What should you measure before automating a business process?

Before automating a process, measure four things for two weeks: how often the step runs, how long it takes from trigger to done, how often it fails or is redone, and how long the customer or colleague waits. Without a baseline you cannot claim a result — and a consultant who claims one without a baseline is guessing.

The four numbers

1. Frequency

Count the real number. Businesses consistently underestimate how often a step runs because it is spread across several people. Fifty quotes a day looks different from "a few".

2. Cycle time

From the trigger (enquiry received) to done (quote sent, order updated). Include the waiting, not just the working. Waiting is usually the larger number.

3. Error and rework rate

How often the output is wrong, corrected, or done twice. Retyping between a chat and a system is where most errors live.

4. Response delay

How long the person on the other end waits. For a sales enquiry in this market, hours cost deals.

How to measure without a project

A shared sheet, two weeks, one person responsible for filling it in. Timestamps, not opinions. If the process runs through WhatsApp and a system, the timestamps already exist — pull them.

What the baseline gives you

A ranking of what to automate first, a number to compare against afterwards, and protection from vendors. In my own company, the health-score work only mattered because we had measured activation and churn first; otherwise it would have been a dashboard nobody trusted.

If the data does not exist

Then establishing it is the first deliverable of any automation work, and claims of savings before that are not credible — including mine.

Next step

If this is the problem you are dealing with, see business process automation consulting, or read Revenue is not resilience.