Choosing an export format and which columns to include
You do not choose a format when you export — every format your plan includes is written, and you choose at download. How to export a subset of the 64 columns, and why fewer columns does not mean fewer credits.
You are not choosing a format at export time
This is the part that catches people who are used to other tools. Every finished export is written in every format your plan includes, from the same rows. The choice happens at the moment you press Download, not at the moment you create the job — which means you can take all of them, and you never have to re-run an export because a colleague wanted a different one.
| Format | On | Pick it when |
|---|---|---|
| CSV | Every plan | It is going into a CRM, a mail tool, or anything scriptable. Read the format note below first — it is tab-delimited despite the extension. |
| XLSX | Every plan | A human is going to open it and work in it. Types and accents survive without any import ritual. |
| JSON | Growth and up | You are feeding it to code and want the JSON-valued columns already parsed. |
| Parquet | Scale and Enterprise | The row count is in the hundreds of thousands and it is going into a warehouse or a dataframe. |
Format is the one place where a tier changes the shape of what you get rather than the amount. It changes nothing about the rows themselves: the same 64 columns, the same values, on every plan.
If a colleague will open it in Excel, send them the XLSX
Our CSV is UTF-8 with a byte-order mark, which Excel handles correctly — but only on some import paths, and only in reasonably recent versions. XLSX removes the whole class of problem, and it costs nothing extra because the file already exists. If you must send CSV, point them at the accented-characters article.
Choosing columns
The export dialog lists 34 column groups with checkboxes, and by default every one is included. Unchecking any of them removes it, and the file is written with only the groups you kept, in the canonical order — the order never changes to match the order you ticked things.
34 groups is not 34 columns. Eight of them produce more than one: Address gives you eleven, All emails gives you nine, Contact pages six, Phone and Website meta three each, and all thirty-four together land at 64 columns. See the 64 columns explained.
Fewer columns does not cost fewer credits
A credit buys a business, not a field. Exporting three columns and exporting all sixty-four costs exactly the same. So the only reasons to trim are file size and the sanity of whoever opens it.
Sensible subsets
These are group names, which is what the checkboxes and the API both use:
- Cold email:
name,main_type,email,website,location_address. - Cold calling:
name,main_type,phone,location_address,working_hours. - Territory analysis:
name,main_type,location_coordinates,reviews,is_closed. - Anything you might have to justify later: keep
google_idandlink. Those two columns are your audit trail — the key the row is stored under and a URL to the public listing it was read from — and they cost nothing to carry. See your lawful basis.
Naming the export
Give it a real name in the dialog. "Dentists Oslo — has email, 4+ stars" is worth the four seconds; "Export 14" is not. Names have to be unique and at least three characters, and the name is what you will search when you need to re-download it months later.
File size, roughly
A full 64-column row varies more than you would expect. Measured across the
sample export, rows run from about 700 bytes
for a business with no website to over 11 kB for one with a long
All emails array. The JSON-valued columns — Working hours,
All emails, All contact pages, Website technologies and
the All … link arrays — are almost the whole of that spread, so an export's size
depends far more on how well enriched your market is than on how many rows you took.
If you want a small, predictable file, untick all_emails and the
All … link groups: the same businesses, a fraction of the bytes, and no change at
all to the credit cost. Large files are
streamed and delivered through a signed link rather than held in the browser —
what happens after you press Export.
Over the API
The export endpoint takes an optional exported_columns array, which must be a
subset of the 34 group names. Omit it and you get all of them. The download endpoint is where
you name the format —
exports over the API.
Updated on: