§ Data Ops

The back-office motions that turn documents into records.

Three overlapping disciplines sit under “data ops.” The titles change by industry; the motions don’t.

The three motions

• Intake — a document arrives (email attachment, scanned form, inbox forward, upload to a shared folder), and someone has to turn it into a structured record in a system. This is the highest-volume motion and the one most commonly understaffed.
• Validation — once a record exists, is it complete? Are the fields consistent? Does the linked vendor actually exist in the vendor master? Does the date make sense? Validation runs against existing records, not new ones.
• Reconciliation — when two systems disagree (the invoice in AP doesn’t match the PO in procurement; the shipment in the warehouse doesn’t match the order in the CRM), someone has to figure out which one is right and fix the other.

Who owns what

In a small org, one person does all three. In a mid-sized org, a data-entry specialist handles intake, a document controller owns validation policy, and a data-quality analyst handles reconciliation. In a regulated industry, all three are formalized and audited.

Where the processes help

Intake is the easiest to partially automate — a repetitive, high-volume motion with clear inputs and outputs. Validation is the easiest to fully schedule — rules, run nightly, report in the morning. Reconciliation is the hardest to automate because it requires judgment, but you can at least produce the variance report automatically so the human starts with the right dataset.

See: hiring a data-entry specialist, hiring a document controller, hiring a data-quality analyst, and the stack-specific pages for Drive, Airtable, and Box.