Did It Actually Work? Measuring the Payoff

The one rule: measure before you automate

You cannot prove a payoff you never baselined. Skipping the before-measurement is the single biggest reason owners cannot defend what their automation is worth. Time the task by hand, three to five times, before you change anything.

The five-step method

1. 1
   Pick one workflow. Map it before you touch a tool.
2. 2
   Time it for a week or two. A Google Sheet with five columns is enough: Date, what ran, minutes spent, did it need a redo, notes.
3. 3
   Write down the averages and set one target.
4. 4
   Automate, then re-measure at 30, 60, and 90 days. Most of the impact shows up inside the first 30 to 60.
5. 5
   Do the math. Payoff = (hours saved per year times your loaded hourly cost) plus any revenue from faster follow-up, minus what the software and setup and upkeep cost.

The spreadsheet is the whole framework. If the work lives in your email or CRM, the timestamps are already being recorded for free.

The handful of numbers worth tracking

Track a few, consistently. Two of them actually change decisions:

• Hours saved, turned into dollars. Value the time at about 1.25 to 1.4 times the wage to cover benefits and overhead. This is the number that moves an owner.
• Lead response time. Inquiry received versus first reply. The most measurable, highest-leverage metric for any business that takes inbound work.

Supporting numbers if a workflow needs them: cost per run, error or rework rate, and end-to-end turnaround time.

The numbers to distrust

Be openly skeptical of vendor ROI claims:

• A 2025 MIT study found 95 percent of business GenAI pilots delivered no measurable bottom-line impact, and only about 5 percent produced fast returns. That is the honest backdrop.
• Ignore eye-popping multiples like "1,275 percent ROI" or "$3.70 back for every dollar." They are vendor-sourced, show no baseline, and do not transfer to your shop.
• One useful finding from that same research: the money tends to get spent on sales and marketing, but the real payoff usually shows up in the boring back office. Measure there too.

The trustworthy number is the one your own before-and-after produces. Auto-Phil will not quote you a multiple; it will baseline your task and show you the after.