How to build a local business directory that keeps itself up to date
A directory goes stale the month it is published. Here is how to start one from places data, add openings each week and flag closures without anyone rechecking listings.
Part of the guide Location intelligence automation
To build a local business directory that stays accurate, start it from a places database rather than from a one-off survey, then run a weekly job that adds the businesses it has not seen before and flags the ones that closed or changed their name. The first list takes an afternoon. The hard part, keeping it true a year later, is what most directories get wrong, and it is the part you can hand to an AI employee.
This guide is for the people who end up owning a directory: a chamber of commerce listing its members' neighbours, a tourism office mapping the shops of an old town, a charity keeping a referral list of local services, a property manager telling tenants what is nearby, or a media site with a "where to eat" section. It covers what to put in a listing, how to fill it, how to keep it current, and the traps that make readers stop trusting a directory.
Decide what a listing holds
Before collecting anything, write down the columns every listing needs. Readers expect the basics; your own purpose adds the rest.
- The basics: name, kind of business, address, phone, website, and a position so it can sit on a map.
- Signals of quality: rating and number of reviews, and whether it is an independent or part of a chain. Many local directories exist to favour independents, so this column matters.
- Freshness: the date you first listed it and the date it was last confirmed open. Without these, nobody can tell a live listing from a ghost.
- Your own fields: a member flag, a short description, opening hours you checked by phone, a category of your own.
Keep your own fields in separate columns from the ones a flow fills, so an automatic update never overwrites a description someone wrote by hand.
Fill the first version
Pick the area and the kinds of business, then pull the list in one go. In xMap Flows, an Area of interest holds the town, district or a shape you draw, and Nearby places searches xMap's places database for up to 12 kinds of place at once, each place with its type, rating, reviews, chain or independent, phone, website, address and position. For a directory of a few streets, a drawn shape is better than a city name: it keeps out the retail park on the ring road.
Send the result to a sheet in Google Sheets or, if the directory runs on your own website's database, to a table there with Save to my database, which can update a row when the same place comes back instead of adding it twice. Then spend the afternoon on the part a machine should not do: reading the list, removing anything that does not belong, and writing the short descriptions.
Keep it current without rechecking every listing
This is where a directory lives or dies. The trick is to search the same area every week and compare with what you already know, so only the differences reach a person. Two blocks do the comparing:
- Only new lets through only the places the flow has not found before, each with first_seen, the day it first appeared. Those are your candidate new listings.
- What changed lets through only the places the flow already knows whose details changed: now marked closed, a new name, a new address, a new phone or website. Each comes with a change line in plain words and the values it had before, so the editor sees both the old name and the new one.
One weekly flow can feed both from the same search:
- Schedule, every Monday at 07:00 in your time zone.
- Area of interest with the area your directory covers.
- Nearby places with your kinds of business, and Only places that are open switched off, so places that closed still come back and can be compared.
- First path: a Filter keeping status Open, then Only new, then a Google Sheets step that adds each newcomer as a row marked "to check".
- Second path: What changed, set to Closed and Name or address, then an Email to the editor listing each change with its old and new values.
The editor's Monday job shrinks from "check 400 listings" to "look at six new rows and two closures". Ready-made versions of this idea include Farmers' markets and organic shops directory and A referral directory of local support services, and the wider pattern is covered in the workflow automation guide.
A made-up example: an old town directory in Cape Town
Say a neighbourhood association in Cape Town keeps a directory of independent cafes, bookshops, galleries and craft shops in a few dozen streets. It draws the streets as a shape, pulls the first list, and the volunteers remove a dozen places that are really offices. The weekly flow then runs on its own. In this invented first month it adds five places, two of them a new gallery and a cafe that opened in a former bank, and flags one bookshop that now shows as closed and one craft shop that changed its name after a new owner took over. A volunteer phones the bookshop, confirms it has gone, and removes the listing the same day.
Traps that cost readers' trust
- Publishing every new row unread. Places data includes home businesses, offices and the odd duplicate. Keep a person between the flow and the public list.
- Treating "not found this week" as closed. A place can drop out of a search for reasons that have nothing to do with closing. Rely on the closed status and a phone call, not on absence.
- Mixing fresh and live data without saying so. Nearby places reads a database refreshed every month. For a directory where a closure must show within days, re-check your current listings with Map business scan, which looks each place up live.
- Letting the area creep. A directory of a quarter that slowly grows into the whole city loses its point. Keep the drawn shape fixed and start a second directory if you need one.
- Duplicates from overlapping searches. If you search two overlapping areas, add Remove duplicates before the rows reach the sheet.
Questions
Where should the directory itself live?
Wherever your readers are. A spreadsheet in Google Sheets is enough to start and can feed a website; a site with its own database can receive the rows directly through Save to my database.
How often should it update?
Weekly suits most local directories. Daily makes sense only where openings are frequent, such as a busy dining district; monthly is fine for slower categories like dentists or lawyers.
Can it tell me when a business closes?
Yes, when the places data marks it closed. What changed reports the change with the date, and you confirm before removing the listing.
Will my own descriptions be overwritten?
Not if you keep them in their own columns. The flow only adds new rows and reports changes; it never edits the text you wrote.
Do I need to code?
No. The flow is built from blocks in the editor; a Custom code block exists if you want to reshape fields, but a directory rarely needs it.
Templates that do this
Read next
More on business and places data · Every article · Templates