The Standard Answer to "How Fast Does It Pay Back?"

Last month, the owner of an auto-parts distributor looked me in the eye and asked: "Lao Mi, be straight with me — how long until this thing pays for itself?" He was looking at a goods-to-person system quoted at $850,000, with the integrator's deck claiming an "18-month payback." I told him: take that number, divide it by two, and then we'll talk — because at least half the assumptions in that deck describe a perfect world that doesn't exist.

"How fast does it pay back?" I've heard that question maybe eighty or a hundred times over twenty years. Here's what I've learned: the problem was never the answer. The problem is that nobody in the room is working from the same math. So let me lay out the math — clearly enough that you can write it on a whiteboard in your next meeting.

First, Get the Total Investment Right — Not Just the Quote

Most payback calculations fail at step one: they only count the equipment quote.

The standard formula is one line: payback period = total investment ÷ annual net savings. Get the numerator wrong and everything after it is fiction.

"Total investment" needs a checklist, and missing any line item means your math is wrong. The hardware is only one piece. On top of it: the integrator's engineering services (design, installation, commissioning), facility work (floor leveling, structural reinforcement, charging areas, fire sprinkler rerouting, safety fencing), software (the fleet management system, integration development with your WMS), training (operator training plus the productivity dip during ramp-up), and the first year's maintenance.

Here's a real-world sense of scale. Last year, an 80,000-square-foot 3PL warehouse deployed 12 material-handling AMRs. Equipment quote: $280,000. Actual money spent: $470,000 — $90K in integration services, $40K in floor and charging-area work, $30K in WMS integration development, over $20K in the first-year maintenance contract, and about $10K in training plus two months of lost productivity. The quote said $280K; reality said $470K. That extra $190K is the part that "mysteriously disappears" from most payback calculations.

Calculating Annual Savings: Price Labor by the Hour, Not by Headcount

The denominator is annual net savings. Labor is always the biggest piece — but you can't calculate it as loosely as "we'll save five people."

Fully loaded warehouse labor in the US runs about $25–$35 per hour these days. That's not the wage on the job posting; it's wages plus insurance, taxes, and benefits. In warehouse-heavy regions like the Inland Empire, where entry-level hiring has been climbing, fully loaded cost sits close to $30 an hour. Use that number — not the hourly rate from the recruiting ad.

The math is simple: annual savings = hours eliminated × fully loaded hourly cost.

Take a two-shift picking zone. It ran with 20 people; after goods-to-person, it runs with 12. That's 8 FTEs eliminated. At 2,080 hours per person per year and $30 an hour fully loaded, that's $499,200 a year — call it half a million.

But labor is only the first piece. Two more get left out constantly. One is error rate. Manual picking typically runs 1–3 errors per thousand; automation can push that below 0.5 per thousand. A single mispick costs about $25 all-in (reshipment, returns, customer service time). For a warehouse shipping 10,000 orders a day, cutting errors from 2‰ to 0.5‰ saves $136,500 a year. The other is overtime. Peak-season overtime pays at 1.5x; once automation lifts peak throughput, you hire fewer temps and pay less overtime — $50K to $100K a year in savings is common.

So: annual net savings = labor savings + error savings + overtime savings − annual maintenance. Note that minus sign. Maintenance is a bill you pay every year, and a lot of people forget to subtract it the moment they finish adding.

Static Payback Is for Meetings; Discounted Payback Is for Yourself

Static payback is total investment divided by annual net savings, ignoring the time value of money. $470K invested, $300K saved per year: 1.57 years. The boss understands it. Good enough for the meeting.

Discounted payback folds future savings back to today at a discount rate — usually 8–12%. It comes out a few months longer. This one isn't for the boss; it's for you. When you're on the fence about whether a project is really worth it, discounted payback forces you to face an uncomfortable truth: savings arrive year by year, and a dollar saved in year four is worth less than a dollar saved in year one. A project with 2-year static and 2.6-year discounted payback is a fundamentally different risk than one with 3-year static and 4-year discounted payback.

My rule of thumb: payback under 2 years — go for it. Two to three years — audit the investment checklist line by line for missing items. Over 3 years — don't sign; go back and recheck utilization assumptions and maintenance costs. Nine times out of ten, the original math was wrong.

A Calculation Sheet the Boss Can Actually Read

Put all of this on one page and project it in the meeting:

Honestly, this is a "looks great on paper" example. In reality, error savings and overtime savings rarely come in this high — and that's exactly why I break every line out. It lets the boss challenge each row in the meeting instead of hiding behind a single number nobody can explain. A number that survives questioning is a number you can sign off on.

The Three Classic "Payback Illusions"

Illusion one: forgetting maintenance. Annual maintenance on automation equipment typically runs 5–10% of the equipment price. On $280K of equipment, that's roughly $20K a year — $100K over five years. A deck that promises "18-month payback" becomes 26 months once maintenance is added back in.

Illusion two: assuming 100% utilization. Integrators model equipment running flat-out around the clock. In reality, your warehouse might only run full tilt for the three peak-season months, with utilization at 40% the rest of the year. Annual savings should be a weighted average, peak and off-peak calculated separately. The worst case I ever saw: the deck assumed 20 hours of daily utilization; the warehouse actually ran 10 hours a day. The payback period doubled.

Illusion three: leaving out the integrator's on-site fees. For the first three to six months after go-live, integrators usually station engineers on-site to work through the shakeout period. Some of this is in the contract; some of it bills by the day — $1,500 to $2,500 a day is typical. Three months of on-site support is $100K and up. If it wasn't written into the contract, it usually gets "forgotten" in the payback math.

One Thing You Can Do Tomorrow

Take the automation project you're currently evaluating and rebuild the calculation using the table above. Check three lines: is maintenance and on-site support missing from total investment? Are annual savings modeled at 100% utilization? Is the per-error cost behind the error savings a real number or a guess? If payback still lands under 2 years, the project is probably solid. If it drifts past 3 years, stop talking about equipment — go back, pull the raw data on utilization and labor costs, and verify them one by one.

Being asked "how fast does it pay back" isn't the scary part. The scary part is not knowing where your own number came from.