The business case for a robot fleet subtracts pickers from a payroll and stops there. The building does something else after go-live. One person still stands at each station, paid the picker’s own median wage, and the site adds a maintenance technician at a higher one. The picker’s federal median is $17.95 an hour, or $37,330 a year [1]. The general maintenance and repair technician’s is $23.84 an hour, $49,590 a year [2]. That works out to 1.33 times the picker’s rate, a premium of $12,260 a year per technician, and most cases book it as zero.

At a glance

  • The shift: Your case subtracts pickers; the system keeps an operator at the picker’s wage and adds a technician at 1.33 times it.
  • The number: 1.33 pickers — one technician’s median wage.
  • The risk: Remove fewer than two pickers a station and the payroll may not fall.

The case subtracts; the site substitutes

The wage evidence is dated rather than assumed. All three medians here come from the 2025 edition of the federal wage tables by occupation, which are re-issued every year, so the ratio can be re-checked against the next edition [2]. The heavier service role is dearer still. An industrial machinery mechanic, the trade an automated storage and retrieval installation draws its service technicians from, medians $31.02 an hour and $64,520 a year [4]. One mechanic costs 1.73 pickers.

Timing explains part of the extra cost. The hours a robot displaces sit in the layer of labour a building rents by the week. Temps were an average 16.6% of the peak-season workforce in the 2025 survey, up from 11.4% a year earlier [5]. The technician is not rented by the week; that role is staffed every week of the year. It is also already stretched: 62% of maintenance and repair workers report a typical week of more than forty hours [3].

The same logic runs the other way at the station. The operator is paid what a picker was paid, so the cost per head at the station does not fall at all. Every dollar of saving comes from heads removed, not from the wage, which is why the headcount ratio decides the case and the wage rate only scales it. On the annual figures the technician’s premium is 683 picker-hours, and that is the part of a substitution a subtraction cannot see.

The added hours are expensive for a second, legal reason. Overtime in the United States is computed workweek by workweek. An employee paid for thirty hours one week and fifty the next must still be paid overtime for the second week, even though the two-week average is forty [6]. The rate on those hours is not less than one and one-half times the regular rate [7]. I read the wage-floor argument on this desk — the legal floor, not the industry average, is what a payback should run at — as stopping at the rate, which is where the role mix begins.

By the numbers

  • 1.73 pickers per mechanic: the industrial machinery mechanic’s $31.02 an hour against the picker’s $17.95 [4].
  • 62% past forty hours: the share of maintenance and repair workers whose typical week runs over forty hours [3].
  • $64,520 a year: the mechanic’s annual median, against the picker’s $37,330 [1].

Where the ratio bites

The break-even is arithmetic a reader can do before the quote arrives. A station has to remove 1 + 1/N pickers, where N is the number of stations one technician covers. At one station per technician that is 2.33. At four it is 1.25, and at eight it is 1.13. The ratio is gentle at scale and unforgiving at the pilot, which is the shape most first deployments take.

Premium hours do not rescue a thin saving. Both roles are hourly, so both collect the same 1.5-times rate, and at a fifty-hour week each costs 1.1 times its base. The money in the case grows and the 1.33 ratio does not move. I would want to see a staffing plan that names the technician before the machines land. The occupation is already past forty hours, and the survey’s own list puts upskilling to operate automation last among its retention practices [5].

The catch

I have been quoting national medians here, and that is the first limit on them. The medians are national, so a high-wage metro pays more for both roles and widens the gap, while a low-wage state may find its pickers closer to a legal floor than to the median. A fleet bought as a service also bundles maintenance into the monthly fee. That moves the technician off the site’s payroll and into a fixed charge that persists whether or not the volume does. I would name the technician in the staffing plan before the machines are ordered, because it is the line an automation case is most likely to leave out and it decides whether a second wave can be staffed at all.

Related reading

Sources