Measure the change, not the number of automations.

Automation ROI gets miscounted more often than it gets calculated wrong. The usual mistake isn't bad math, it's measuring the wrong things, or measuring them without a baseline to compare against.

Five dimensions carry the real return

• Time: cycle time, waiting time, and manual effort.
• Quality: error rates, rework, completeness, and consistency.
• Risk: missed controls, unauthorised actions, and exception handling.
• Adoption: usage, completion, and successful handoffs.
• Business outcome: conversion, capacity, revenue, cost, or service level.

A baseline separates real change from launch noise

Without a record of the current process over a representative period, it's difficult to tell a real improvement apart from a temporary launch effect or an unrelated change in workload. The baseline is what the after-numbers get compared against, and it has to exist before the change does, not be reconstructed from memory afterward.

The full cost includes what happens after go-live

Implementation is only one line of the cost. Licences, integrations, maintenance, monitoring, training, and the time spent reviewing exceptions all belong in the same total. A workflow that saves time but creates expensive operational risk isn't delivering a complete return, even when the launch metrics looked good.

The real picture emerges after adoption settles

Early numbers reflect a workflow still being learned, not the workflow as it will actually run day to day. The results that matter are the ones measured again once adoption has settled, reviewed with the people who operate it, and used to adjust the flow based on the exceptions that actually occurred rather than the ones assumed during design.