Cohorts

Import your own identifiers from a file, an upload, or an audience. Conventions apply to every command here.

Create

Import your own identifiers as a reusable dataset. See Create Cohort.

Exactly one source: --file-uri, --upload-reference, or --audience-id. The whole body can come through --file instead.

The accepted --file-format and --identifier-type values are on Create Cohort.

A --file-uri ending .csv, .gz, or .parquet is imported as one file, and anything else as a folder, even when it names a single file. The match is case-sensitive, so Q3.CSV imports as a folder. The same rule applies to --upload-reference, through the name the file was uploaded under (the uploads put argument, or --filename on uploads reserve), cut to its first 100 characters. Rename a file to end .csv, .gz, or .parquet before uploading it, because a preview reads the upload whatever its name. No credentials travel with the request, so the location must already be readable by Intuizi. Your Account Manager can help set that up.

An audience source takes the audience’s own name and project, so --name and --project-id are rejected with it, as are the flags that describe a file import. A regular audience yields at most one cohort, and a Lookalike Model audience can yield several.

intuizi cohorts create --name "Q3 customers" \
  --file-uri s3://example-bucket/cohorts/q3.csv --file-format csv \
  --identifier-type hem_sha256 --identifier-column email_sha256

intuizi cohorts create --audience-id <id> --device-limit 1000

No flag writes the other limits. Capping an audience cohort by visit frequency, by distance or, for a Lookalike Model, by score range instead of by device count needs the whole body through --file, and so does Match Strictness (max_devices_per_ip) on an SCID file import. See Create from an audience for the audience body.

A create returns as soon as the import is queued, and no cohort command takes --wait. Re-run intuizi cohorts show <id> until the status reaches 4 Completed, as the create’s note on stderr says. The table shows the status name, and --json gives the id at .data.status.id. Status 5 Not Available means the import failed: stop polling and fix the cause before creating again. A cohort that failed before failures were reported as 5 can show Unknown instead, and has failed too. For a --file-uri source, check the file and run the create again. For an --upload-reference source, upload the file again for a new reference, because the failed create used the old one up. A regular audience keeps its failed cohort, so run intuizi cohorts delete <id> before creating from that audience again. A Lookalike Model audience needs no delete. See Cohort status scale.

Preview

Read the first rows of a staged file and report its columns, so --identifier-column can be chosen with confidence. See Preview Cohort File.

Exactly one of --file-uri or --upload-reference, unless the whole body comes through --file. A preview returns sample rows rather than a resource, so --quiet is rejected on it.

A preview reads the file at a --file-uri whatever its suffix, and falls back to a folder only when no file is there. The create goes by the suffix alone, so a single file needs one of the suffixes above to import as a file.

intuizi cohorts preview --file-uri s3://example-bucket/cohorts/q3.csv

List

Page through the cohorts in your account. See List Cohorts.

Show

Read one cohort by id, including its import state. See Get Cohort.

A cohort moves through 1 Uploading, 2 Initiating, 3 Processing, and 4 Completed, a scale of its own rather than the audience one. A failed import ends at 5 Not Available instead. Read .data.status.id from --json when scripting. See Status Codes.

Delete

Remove one cohort. This cannot be undone. The cohort’s stored data is deleted with it unless it was built from a regular audience that still exists, so a file-imported or Lookalike Model cohort loses its data. Once the data is gone, an audience whose only dataset is this cohort can no longer be activated, and a schedule whose audience uses the cohort fails from its next cycle. Audiences that combine the cohort with other datasets keep their built results. Delete or deactivate those schedules, and deliver those audiences, before you delete the cohort. See Delete Cohort.