Import contacts from CSV

If you already keep customers in a spreadsheet, a booking system or another tool, you can bring them into DMLY as Contacts with a CSV file. The importer matches rows against people you already have, so running it twice does not give you two of everyone. Prefer to watch first? This walkthrough covers a whole import end to end.

Two things to know before you start, because they decide what actually lands:

Every row needs a phone number. Rows with a blank Name, a blank phone, or a phone DMLY can’t read (letters in it, or no digits) are skipped, and DMLY always tells you how many: you land back on Contacts with “N contacts imported. M row(s) skipped — missing a name or a phone number.” If every row is skipped, nothing is imported and you stay on the upload screen with an error rather than a green zero: “No contacts were imported — all N row(s) were missing a name or a phone number. Check the first row is a header containing name, country_code and phone columns (download the sample for the exact format).” A CSV of email-only contacts is exactly that case, so it fails loudly.You must have a channel selected. The importer attaches each contact to one channel. With no channel active, it stops with “Select a channel before importing.”

Import a list

Open the importer

Go to Contacts and select Import. The page is headed Import contacts and tells you where they are going: “Contacts will be imported into” your current channel.

The Contacts header with the Import button highlighted

If you don’t see Import, your role doesn’t include it. See Roles and permissions.

Start from the sample

Select Download sample to get salesbot-contacts-sample.csv. It has the exact header row the importer expects. Replace the two demo rows with your own data and keep the header as it is.

Upload the file

Drop your file on the upload area or select it. Only .csv files are accepted.Preview shows the first 10 rows of your file. It only renders the seven sample columns, so an email column is read on import but won’t show up here.

It is still your chance to catch a shifted column: if a name is sitting in the phone column here, it will import that way too.

Import

Submit the file. You land back on Contacts with a count of how many contacts were imported.

The columns

The header row is what matters: column order follows your header, and capitalisation and surrounding spaces are ignored. Anything not in this list is ignored.

Column	What it does
`name`	Required. A blank name skips the row.
`country_code`	Joined onto the front of `phone` to make the full number.
`phone`	Required. A blank phone skips the row.
`email`	Saved on the contact, and used for matching. Not in the sample file, but it works.
`tags`	One or more tag names, separated by `\|` or `,`. Tags that don’t exist yet are created.
`stage`	A pipeline stage, matched by its exact name.
`notes`	Saved as a note on the contact.
`custom_fields`	A JSON object of custom fields, for example `{"company":"Acme"}`.

The sample file looks like this:

name,country_code,phone,tags,stage,notes,custom_fields
Jane Doe,1,5550100,VIP|Buyer,Lead,Met at expo,{"company":"Acme"}
John Smith,44,7700900000,Lead,Subscriber,,

Don’t trust the sample’s second row as a model. It puts Subscriber in the stage column, but the stages a new Workspace starts with are Lead, Engaged and Customer, so that stage matches nothing and John Smith arrives with no stage at all. Use your own stage names, exactly as they appear on the Pipeline.

Quote any value that contains a comma. An unquoted comma shifts every value after it one column to the right and pushes the last columns off the end of the row. With the sample header, a note written as Met at expo, wants a quote imports Met at expo as the note, wants a quote into custom_fields where it fails to parse, and drops {"company":"Acme"} entirely. Write it as "Met at expo, wants a quote".

What happens to people you already have

Before creating anything, the importer looks for an existing contact anywhere in your Workspace with the same phone number or the same email address. If it finds exactly one:

No duplicate is created.
The import channel is added to that person as another channel identity.
Name, email, phone and stage are only filled in where they are empty. If the contact has a name and your CSV has a different one, the existing name stays.
Tags are replaced with whatever the tags column says, not added to. The import does not start any automations for tags or fields it changes (see the note below).
Custom fields are merged into what’s already there, and the CSV value wins on any key it sets. This is the one field the importer overwrites.
A notes value is added as a new note, on every run.

Rows that match nobody create a new contact, recorded with a source of CSV import. So a re-import doesn’t duplicate your list, but it isn’t read-only either. It tops up blank names, emails, phones and stages, and leaves everything else alone, while replacing tags, overwriting the custom fields your CSV sets, and appending the note again.

A blank or missing tags column strips tags. Tags are replaced, so a re-import whose tags column is empty, or that has no tags column at all, removes every existing tag from every contact it matches. If you are re-importing to update something else, carry your current tags in the file.

An import never triggers automations. New contacts don’t start “contact created” flows, and tags or custom fields the import adds, removes or changes don’t start tag or field automations either. The success message says so: “Automations that watch for new contacts were not run.” Outbound webhooks (contact.created, contact.tagged, contact.untagged) still fire.

Export on the Contacts page writes the same seven columns the importer reads, so you can export, add a column of tags in a spreadsheet, and import it back to apply them. Remember this won’t start any tag-based automations.

When rows don’t arrive

The importer tells you both numbers: “N contacts imported. M row(s) skipped — missing a name or a phone number.” Read that first: if the skipped count matches your own tally of empty cells, nothing else is wrong. If the numbers still don’t add up, work down this list.

Fewer contacts imported than rows in the file

Almost always a blank name or a blank phone: either one skips the row, and the message that comes back counts them for you. Sort your spreadsheet by each of those columns, fix the empties, and re-run.

Nothing imported, and you're still on the upload screen

You got “No contacts were imported — all N row(s) were missing a name or a phone number.” Every row looked blank to the importer, which is usually a header problem rather than empty data. The first row of the file is read as the header and only the eight column names in the table above are recognised, so a renamed column (Full Name instead of name, Mobile instead of phone) leaves every row with no name or no phone and skips the lot. Capitalisation and stray spaces are fine, and Excel’s “CSV UTF-8” byte-order mark is stripped for you, so it is the name itself to check. Select Download sample, copy its header row exactly, and re-run.

The import was blocked before it started, and nothing was imported

Your plan has a cap on how many contacts a Workspace can hold. The whole file is checked before a single row is written, so an import that would take you over the cap imports nothing at all; there is no partial result to clean up.The check is deliberately cautious: rows that would only have matched an existing contact still count toward the estimate, so a large re-import of people you already have can be blocked even though it wouldn’t add anyone. See Plans.

Contacts imported with no stage

The stage value must match a stage name on your Pipeline exactly. Anything else is ignored quietly. Open Pipeline stages and copy the names from there.

Phone numbers look wrong

DMLY joins country_code and phone with nothing in between. If your phone column already includes the country code, leave country_code empty.On a WhatsApp or SMS channel, the joined number is then tidied up: a leading + or 00 is removed, a number that already starts with your connected number’s country code is kept, and a local number (with any leading trunk 0 dropped) gets that country code added. If the number can’t be tidied (for example, it’s too long), DMLY stores its bare digits, so a country code entered twice still ends up in the number twice. On other channels the number is only lightly cleaned.A phone value containing any letter (such as ext 4) or no digits at all is refused. That row is skipped and counted as “missing a name or a phone number”, even though its phone cell isn’t blank.

Everything imported, but you can't message them

Imported contacts are attached to the channel that was selected at import time. Whether you can actually reach someone on WhatsApp depends on your connected number and WhatsApp’s messaging rules, not on the import. See WhatsApp limits.

Tags and segments

Group the contacts you just imported into a reusable audience.

Send a broadcast

Message a segment on WhatsApp.

Custom fields

Set up the fields your custom_fields column fills in.

Contacts overview

How one person can span several channels.

Getting started

Messaging

Business

Workspace

Compliance

Troubleshooting

Import contacts from CSV

Import a list

The columns

What happens to people you already have

When rows don’t arrive

Next

Tags and segments

Send a broadcast

Custom fields

Contacts overview

​Import a list

​The columns

​What happens to people you already have

​When rows don’t arrive

​Next

Tags and segments

Send a broadcast

Custom fields

Contacts overview

Import a list

The columns

What happens to people you already have

When rows don’t arrive

Next