This guide is the cleanup, not the shopping list. Which directories are worth claiming, and in what order, is already laid out in the article this pairs with, which has the tiers and the reasoning. No point repeating it here.
What that article doesn't do is walk you through finding the listings you didn't create, working out which ones are wrong, and getting them fixed in a way that sticks. That's this, and the last part matters more than it sounds, because the most common outcome of a listings cleanup is that some of it quietly undoes itself within a few months.
The worked example is your own business, since the output is a sheet about you. Set aside three hours for the first pass and expect the number afterwards to be closer to fifteen minutes.
Decide what correct looks like, before you search
You can't audit consistency until you have defined what you're being consistent with. Do this first and write it down, because otherwise you'll make judgement calls listing by listing and end up with three versions of correct.
Fill in the declaration block below and treat it as the source of truth. Three rules on what goes in it, from Google's guidelines:
- Name is your real-world business name, as used consistently on your storefront, your website and your stationery, and as customers know you. Nothing else. Adding a city or a service to the field isn't permitted and can get a profile suspended.
- Address is a precise, real location you actually operate from. A rented mailing address you don't work out of isn't eligible.
- Phone should be the single authoritative number for the business. One number, the one that rings.
Then decide the fiddly bits and commit: Street or St., Suite 4 or #4, whether the legal Ltd appears. There's no right answer. There's only the answer you use everywhere from now on.
Last, check your own website against the block you just wrote. Your footer and your contact page are citations too, and they're the two you fully control. If they disagree with each other, fix that before you go anywhere else.
How to know it worked One written block exists, and your Google Business Profile, your site footer and your contact page all match it exactly. If those three disagree, you aren't ready for step 2.
Common trap Picking a canonical name that isn't what your sign says, usually the legal entity name. Match the real-world name customers see, not the one on your incorporation paperwork.
Find the listings you never made
Here's the part people get wrong. They audit the listings they remember creating. The problem is almost always in the ones they didn't.
Directories buy and scrape business data constantly, so you have listings you never signed up for, built from records of varying age. That's where the disconnected phone numbers and the previous tenant's address live.
So don't search your business name and stop. Run the full set from the search block below, and pay particular attention to two of them:
- Your phone number in quotes. The single highest-yield search here. It finds listings that don't carry your name in a form you'd have thought to search for.
- Your old phone number and old address in quotes. If you have ever moved or changed providers, this is where the stale records surface. Nothing else finds them.
Every hit goes in the sheet as a row: platform, URL, and what the listing currently says for each of the three fields. Copy what's actually there, not what you expect. The whole audit depends on that column being honest.
| Search | Finds |
|---|---|
| Business name | The listings you already knew about |
| "Current phone number" | Listings under names you'd never search1 |
| "Old phone number" | Stale records still in circulation2 |
| "Street address" | Previous tenants, and you under a variant name |
| Name without the legal suffix | Older listings from before you incorporated |
| Name plus the word reviews | Aggregator pages you never created3 |
How to know it worked Your sheet contains at least one listing you didn't know existed. If every row is something you personally created, you haven't searched hard enough. Go back and run the phone number in quotes.
Common trap Searching while signed in and from your own premises, which personalises results heavily. Use a private window, and check page two.
Score every row against canonical
Now turn the list into a work queue. Go row by row and give each one a verdict against the block from step 1. Four possible outcomes, and no row leaves without one:
- Match. Everything agrees. Mark it and never think about it again.
- Mismatch. One or more fields are wrong. Note which field, because that drives the order in step 4.
- Duplicate. A second listing for you on a platform where you already have one. Step 5.
- Dead. A listing for a business you no longer are, or a directory that no longer functions. Decide whether it's worth the effort at all.
Add a claimed column while you're here: can you log into this, or is it a page that exists about you with no owner? That distinction determines whether fixing it takes two minutes or a support ticket.
Resist fixing anything yet. Auditing and fixing at the same time is how people spend three hours and correct four listings, having never discovered the one that mattered.
How to know it worked Every row has a verdict and a claimed flag, and you can sort the sheet to see your mismatches grouped by which field is wrong. That sorted view is your actual to-do list.
Common trap Marking a row as a match after checking only the name. Check all three fields on every row. Name is the field most likely to be right and least likely to matter.
Fix in damage order, not alphabetical order
Not every wrong field costs the same, so work the sheet by damage rather than top to bottom.
- Wrong phone number. Always first. A wrong number isn't a ranking problem, it's a customer who tried to call you and reached nobody, or worse, reached somebody else. This is the only category costing you money today.
- Wrong address. Second. Sends people to the wrong door and contradicts your strongest location signal.
- Wrong or missing website. Third. Costs you the click and the link.
- Name variants. Last. Genuinely worth fixing, genuinely not worth doing before the three above.
Fix each one, then write the date in the sheet. That date matters, because platforms update on their own schedules and some queue edits for review. You aren't finished when you click save, you're finished when you reload the public page and see the change.
Re-check each fix after a week. Put it in the sheet as a second date column. What you're watching for is anything that reverted, because a revert is the single most useful diagnostic in this whole process and step 6 is what to do about it.
How to know it worked Every mismatch has a fixed date, and a week later you have reloaded each public page and confirmed the correction is still there. Not the edit screen. The page a customer sees.
Common trap Starting with the platform you find easiest to log into. It's almost never the one doing the damage. Sort by wrong phone number and start there, even if it means a support ticket.
The free audit checks your listings across the web and flags the conflicts, which is a faster way to build the sheet in step 2.
RUN THE FREE AUDITDeal with the duplicates
Two listings for one business on one platform is worse than one listing with a typo. A typo is a weak signal. A duplicate is two competing claims about who you're and where.
Google's position is direct: there should only be one profile per business, because duplicates cause problems with how your information displays on Maps and Search.
Duplicates usually come from a predictable place. Somebody created a listing not knowing one already existed, often a new manager, a franchise head office, or an agency. Or you moved and a new listing was created at the new address while the old one carried on.
The order to handle them:
- Work out which one to keep. Usually the one with the reviews and the history, not the one with the correct address. You can fix an address. You can't move reviews to a different listing.
- Merge rather than delete where the platform offers it, since merging preserves what the loser accumulated.
- If you can only delete, correct the keeper first, then remove the other, then re-check in a week.
Expect this to be the slowest step. Duplicate removal often needs a support request and a wait, and it can't be rushed.
How to know it worked Search your business name in a private window and get one result per platform. The survivor is the one that kept the reviews.
Common trap Keeping the listing with the right details and deleting the one with fifty reviews. The details take a minute to correct. The reviews are gone permanently.
Work out which listings are fed by something else
This is the step that separates a cleanup that holds from one that unwinds, and it's the reason people do this work twice.
Many directory listings aren't typed in by anyone. They're generated from a data feed, and the directory is a display layer over a record held somewhere upstream. Edit the page you can see and you have changed the display. The source keeps publishing the original, and in a few weeks or months it overwrites your correction.
That's why a listings cleanup can look finished in March and be half wrong by June, with nobody having touched anything.
You identify these by their behaviour rather than by any label:
- The revert. Anything that went back to the old value after you fixed it is fed by something. This is why step 4 has a re-check column.
- No owner and no login. A page about your business with no way to claim it is almost always generated.
- The same wrong value in several places. If four unrelated directories all show the same disconnected number, they didn't each make the same mistake. They share a source.
That third pattern is the useful one, because it turns a list of symptoms into one cause. Fixing the upstream record fixes the whole cluster, and it's the part of this job most worth paying somebody for, since the data layer is tedious to reach by hand.
How to know it worked You can point at the rows in your sheet that reverted or share a wrong value, and say which cluster they belong to. Naming the cluster is the win here, even if fixing the source takes longer.
Common trap Re-fixing a reverted listing a second time and considering it handled. It will revert again on the same schedule. The revert was the diagnosis, not the problem.
Watch for Google changing your profile back
The same thing happens on Google, from a different direction, and it surprises people because they assume a verified profile is theirs to set.
It mostly is, but not entirely. Google's documentation says that if it receives reports your business information isn't accurate, it might update your profile, and that when it does you get a notification and find an alert in the Edit profile section.
So two habits, both cheap:
- Don't ignore Google's notification emails. They look like marketing and occasionally aren't.
- Open Edit profile during your quarterly check and look for the alert, rather than assuming no news means no change.
When a change is wrong, correct it and get on with your day. When the same change keeps coming back, that's a signal rather than a nuisance: something on the open web is telling Google the other version is right, and it's usually one of the feed-fed listings from step 6, or your own website contradicting your profile. Which is why step 1 started with the footer and the contact page.
How to know it worked You know where the alert appears in Edit profile, you have looked at it once, and your profile currently matches the canonical block.
Common trap Treating a Google-applied change as vandalism and repeatedly reverting it. Sometimes it is correcting a genuine contradiction you published somewhere else, and the fix is upstream rather than on the profile.
Set the cadence, because this drifts
The first pass is the expensive one. Keeping it right isn't, provided it actually happens.
Listings drift for reasons that have nothing to do with you. Data gets bought and resold, directories merge, feeds refresh from records you've never seen. Nothing notifies you, so it's a calendar item or it doesn't get done.
Quarterly is right. Fifteen minutes, and only three things:
- Re-run the phone number search in quotes. New listings appear on their own.
- Spot-check the rows you fixed last time, especially anything in a step 6 cluster.
- Open Google's Edit profile and check for an alert.
Add a row to the sheet each time with the date and what you found. After a year that column tells you which platforms are stable and which need watching, and you can stop checking the stable ones.
How to know it worked A recurring quarterly reminder exists with a link to the sheet, and the sheet has a dated check log. A reminder with no link to the sheet gets dismissed.
Common trap Treating this as a project that finished. It's a subscription. The businesses whose listings are right are the ones that look four times a year, not the ones that did a heroic cleanup once.
Copy and paste
Three artifacts. The sheet is the one that matters, and the search block is what fills it.
CANONICAL NAP ....... decided 2026-__-__
This is the only correct version. Everything gets
matched against this, including my own website.
NAME ......... [real-world name, as on the sign]
no city, no services, no keywords
ADDRESS
Street ..... [ ]
Unit ....... [ ] format: Suite 4 / #4 -> pick one
City ....... [ ]
Province ... [ ]
Postal ..... [ ]
Street/St. ..... pick one: ______
Ltd / Inc ...... include? yes / no
PHONE ........ [ ] the one that rings
OLD PHONE .... [ ] for searching only
OLD ADDRESS .. [ ] for searching only
WEBSITE ...... https://
CHECKED AGAINST
[ ] Google Business Profile
[ ] website footer
[ ] contact page
These three must agree before step 2.Private window. Not signed in. Check page two.
1. Business Name
2. Business Name + city
3. "555-555-5555" <- current phone, in quotes
HIGHEST YIELD SEARCH
4. "555-111-2222" <- OLD phone, in quotes
5. "123 Example Street" <- address, in quotes
6. Business Name without Ltd / Inc
7. Business Name + reviews
Also check directly, since search does not always surface them:
Apple Maps Bing Maps Facebook Yelp
Every hit becomes a row. Record what the listing
ACTUALLY says, not what it should say.PLATFORM | URL | NAME SHOWN | ADDRESS SHOWN | PHONE SHOWN |
VERDICT | WRONG FIELD | CLAIMED | FIXED | RECHECKED | NOTES
VERDICT ...... match / mismatch / duplicate / dead
WRONG FIELD .. phone / address / website / name
CLAIMED ...... yes / no / no login offered
FIXED ........ date you made the correction
RECHECKED .... date you reloaded the PUBLIC page
a week later. Blank = not done.
FIX ORDER (sort by WRONG FIELD, work down)
1. phone .... costs you customers today
2. address .. sends people to the wrong door
3. website .. costs the click and the link
4. name ..... real, but last
REVERT WATCH
If RECHECKED shows the old value again, that row is
fed from somewhere upstream. Do not just re-fix it.
Look for other rows with the SAME wrong value: that
cluster shares one source, and the source is the fix.
QUARTERLY LOG
DATE | NEW LISTINGS FOUND | REVERTS | GOOGLE ALERT?The whole cleanup on one page
Where this stops
The first pass is completely doable and mostly just tedious. Three hours, a spreadsheet, and a tolerance for support forms. Nobody needs to be paid for the searching and the typing, and anyone telling you otherwise is selling the easy half.
Two things genuinely break DIY here, and neither is difficulty. The first is the data layer in step 6: when a cluster of listings all regenerate from one upstream record, reaching that record isn't something you can reliably do by hand, and that's the honest case for paying somebody. The second is that this never finishes. Listings drift whether or not you're paying attention, so the value isn't in the heroic cleanup, it's in the fifteen minutes every quarter forever. Most owners do the first and not the second, which is why the cleanup gets done twice.
If you'd rather it stayed right without the calendar entry, that is what the local SEO work covers. If you just want to know how bad it currently is, the free audit checks your presence across the web and flags the conflicts, which is a much faster way to build the step 2 sheet than searching by hand.
Questions people ask
Why did my corrected listing go back to the old information?
Because something else is still publishing the old version. Many directory listings aren't typed in by a person, they're generated from a data feed, so editing the page you can see changes the copy while the source keeps republishing the original. The same thing happens on Google from a different direction: Google says that if it receives reports your business information isn't accurate it might update your profile, and you get a notification and an alert in the Edit profile section when it does. A revert is therefore information rather than a failure. It tells you that listing is downstream of something you haven't found yet.
Does Street versus St. actually matter?
Less than most advice implies, and it's worth being honest about that. A search engine reconciling two records is weighing everything at once, and an abbreviation is a far weaker signal than a phone number that rings somewhere else. Pick one format, use it everywhere from now on, and correct the variants when you're already in the listing for another reason. What you shouldn't do is spend an afternoon normalising abbreviations while a disconnected phone number sits on a directory that ranks for your business name. Fix the wrong data first, tidy the formatting second.
There's a phone number on my listing that I've never used. Where did it come from?
Usually one of three places. It's a tracking number a former marketing agency or directory sales rep added so calls could be counted, and it stayed after they left. It's a genuinely old number from a previous owner or a previous premises that a data feed still holds. Or it is a call-forwarding number a lead generation site attached so it could resell your calls back to you. All three are worth removing, and the last one is worth removing urgently, because a number you don't control is a customer relationship you don't own.
What should I do if my business has moved address?
Update the existing listings rather than creating new ones, everywhere you possibly can. The instinct is to start fresh at the new address, but a new listing begins with none of the history the old one accumulated, and the old one carries on existing and contradicting it. Update in place, starting with Google, then work outward through the sheet. Expect the old address to keep resurfacing for months from feeds that haven't caught up, which is exactly the aggregator problem in step 6, and budget a re-check at 30, 60 and 90 days rather than assuming one pass finished it.
Should I pay a citation building service?
For cleanup and aggregator submissions, sometimes yes. For bulk listing creation, no. The genuine value in a paid service is reach into the data layer underneath the directories, which is tedious to work by hand and is where reverts come from. What isn't worth paying for is a package promising hundreds of new listings, because volume was never the thing that helped and a hundred fresh listings built from one wrong record just multiplies the error. If you buy anything, buy the cleanup and the aggregator submissions, and check the list of sites it covers before you pay.
