Getting a clean file out of LeadForge.

Decide what to export before you click

Every export button writes what is currently on screen, which means your filters are the selection mechanism. Narrow by category, city, distance or the presence of a phone number or website, and the export follows. The one exception is Export selected, which writes every record you have ticked — including ones a filter is currently hiding. That is deliberate: a shortlist you built by hand should not silently lose rows because you changed a filter afterwards.

Five outputs are available from the results toolbar:

• Export all — every record matching the current filters.
• Export selected — only the records you ticked.
• Phone list — the filtered records that have a phone number.
• Email list — the filtered records that have at least one e-mail address.
• GeoJSON — the filtered records as geographic features.

All of them are produced in your browser and download immediately — nothing is uploaded, and no copy is kept anywhere else. The phone and e-mail lists use the same columns as the full export; they are convenience filters, not different formats, so you keep the address and category context instead of ending up with a bare column of numbers.

What the CSV contains

Every CSV export uses the same seventeen columns, in this order:

• business_name, category, city, country, address — the identity of the record.
• phone, email, website, opening_hours — the contact fields, where they exist.
• distance_km — straight-line distance from your search centre, to two decimal places.
• latitude, longitude, osm_url — the location, and a link back to the underlying map record.
• emails, socials — every address and social profile found, separated by a semicolon and a space, so a record can carry more than one.
• source, source_id — where the record came from and its identifier in that source.

The source column is worth understanding. Records found by searching are openstreetmap; ones you added by hand are manual; ones you brought in from a file are imported. Keeping that column, together with source_id and osm_url, is what lets you answer “where did this row come from?” six months later — and it is what makes the attribution obligations described in the data guide practical to meet.

Note that email and emails are not redundant. The first is the single address recorded in the map listing itself; the second is the full set, including anything contact discovery found on the business website.

Two CSV details that surprise people

Leading apostrophes. Any value beginning with =, +, - or @ is written with a leading apostrophe. This is a deliberate safety measure: spreadsheet applications interpret those characters as the start of a formula, and a maliciously crafted field in an untrusted list could otherwise execute when the file is opened. The most visible effect is on international phone numbers, which start with +. The apostrophe is a spreadsheet text-marker rather than part of the value, and it can be stripped in bulk with a find-and-replace once the file is open — do that after import into your spreadsheet, not by editing the raw file.

Everything is quoted. Fields are wrapped in double quotes and internal quotes are doubled, which is standard CSV escaping. Any tool that reads CSV properly will handle it. If a column looks wrong, the usual cause is opening the file by double-clicking rather than using your spreadsheet’s import dialogue and specifying UTF-8 and comma separation — worth doing deliberately if your list contains accented business names.

When to use GeoJSON instead

The GeoJSON export writes a standard feature collection: one point feature per business, coordinates in longitude-latitude order, with the record’s fields carried as feature properties. Use it when the question is geographic — territory planning, mapping coverage against your existing customers, or loading results into GIS software or another mapping tool. Use CSV when the question is about contacting people.

The property set mirrors the CSV closely, with one difference worth knowing: it carries the record’s internal identifier and omits the map link. If you need the link back to the map record, export the CSV as well.

Importing a list back in

Import CSV accepts a comma-separated file up to 5 MB, and matches columns by header name rather than position. It recognises business_name (or business name, or name), category (or business_type), city, country, address, phone (or phone_number), website (or url), email, and latitude/longitude — including the lat, lon and lng abbreviations. Header matching is case-insensitive, so a file exported from another tool often works without editing.

Three behaviours to plan around:

• A business name is required. Rows without one are skipped silently, so a row count that comes back lower than you expected usually means a blank or mislabelled name column.
• Coordinates matter. A row without usable latitude and longitude is imported, but it has no real position and will not sit anywhere meaningful on the map. If you want the map view to be useful, keep the coordinate columns.
• Import replaces the current result list and does not save anything by itself. Select the records you want and save them, exactly as you would after a search.

A round trip is not lossless. Re-importing a LeadForge CSV preserves the name, category, city, country, address, phone, website, e-mail and coordinates; opening hours, social links, distance and the original source identifiers are not read back, and every imported row is marked imported. Treat exports as your archive, not as a save format to work from.

Cleaning a list before it goes anywhere

The tool removes duplicates within search results — records that share a source identifier, or a name and address, or a name and website domain, are merged, with populated fields kept over empty ones. That does not extend to imported files, and it cannot catch everything. Once your file is in a spreadsheet, a short cleaning pass pays for itself:

• Sort by name and scan for near-duplicates — “The Corner Café” and “Corner Cafe” merge on the eye but not on a key.
• De-duplicate on website domain, which catches chains and franchises listed under slightly different names.
• Check for branch versus head office where several rows share a domain, and decide which one you actually want to contact.
• Normalise phone numbers into one format before a dialer or CRM sees them.
• Drop columns you will not use. A calling list does not need coordinates or social links.
• Delete rows that failed verification rather than leaving them in with a note. Notes get ignored during a mail merge.

Verify before a CRM import, not after

A CRM is the worst place to discover that a record was wrong, because bad rows outlive the campaign that created them: they get assigned, reported on, re-contacted, and merged into other records. Do the verification described in Responsible business outreach first, on the file.

When you do import, map the provenance columns to real fields rather than dropping them — a source field and a link back to the original record turn an unexplainable row into a traceable one. Import into a staging list or with a distinguishing tag on the first run, check a sample by hand, and only then merge into your main database. And decide in advance what happens when an imported business already exists in the CRM; letting the system create duplicates is how a clean list becomes a mess.

Look after the file

While records are in LeadForge they stay in your own browser. Once you export, that protection ends and the file is an ordinary document containing contact information about real businesses. Treat it accordingly: store it somewhere access-controlled rather than a shared drive the whole company can read, avoid mailing it around as an attachment, and do not upload it to third-party services casually — a “free enrichment” tool that takes your list is being paid in your list.

Delete exports when the work they were made for is finished. Old lead files accumulate quietly in download folders, decay into inaccuracy, and are exactly the sort of thing you do not want to still be holding without a reason. Inside the application, the same principle is one action: Clear my local data on the About page removes saved leads and search history from your browser.

Re-run rather than reuse

Map data changes: businesses open, close, move and rebrand. An export is a snapshot of a moment, not a maintained list, and its accuracy declines from the day it is created. For anything time-sensitive, re-run the search and compare against your previous file instead of trusting the old one — the difference between the two is often the most interesting part.