Skip to content
What’s my data worth?

Dataset quality

What makes business data useful?

Useful records show complete work and can be checked. Explain where they came from, what they include, and what is missing.

Where records came from

Name the original system, owner, dates, and any changes made to the records.

  • Original system, owner, dates, and export method
  • Sharing permission and restrictions
  • Version, changes, exclusions, and related records

Request, work, and result

Keep each request with its actions and result. Explain the fields and separate what happened from what was suggested.

  • The request and useful details
  • Actions, decisions, handoffs, and rules
  • The result and evidence it was completed
  • Roles and order of events, where permitted

Copies and missing results

Count complete tasks as well as messages. Note repeated records and missing results.

  • Copies and closely related versions
  • Unfinished tasks and missing decisions
  • Automatic replies and repeated templates
  • How examples were chosen and what is missing

Clear review rules

Write down the checks. Use labels consistently and explain how uncertain cases are handled.

  • Missing fields, consistency, and accuracy
  • Which records pass, fail, or need review
  • How to handle disagreements and other valid answers
  • Who checked each version and what changed

Keep training and tests separate

Use separate records for learning and testing. Keep copies of the same task in one group.

  • What each group is for and how it was chosen
  • Checks for the same task in both groups
  • Test inputs kept separate from expected answers
  • A dated list of records in each group

Agree the format

Choose a format the buyer can use. Include field meanings and notes about changes.

  • JSONL: one structured record per line
  • CSV: a table of rows and columns
  • Field meanings, version notes, and handling instructions
  • Agreed transfer and approval steps

Agree what
the sample should show.

Choose a small sample. Agree who can see it, what they may do with it, and when access ends.

Check it against the buyer’s task and agreed rules. Sample access does not grant rights for every future use.

Read the evaluation guide

Sample checks

  • Does it show the task and needed details?
  • Are fields clear and the result supported?
  • Was restricted information left out?
  • Are missing details and limits noted?
  • Who approves it, and how are changes handled?

See a record structure.

Three invented records show a support case, a workflow, and an AI correction. The pack contains no real customer data.

Download example pack Read the format guide

JSONL records, field definitions, and a format guide. These examples show a structure, not available inventory or measured results.

Example record, shortened
{
  "synthetic": true,
  "record_type": "operational-workflows",
  "request": "Reschedule a shipment after a route closes.",
  "work": [
    "Check approved alternatives.",
    "Record the required approval."
  ],
  "result": {
    "status": "completed-for-example",
    "outcome": "A feasible route is approved."
  },
  "evaluation": {
    "criteria": [
      "Uses a permitted route.",
      "Records the approval."
    ],
    "acceptable_alternatives": [
      "A plan requiring approval before release."
    ]
  }
}
Download JSONL only

License your data. Get paid.

Approved buyers pay for records like these. Find out what your data could earn in a few minutes. It costs you nothing.

What’s my data worth?