If you don't measure the starting point, you can't show a return. Do this before you build anything.
Step 1: Record today's numbers. For one workflow, over a normal month:
| Measure | Your number |
|---|---|
| Items per month (e.g. SoAs, submissions, bordereaux lines) | |
| Minutes of staff time per item | |
| People involved, and their roles | |
| Average hourly cost of those people | |
| Share of items needing rework | |
| Time from start to finish |
Step 2: Work out the monthly cost. Items × minutes ÷ 60 × hourly cost.
Step 3: Set a target. How much should each number move? Be honest. Use your own files, not a vendor's claims.
Step 4: Add the cost of the build and the running costs. Build cost, software and AI usage fees, and the time your own people will spend.
Step 5: Work out the payback. Build cost ÷ monthly saving.
An example, with made-up numbers for illustration only. 100 items a month at 90 minutes each is 150 hours. At $60 an hour that is $9,000 a month. If the target is to cut the time by 40%, that is $3,600 a month saved. A $20,000 build pays back in a little over five months. These are not benchmarks. Your numbers will differ.
Two warnings
- - Time saved is not money saved unless the time goes somewhere useful. Decide where it goes.
- A tool nobody uses saves nothing. Track how many people use it each week.