ROI Math: How to Value an Automation Before You Build It

If you can't put a dollar figure on what an automation will save you before you build it, you're guessing. And guesses, in operations work, are how five-figure projects turn into things nobody uses.

Here's the math I apply on every audit, in three layers — from the back-of-napkin version to the version that holds up in a budget conversation.

Layer 1: The five-second sanity check

For any candidate automation, multiply two numbers:

• →Hours per month spent on the manual version × fully-loaded hourly cost of the person doing it

That's your monthly leak. Multiply by 12 for the annual leak. If the annual leak is less than the build cost, the math doesn't pencil and you should pass.

Example: lead follow-up takes the owner roughly 5 hrs/week. Owner's effective rate is $80/hr (this is the rate she'd command if she were doing billable work instead). 5 × 4.3 × $80 = ~$1,720/month, ~$20,640/year. That's the leak.

Build cost for an AI-powered lead follow-up automation: roughly $5,000–$8,000 in an Implementation Sprint. Annual savings: $20,640. Break-even at ~3 months. Easy yes.

Layer 2: The actual value (which is bigger than the time savings)

Time savings are the floor, not the ceiling. The real value of most automations comes from one of three places:

Revenue capture. Faster lead response wins jobs that would have gone to competitors. *"Lead response time goes from 4 hours to 5 minutes"* often means a 20–40% increase in conversion rate — which on a business doing 50 leads/month at $400 average job value is roughly $4,000–$8,000/month in *new revenue*, on top of the time savings.

Customer quality. Automated reminders reduce no-shows by 30–50%. A no-show isn't just a wasted slot — it's the lost revenue from that job *plus* the opportunity cost of what you could have scheduled instead. Quantify both.

Owner sanity. Genuinely hard to dollarize, but real. Owners burn out. Burned-out owners make bad decisions. An automation that takes invoice-chasing off the owner's plate has a value beyond the four hours/week — it changes what the owner thinks about during evenings.

Layer 3: The risk-adjusted version

Once you've got the value estimate, discount it for two things:

Adoption risk. A perfectly built automation that nobody uses produces zero ROI. Discount the projected value by 20–40% if there's any change-management risk — anyone who has to learn a new workflow, anyone whose habits have to change.

Maintenance load. Automations aren't free after build. They need monitoring, occasional tuning, and updates when underlying systems change. Budget 5–15% of build cost annually for maintenance. (This is one reason retainer plans exist — the maintenance is real.)

What to do with the math

For each candidate automation, you'll end up with three numbers:

• →Annual leak (raw dollar value of doing nothing)
• →Build cost (one-time)
• →Annual maintenance (ongoing)

The decision rule: build it if annual leak > 1.5x (build cost / payback period in years) + annual maintenance. The 1.5x buffer is for the adoption risk discount.

What this looks like on an actual audit

Every audit deliverable I produce has an opportunity table that runs this exact math, opportunity by opportunity. You see the leak, the build cost range, the payback timeline, and the adoption risk score for every candidate. Then we sequence them by ROI — the obvious-yes ones first, the marginal ones flagged, the negative-ROI ideas killed.

That's why audit clients usually recover the audit cost from the *first* automation alone, often within 30 days of implementation. The math isn't optional — it's the whole point.