2 directories, tier 3 directories, ranked by domain authority or perceived importance. The problem with tiers is they describe the directory, not your relationship with it. A tier 1 directory you have no control over is a worse asset than a tier 2 directory whose data you can update in real time.
CANVAS describes functions, not rankings. Yelp is both Authority (because it feeds other systems and dominates branded searches) and Verification (because users treat it as a trust check). A healthcare directory might be Niche and also part of your Verification story if it requires licence number confirmation. The same platform can sit in two layers, and that is fine.
When this framework applies (and when it doesn't)
CANVAS is built for businesses with a defined service area and a category that maps cleanly to existing directories. It works well for trades, professional services, healthcare, hospitality, retail with physical locations, and most local service businesses.
It is less useful for purely online businesses (citations are not really the right signal there), for businesses operating in jurisdictions where the major directories do not have meaningful presence, and for businesses whose category is so new that niche directories do not yet exist. If you sell drone insurance to film productions, the Niche layer is going to be very thin and you should weight your effort toward Core and Authority instead.
Building the core and authority layers
The Core and Authority layers are where about 70% of the ranking benefit lives, in my experience. They are also where most of the consistency damage happens, because they are the most actively scraped and resold.
xychart-beta title "Austin dental practice local pack position after citation fixes" x-axis [Month0, Month1, Month2, Month3, Month4] y-axis "Local pack position (lower is better)" 0 --> 9 line [8, 6, 5, 3, 2]
Identifying your non-negotiable platforms
The non-negotiables for a US-based local business are Google Business Profile, Bing Places, Apple Business Connect, Facebook, Data Axle, Foursquare, Yelp, and Yellow Pages. For a UK business, swap in Royal Mail's PAF-aligned address as a canonical reference, plus Yell, Thomson Local, and 192.com alongside the global platforms. For Australia, add True Local and Hotfrog. The list is regional and you should resist the temptation to copy a US list into a UK strategy.
The test for "non-negotiable" is straightforward: does the platform either (a) display directly in search results, (b) feed data to other platforms that do, or (c) get queried by users as a trust check? If yes to any of those, it goes in Core or Authority.
Google, Bing, Apple, and the big four
Google Business Profile is the canonical record for most local searches. Treat it as the source of truth and make every other platform match it, not the other way around. Bing Places now has a useful import-from-Google feature, but I would still verify each field after import; the import is not always clean for suite numbers and special characters.
Apple Business Connect matters more than people realise because of CarPlay, Siri, and Apple Maps integration into iOS. If you have ever asked Siri for a nearby plumber, you have used Apple's local index. It is smaller than Google's but the users have higher purchase intent on average, and there are far fewer competitors fighting to be in there.
Facebook is the one I am most cynical about. Its local search functionality has decayed. But the page still gets indexed, still surfaces in branded searches, and is still a place people check to see if you exist. Keep it accurate; do not invest creative energy in it.
Myth: If my Google Business Profile is correct, the other platforms will eventually catch up. Reality: Aggregators pull from multiple sources, including public records, old domain registrations, and user submissions. They do not treat Google as authoritative. You have to push corrections into each major platform directly, or use a managed sync service.
Worked example: a dental practice in Austin
Let me walk through a real shape of a project (details slightly anonymised). A two-dentist practice in Austin, Texas, opened in 2014, moved suites in the same building in 2019, and added a partner in 2022 which prompted a slight name change.
When I ran the audit in early 2023, I found 47 inconsistent listings. The name appeared as "Austin Family Dental", "Austin Family Dental PLLC", "Dr Garcia Family Dental", and on one ancient health directory simply as "Garcia, J DDS". The suite number appeared as "Suite 210", "Ste 210", "#210", and on Foursquare it was missing entirely. The phone number was correct on Google, but two directories still listed a Vonage number the practice had dropped in 2017.
The fix sequence was: define the canonical record (more on that below), correct Google first, correct the three other Core platforms within the same week, then move to Authority. The aggregators took six to nine weeks to propagate corrections. Two listings on small directories never got corrected and we eventually got them removed because correction requests went unanswered.
Ranking impact: the practice moved from positions 7 to 9 in the local pack for "dentist near me" type queries to consistent positions 2 to 4 over about four months. I would love to claim that was all citations, but they also added 23 reviews in that window. Both contributed. Citations did the foundational work; reviews provided the competitive edge.
Niche directories and verification mechanics
The Niche layer is where you separate from competitors who stopped at Core and Authority. It is also where the rules get weird, because industry-specific directories often have stricter verification requirements and idiosyncratic data formats.
block-beta columns 3 A["Postcard 5-14 days"] B["Phone instant"] C["Video walkthrough"] D["Document: utility bill, lease, licence"]:3 E["NPI / licence"] F["Bar admission"] G["Proof of insurance"]
Industry-specific submission criteria
Healthcare directories typically require NPI numbers, state licence verification, and sometimes a manual review by directory staff. Legal directories like Avvo and Justia want bar admission details. Trade directories sometimes want proof of insurance or licensing. This is good news, not bad: the friction filters out competitors who cannot be bothered, which means the citation is worth more when you earn it.
For a contractor, the Niche layer might include Angi, HomeAdvisor, Houzz, Porch, BuildZoom, and Thumbtack. For a restaurant: OpenTable, Resy, Zomato, TripAdvisor, and Yelp's restaurant-specific features. For a vet: PetMD's directory, the AVMA, and local pet rescue networks that often run informal listing pages.
The criterion I use to decide whether a niche directory is worth pursuing: does it show up on page one of Google for at least three competitor names in my client's category? If yes, it is influencing branded searches and should be claimed. If no, skip it.
Phone, address, and entity verification methods
Verification methods vary, and the variation matters because some methods cannot be redone quickly if you move. Postcard verification (still common on Google for certain categories) takes 5 to 14 days. Phone verification is instant but requires the phone number listed to be reachable, which is a problem if you used a tracking number that has since been rerouted. Video verification (now common on Google) requires you to walk around the location with a phone camera proving the business exists; this is genuinely useful for ruling out fake competitors but mildly humiliating to perform.
Document verification, used by many niche directories, typically wants a utility bill, lease, or business licence with the address you submitted. If your business licence still shows an old address, this becomes a chicken-and-egg problem. Update the licence first.
Quick tip: Before submitting anywhere, photograph your business licence, lease or rates bill, and most recent utility bill. Store them in a single folder with the canonical address visible. You will need them across multiple verification flows and gathering them once saves hours later.
Worked example: a multi-location HVAC company
Multi-location is where verification mechanics get interesting. I worked with an HVAC company running four locations across two states. Each location had its own phone number, address, and Google Business Profile. The mistake the previous SEO had made was treating them as variations of the parent brand, which meant Google sometimes merged the listings or treated location B as a duplicate of location A.
The CANVAS approach: each location got its own canonical record. Same brand name, identical formatting of brand portion, distinct address and phone. The parent website used location-specific landing pages (one per city) with embedded maps pointing to the matching GBP. Schema markup on each landing page used LocalBusiness with the corresponding location's NAP, not the parent organisation's address.
{ "@context": "https://schema.org", "@type": "HVACBusiness", "name": "Acme Heating and Air", "address": { "@type": "PostalAddress", "streetAddress": "1420 W Maple St, Suite 3", "addressLocality": "Springfield", "addressRegion": "MO", "postalCode": "65802" }, "telephone": "+1-417-555-0142", "url": "https://acme-hvac.example/springfield"
}One thing I got wrong on this project, and want to flag because it is the kind of mistake nobody admits: I initially recommended consolidating two locations under one phone number with extension routing, because the client found it easier to manage. Six weeks later, the GBP for the smaller location lost prominence because the phone number was no longer unique to that address in the aggregator data. We reverted. Lesson: phone uniqueness per location is not negotiable, even if the client's call routing setup wants to disagree.
Myth: Multi-location businesses should use a single national phone number with location-based routing. Reality: For local search, each location needs a phone number that resolves to that location's data. Routing technology may make consolidation feel efficient, but the algorithms use phone-address association as a corroborating signal and shared numbers dilute it.
NAP consistency rules under CANVAS
NAP stands for Name, Address, Phone. That is the minimum. In practice, you also need to standardise the website URL, hours, primary category, and business description. TrueFuture Media's guide calls the expanded version NAP+W (with the W being the website URL) and proposes a seven-field standardisation list: business name, street address, suite/unit, city, state, ZIP, and primary phone. I use a slightly extended ten-field version that adds website, primary category, and hours.
mindmap
root((NAP consistency under CANVAS))
Name
Trading name not legal entity
DBA handling
No token variations
Address
USPS abbreviation
Suite vs Ste vs hash
No trailing period
Phone
Unique per location
One format only
No tracking number drift
Website plus
NAP+W canonical URL
Primary category
Hours
Defining your canonical record
The canonical record is a single document, owned by the business, that states the exact format of every NAP field. Not roughly. Exactly. Down to whether the suite number is preceded by "Suite", "Ste", "#", or nothing. Down to whether the state is "NJ" or "New Jersey". Down to whether the phone uses dashes, dots, or spaces.
Here is a template I hand to clients on day one:
| Field | Canonical value | Acceptable alternative | Forbidden variation | Reason |
|---|---|---|---|---|
| Business name | Acme Heating and Air | Acme Heating & Air (where ampersand is forced) | Acme HVAC, Acme H&A, Acme Inc. | Algorithms treat name token differences as separate entities |
| Street address | 1420 W Maple St | 1420 West Maple Street (only where directory disallows abbreviations) | 1420 W. Maple St., 1420 Maple W | USPS-style abbreviation, no trailing period |
| Suite | Suite 3 | Ste 3 (where character limit forces it) | #3, Unit 3, 1420-3 | Suite/Ste/# are read as distinct strings |
| State | MO | Missouri (only on directories requiring full name) | Mo., Mo, missouri | Pick one casing and stick with it |
| Phone | (417) 555-0142 | +1 417-555-0142 (international fields) | 417.555.0142, 4175550142 | Format is parsed; mismatched formats can create duplicate records |
The "acceptable alternative" column is the practical concession. Some directories truncate at 60 characters. Some force ampersands. Some require full state names. The canonical record should anticipate these and tell the person doing the submission which alternative is allowed and which is not.
Handling suite numbers, abbreviations, and DBAs
Suite numbers are where I have seen the most damage done quietly. Google Business Profile treats "Suite 210" and "Ste 210" as different strings for the purposes of internal entity matching, even though humans read them identically. The Jasmine Directory analysis on why NAP consistency matters puts this bluntly: the algorithms are literal. They do not know "St." is short for "Street". They know that one string contains a period and the other does not.
For DBAs (doing business as) and legal entity names, the rule is: use the customer-facing trading name in citations, not the legal entity. If your invoices say "Acme Heating LLC" but customers know you as "Acme Heating and Air", the citation goes in the trading name. The legal name lives on your About page and in your invoicing, not in your directory listings. Mixing them is the single most common cause of split-name citation pollution I see.
Did you know? In one audited dental practice the same business name appeared in four variations across major platforms: "Smith and Sons", "Smith & Sons Ltd", "Smith & Son's", and simply "Smith". Each variation was indexed separately. The practice's local pack visibility recovered only after all four were resolved to a single canonical form. Source: Jasmine Directory citation audit case study.
Resolving conflicts across legacy listings
Legacy listings are the citations you did not create, mostly cannot edit, and definitely cannot ignore. They typically come from three sources: aggregator data flowing from old public records, automated scraping of your website during a previous design, and submissions made by previous owners or marketing agencies who are no longer reachable.
The resolution sequence I use:
- Identify the listing using a discovery tool. Whitespark's Local Citation Finder is the standard option here; for quick checks I also use the free NAP Hunter Chrome extension to compare what appears on each platform side by side.
- Attempt to claim the listing via the platform's standard ownership flow.
- If claim is rejected (common when the previous owner is unreachable), submit a correction request with documentation.
- If correction is ignored after 30 days, submit a removal request citing duplicate or out-of-date data.
- If none of the above work, file with the data aggregator upstream. Often the listing is being regenerated from Data Axle or Foursquare data; correcting at source eventually starves the downstream listing.
This is slow, unglamorous work. There is no shortcut. A managed service like Yext, Moz Local, or BrightLocal can compress the timeline by pushing into many platforms at once, but the legacy listings that pre-date your management of the brand will still need manual handling.
Complete walkthrough: relocating a service business
Citation work gets stress-tested when a business moves. I want to walk through a full project to show CANVAS in motion. Names are changed; the structure is real.
pie title Citation profile cleanup outcomes, 47 listings (cleaning service relocation) "Already matching NAP" : 41 "Corrected (Yell, suite, phone)" : 3 "Removed, inaccessible" : 2 "Persists with old data" : 1
Client: a residential cleaning service in suburban Manchester, UK. Operating since 2016. Moving from a leased office to a new building 1.4 miles away in the same postal town. Total citations identified in the audit: 47. Time from move announcement to full citation consistency: 11 weeks.
Pre-move citation audit
Six weeks before the move, I ran a full discovery sweep. The starting state was actually quite clean by usual standards: 41 of 47 citations had matching NAP. The six exceptions were a Yell listing with an old phone number from before a 2020 system change, two industry directories with abbreviated suite information, and three listings on small directories that I could not access because the original submission email no longer existed.
I built the canonical record for the new address, including standardised formatting for the street name (which contained a saint's name, frequently rendered as both "St" and "Saint" by different directories). I confirmed Royal Mail's official address format for the new building and used that as the master.
Sequenced update rollout across 47 directories
The rollout sequence was deliberate. Week of move minus two: update the website with the new address, including LocalBusiness schema, footer, contact page, and any embedded maps. Update Google Business Profile, but use Google's "move" function rather than editing the address field, because the move function preserves review history and signal continuity. Bing Places and Apple Business Connect updated the same day.
Week of move: update Facebook, Yell, and the three largest niche directories. Submit data correction requests to Data Axle UK and Foursquare. These take time to propagate but pushing early means downstream listings start updating in parallel with the move itself.
Week of move plus one to four: manual correction of the remaining 30+ listings. This is grunt work, done in batches of 5 to 8 per day, with screenshots logged for each to confirm successful update. Two listings required emailing the directory's support; one required a phone call.
Week of move plus four to eight: audit propagation. Run discovery sweeps weekly. Identify any listings that have not updated or that aggregator-pushed corrections did not reach. Manually intervene on the laggards. Two listings reverted to old data twice during this window because the directories were pulling from cached aggregator data; we pushed the correction again until it stuck.
Week of move plus eight to eleven: final cleanup. The three listings I could not access during the audit had to be handled via removal requests rather than corrections. Two were removed; one persists to this day with old data and is genuinely beyond reach.
Post-move ranking recovery timeline
Rankings dipped predictably in week two and three after the move, recovering to the prior baseline around week six and exceeding it by week ten. The temporary dip was not caused by citation inconsistency (we had been disciplined); it was caused by Google reassessing proximity to user queries after the GBP address change. That is unavoidable. What we avoided was the longer-tail dip that comes from a messy citation profile, which I have seen extend for months on relocations done without an audit.
What if... you only have two weeks between learning about the move and actually moving? The honest answer: skip the niche directories on the first pass. Update website, GBP, Bing, Apple, Facebook, and submit aggregator corrections within those two weeks. Accept a longer ranking dip. Handle niche directories in weeks 3 to 8 after the move. The Core and Authority layers carry most of the load; the Niche layer can lag without catastrophic damage as long as the data eventually converges.
Edge cases and honest limitations
I want to spend the last section on where CANVAS gets uncomfortable, because every framework has edges, and pretending otherwise is how practitioners end up promising things they cannot deliver.
Service-area businesses without storefronts
Service-area businesses (plumbers, mobile pet groomers, locksmiths) have an awkward relationship with citations. Google Business Profile lets you hide your address if you serve customers at their locations, but many directories were built for businesses with public-facing premises and do not have a good way to handle "address present for verification, not displayed publicly".
The practical workaround: use the operator's home or office address as the canonical record (with consent and proper documentation), hide it on platforms that allow hiding, and accept that on platforms that do not allow hiding, your address will be visible. For locksmiths and security businesses, this matters; for most others, it is a minor privacy compromise. Do not invent an address. Do not use a UPS Store or virtual office; Google's spam team is reasonably good at catching these and the penalty when they do is severe.
The other workaround, which I see suggested often and consider mostly wishful thinking: registering at a coworking space and using that as your address. It works until Google's verification team realises forty businesses share the same address, at which point all forty start to suffer. I would avoid it.
Mergers, rebrands, and franchise structures
Rebrands are the hardest citation problem. When a business changes its trading name, every existing citation now contains stale data. The instinct is to update everything to the new name immediately. The better approach is to use the rebrand as an opportunity to claim listings you had abandoned, while preserving the old name as a known alias on platforms that support it (Google does, via the "from the business" description and historical context; most directories do not).
For franchise structures, the rule is: each franchise location is a separate citation entity. They share branding but not NAP. The corporate parent has its own canonical record, often with a registered office address that does not appear in local citations at all. Confusion happens when franchisees use corporate marketing materials that include corporate's phone number; the corporate number then leaks into local citations and you end up with one phone number associated with 400 addresses, which the algorithms hate.
Myth: After a rebrand, you should ask directories to delete old listings to avoid confusion. Reality: Updating an existing listing preserves its age, review history, and any backlink equity. Deleting and recreating discards all of that. Update in place wherever possible; only request deletion when the listing genuinely cannot be updated, such as orphaned listings tied to email accounts you no longer control.
Where CANVAS breaks down
CANVAS assumes you can identify a meaningful set of niche directories. In some categories, you cannot. Brand new categories (specialised AI services, niche subscription products) lack established directory ecosystems. In those cases the framework collapses to Core, Authority, and a smaller Audit step.
CANVAS also assumes the platforms cooperate. They do not always. Some platforms (and I will name names: Foursquare has historically been worst at this) overwrite submitted data with their own preferred format. You submit "Suite 3" and a week later it reads "Ste. 3" because that is the format their cleaner script enforces. You correct it; it reverts. Eventually you give up and align your canonical record to whatever the largest immovable platform insists on, which is a small surrender but a real one.
Citation work also has a ceiling. Honest practitioners acknowledge that consistency alone cannot rescue a business with weak reviews, an underpowered website, or a category in which it simply does not compete. I have had clients arrive expecting citations to undo years of being outranked by a competitor with better content and a stronger backlink profile. Citations do not do that work. They make sure you are in the running, not that you win.
One resource I keep returning to when explaining all this to clients is the curated set of business directories listed in the Jasmine Directory, which is a reasonable starting point for the Authority layer in particular. The point is not that any single directory matters enormously on its own; it is that a maintained, accurate presence across a chosen set is what builds the entity confidence search engines look for.
If you take one thing from this article, take this: stop counting citations and start curating them. Sixty accurate citations beat 200 messy ones, every time. Pick your Core, lock down your Authority, choose your Niche carefully, treat Verification as a privilege not a chore, run Audits quarterly, and Sync your canonical record outward whenever it changes. The work is dull. The results are not.
Block out two hours next Tuesday morning, before you do anything else, to run a Whitespark audit on your own primary location. Whatever you find will be worse than you expect. Fix the worst three discrepancies that week. Then put a recurring calendar entry for the same exercise in 90 days. That is how this work actually gets done, not in a campaign, but as a habit.
