How County Procurement Bids Cleaning Robot Leasing for Schools

How does a county procurement office bid out cleaning robot leasing for school campuses?

A county procurement office that wants to lease cleaning robots for school buildings usually begins with a formal solicitation, not a demonstration. The solicitation, whether an invitation for bids or a request for proposals, has to say what work the lease covers, where that work happens and how the county will judge whether it was done. Vendors can only answer what they can measure, so the scope is the most important part of the document.

This page covers four things in order: how large the cleaning job is on a school campus, which staffing and injury figures belong in the justification, how the county code may affect the bid method, and what the steps after award look like. The figures come from federal labor and injury statistics, national school counts and an industry staffing benchmark, and each one is named where it appears. Where a source is dated or secondhand, the text says so.

We are one of the firms that answer bids like this, and we are not a robot manufacturer. We choose the unit that fits each site across makers, lease it through financing, deploy and integrate it, train staff, and service every unit through regional service engineers under our service plan. The payment comes in a quote after a free site assessment, and that assessment is where the county gets real floor data for its specification.

How large is the school cleaning job across the US?

Public schools are a large and steady market for floor care. A secondary site citing the National Center for Education Statistics Common Core of Data counted 99,073 public schools operating in the 50 states and D.C. in 2024-25, serving 94,177 reporting students across 17,648 districts. An industry publication reported 98,609 public schools in operation for 2020-21, so the count changes slowly from year to year, which helps a county write a multi-year scope.

The school count alone does not tell a bid writer how much floor a robot must cover. The county should take building square footage from its own facilities records, split it by building type, and apply a cleaning level to each type before the scope is written.

What staffing benchmarks should the specification use?

A community college facilities page that reproduces the APPA Custodial Staffing Guidelines gives square feet per custodian for classroom hard floors at five cleaning levels: 8,500 at Level 1, 16,700 at Level 2, 26,500 at Level 3, 39,500 at Level 4 and 45,600 at Level 5. A lower number means a higher cleaning level, so a school held to Level 1 needs a much heavier crew for the same floor than a school held to Level 5.

A Virginia school district's allocation document, citing NCES with the U.S. Department of Education, puts the range at about 18,000 to 25,000 square feet per custodian in an eight-hour shift, and calls 25,000 the normal amount for most school facilities. That range sits between the Level 2 and Level 3 APPA values, so the county should compare its own staffing against more than one benchmark and state which one the bid uses.

The same community college page includes a local analysis, which is not an APPA rule, warning that a drop to 80 percent attendance could push service down a level or more. For a school campus, that matters because a robot schedule built on a full staff can fall short when absences rise. The specification should state the attendance assumption it uses and ask vendors to show their coverage at a lower staffing level.

Slow projected growth in the labor pool means a county cannot count on hiring alone to fill cleaning shifts. The Bureau of Labor Statistics counted 2,447,700 janitor and building cleaner jobs in 2024 and projects about 47,800 jobs added from 2024 to 2034, a growth rate of 2 percent that is slower than average. We did not verify an official turnover rate for this occupation, so the bid should not state one.

Which injury and illness figures belong in the justification?

Injury data is a legitimate reason to scope robots, but the figures have to match the job. A Bureau of Labor Statistics occupational injury snapshot covering 2015 to 2019 reports 18,680 janitor cases involving days away from work in 2019, with an incidence rate of 157.4 per 10,000 full-time equivalent workers. That is a rate per worker, and the bid should describe it that way rather than as a count of injuries at a school.

A second BLS snapshot, for janitors and cleaners excluding maids and housekeeping cleaners, reports 21,100 private-industry nonfatal cases involving days away from work in 2020. The two snapshots use different years and occupational scopes, so the bid should cite each with its year and scope rather than merging them into one trend.

We did not find a 2024 nonfatal rate for janitors, so the bid should use the most recent year available and say so. Fatality totals for the wider building and grounds cleaning group are counted on a different basis from nonfatal rates, and they should stay out of any comparison with robot performance.

The better use of injury data is to tie it to a task. If the scope names repetitive heavy work, such as pushing equipment through the same corridors every night, the bid can ask that those routes be covered by the robot and that the custodian hours released be measured. Avoid asking vendors to promise fewer injuries, because no vendor can verify that for a county.

Does a county lease of cleaning robots need sealed bids?

The answer depends on the code, and the code often treats a lease differently from an acquisition. The Snohomish County code in Washington applies formal sealed bidding to acquisitions and leases of personal property, subject to listed exceptions. A county under a similar rule should check those exceptions before treating a robot lease as an ordinary operating expense.

That is one state's rule, not a national one. Other codes may allow a lease through a request for proposals, a cooperative contract or a separate approval, and some may set thresholds below which bidding is not required. Counsel should confirm the path, the threshold and any exception before the procurement schedule is set, because a bid method chosen in error can delay deployment.

A workable structure is one bid covering the lease, deployment, training and service plan for a defined set of buildings, with an option to add units or buildings on the same terms. That lets the county test a pilot with real coverage data before committing to the full campus list.

What should a school campus cleaning robot bid specification include?

Write the specification as a set of measurable requirements, not as a list of desired features. A vendor that can meet a measurable requirement can bid on it, and a vendor that cannot can say so in its response. The elements below give a bid writer enough detail to compare proposals fairly.

• Floor area by building, split into hard floor corridors, cafeterias, gymnasiums and other open areas, with the source of the square footage named
• Cleaning level for each area, stated against the staffing benchmark the county uses
• Operating window, including whether runs happen overnight, during occupied hours, or both, and how students and staff are kept clear
• Access rules covering doors, alarms, key control and the person who opens each building
• Route and reporting requirements, including coverage maps, exception logs and a record of completed runs the county can audit
• Service response, covering the time to respond to a stopped unit, spare unit coverage and who handles remote triage
• Training for custodial staff and a go-live acceptance test that the county can run itself

Where do cleaning robots fit on a school campus, and where do they not?

Robots fit the large open hard floors that take the most custodian time at night, such as main corridors, cafeteria floors and gymnasium floors. These areas are wide, mostly free of fixed obstacles when the building is empty, and they can be covered on a repeating route. The scope should name them first so the bid is tied to the work that robots can take over.

Restrooms, classrooms with desks and loose furniture, stairs, carpets, and areas with frequent spills still need people. A floor unit cannot clean fixtures, wipe surfaces or handle a classroom where chairs are stacked in unusual places. A bid that expects robots to cover those areas will disappoint, so it should keep them with custodians and say so plainly.

Campus layout matters too. Long connected corridors and a clear route between buildings let a unit run a full loop, while a building with many small rooms, ramps and changes in floor level needs a site survey before its route is drawn.

What are the steps after a county awards a robot lease?

After award, the work moves from the bid document to the building. The first step is a site assessment, in which the vendor surveys the buildings, maps the floors and confirms the access and alarm rules the bid described. That survey is what turns the specification into a lease quoted after a site assessment, and it should be finished before the county signs.

We run the sequence in the same order at every site. We assess the site at no charge, then quote the lease through financing, with the payment stated in that quote. Once the lease is signed, we deploy and integrate the units, train custodial and facilities staff, and hand over the site under our service plan, with regional service engineers in every state we service.

Go-live is the point where the county should check the work, not the vendor. The acceptance test in the bid should compare the coverage log against the area list, confirm that the exception rules work, and confirm that custodial staff can clear a stopped unit without help. Only then should the units run without an operator.

How should the contract handle service, downtime and replacement?

A school campus cannot tolerate a cleaning route that quietly stops. The contract should say how fast a stopped unit gets a response, whether a remote check comes first, and when an on-site visit follows. Ask each bidder to state those times in writing, because a promise made only in a sales meeting gives the county nothing to enforce.

Spare coverage matters as much as the response time. A county with one unit per building has no cover while that unit is in for repair, so the bid should ask whether the vendor provides spare or loaner units and how quickly they are placed. Ask also whether the spare pool is shared across buildings under the same lease.

Do not accept a promise of zero downtime. A vendor can commit to response and replacement times, but no one can promise that a unit never stops, and a bid that asks for that will score marketing rather than service. The more useful test is what happens in the first hour after a unit fails.

How does the county know the robots are doing the work?

Measure the work by area and by log, not by how the unit looks in a demonstration. The coverage log should show which zones ran, when they ran and how long each took, and the exception log should show every stop with its cause. Compare those logs against the area list in the scope during the first weeks of operation, while the problems are still easy to trace.

Ask custodial leads to report each morning after a run, including any corridor that was not finished, any spill the unit missed and any hazard a person had to handle. Those reports show whether the robot scope is set correctly or whether the cleaning level in the bid needs adjustment. At the end of the first term, use the logs and the staffing data to decide whether to add units or buildings.