It is 9:40 p.m. Your account manager at a 62,000 square foot medical office building calls to say the lobby glass was not done and the trash in Suite 210 is still there. You check your paper time sheet the next morning and it says the two-person crew worked 10:00 p.m. to 2:00 a.m., signed in blue ink.
You have no idea whether that is true. Not because your people are dishonest, but because a signature on a clipboard is not evidence of anything. It is a statement of intent written at the end of a shift when everyone is tired and the sheet has to be filled in before it goes back to the office.
That gap is what GPS clock-in solves, and it is also where most operators buy the wrong tool. They shop for "tracking" when what they actually need is a defensible payroll record tied to a specific address, a specific scheduled shift, and a specific person. This guide walks through how to evaluate that, what the technology can and cannot do, and how to roll it out to a crew that has never used it.
What GPS clock-in actually proves, and what it does not
A geofenced clock-in captures a latitude and longitude at the moment an employee taps "clock in," compares it to the coordinates of the job site, and either accepts the punch or flags it. That is the whole mechanism. It is simple, and it is powerful precisely because it is simple.
Here is what it gives you: a timestamped record that a specific device was within a defined distance of a specific building when the shift started and when it ended. That single fact resolves most of the disputes that eat your margin.
Here is what it does not give you. It does not prove the restrooms were cleaned. It does not prove the person holding the phone is the person on the schedule. And unless the vendor specifically offers continuous location logging, it does not tell you where anyone was between punches.
According to GPS.gov, GPS-enabled smartphones are typically accurate to within about a 4.9 meter (16 foot) radius under open sky, and accuracy degrades near buildings, bridges and trees. In a downtown core surrounded by towers, or inside a concrete stairwell, a phone may fall back to Wi-Fi and cell tower positioning that can be off by tens or even hundreds of meters.
That single technical fact drives almost every configuration decision below. If you set a 25 meter geofence on a high-rise account, you will spend your evenings approving failed punches from honest employees.
The real dollars at stake
Time theft in janitorial rarely looks like fraud. It looks like drift: a crew that arrives at 10:12 and writes 10:00, or clocks out at the truck at 2:05 after finishing at 1:50.
Run the arithmetic on your own book. Assume 10 minutes of unearned time per shift, five shifts per week, 50 weeks per year. That is 41.7 hours per employee per year. At an illustrative $17 per hour, that is roughly $708 per person, or about $14,200 across a 20-person crew, before payroll taxes and workers' compensation premium, which scale off payroll.
Those assumptions are illustrative, not a benchmark. Substitute your own wage rate: the Bureau of Labor Statistics publishes median hourly wages for Janitors and Cleaners, Except Maids and Housekeeping Cleaners (SOC 37-2011) in its Occupational Employment and Wage Statistics program, by state and metro area.
The second dollar figure is the one operators forget: the cost of not being able to prove service. When a client threatens to cancel because "nobody showed up Tuesday," a geofenced punch log with a timestamp is the difference between a credit memo and a closed conversation.
The nine requirements that actually matter for janitorial
Most time clock apps were built for construction, field service or restaurants. Cleaning has a specific shape: night work, multiple short stops per shift, high turnover, personal phones, and buildings that swallow cell signal. Score every vendor against these nine.
| Requirement | Why it matters in cleaning | Question to ask the vendor |
|---|---|---|
| Per-location geofence radius | A strip mall unit and a hospital campus cannot share the same setting | Can I set a different radius per site, and what is the default? |
| Multi-stop shifts | One cleaner may hit four buildings between 6 p.m. and 11 p.m. | Can one person clock in and out of several sites in one night without confusion? |
| Punch tied to a scheduled shift | An unmatched punch is data. A punch attached to a work order is accountability | Does the system flag a punch with no matching schedule? |
| Zero-install access | Crews using personal phones resist downloads, and turnover means constant reinstalls | Does it run in the phone browser, or does every hire need an app store account? |
| Edit audit trail | Manual corrections are legitimate and constant. Undocumented ones are a liability | Does every edit record who changed what, when, and why? |
| Missed clock-out handling | People forget. Nightly. It is the single most common exception | What happens at hour 12 of an open punch? Alert, auto-close, or nothing? |
| Language support | Much of the front-line workforce is more comfortable in Spanish or Portuguese | What languages does the employee-facing screen support? |
| Exportable records | Payroll, client disputes and wage audits all need raw data | Can I export punch detail to CSV without contacting support? |
| Predictable pricing | Janitorial turnover makes per-seat billing volatile | Is the price per user, per location, or a flat plan tier? |
Setting the geofence radius correctly
This is the setting that determines whether your crew trusts the system or fights it. Too tight and honest employees get locked out. Too loose and someone can punch in from the parking lot of the building next door.
The radius has to absorb three things: GPS error at that specific site, the building footprint, and the walking path from where crews park to where they get signal. Start generous, then tighten with real data after 30 days of punches.
| Site type | Practical starting radius | Reasoning |
|---|---|---|
| Standalone suburban office, own lot | 100 to 150 m | Open sky, good accuracy, but parking may sit far from the service door |
| Retail unit in a shared plaza | 75 to 100 m | Tighter to avoid capturing neighboring tenants you do not service |
| Downtown high-rise | 150 to 250 m | Urban canyon effects degrade accuracy badly. Expect drift |
| Warehouse or distribution center | 200 to 300 m | The building footprint alone can exceed 200 m end to end |
| Multi-building campus (medical, school district) | Separate geofence per building | One giant radius destroys your ability to tell which building was serviced |
| Small site with homes nearby | 50 to 75 m | Keeps the fence off residential property lines |
Native app versus phone browser
Operators assume they need an app store download. In practice, a browser-based clock-in page often gets adopted faster in this industry, for three unglamorous reasons.
- No install friction: A new hire starting tonight can bookmark a link in 30 seconds. No app store password to recover, no storage space excuse.
- No BYOD storage fight: Many cleaners run older phones that are already full. "I could not download it" disappears as a reason.
- Faster turnover recovery: When you replace 3 people in a month, link-and-login beats install-and-configure every time.
The trade-off is real and you should know it. A browser page generally needs live connectivity at the moment of the punch, and it cannot capture location in the background. Native apps can queue punches offline and sync later.
For most cleaning routes this is a non-issue if you train one rule: clock in at the entrance before you go inside, clock out at the entrance on the way to your vehicle. Signal at the door is nearly always fine. Signal in a sub-basement mechanical room is not, and no app fixes that.
Legal guardrails you need in place before day one
None of this is legal advice, and you should run your policy past counsel in your state. But these are the four issues that come up repeatedly.
Recordkeeping. The Fair Labor Standards Act requires employers to keep accurate records of hours worked. Under 29 CFR Part 516, payroll records must be retained for three years and the underlying time cards and work schedules for two years. A GPS system helps you comply, but only if you can export and archive the data.
Rounding. Federal regulation at 29 CFR 785.48(b) permits rounding recorded time to the nearest increment, commonly a quarter hour, provided the practice is neutral over time and does not systematically shortchange employees. Some states are stricter. Once you have exact GPS timestamps, ask yourself why you are still rounding at all.
Employee notice of electronic monitoring. Several states regulate this. New York requires private employers to give written notice of electronic monitoring to new hires (Civil Rights Law 52-c), and Connecticut requires notice of electronic monitoring under Conn. Gen. Stat. 31-48d. Give written notice everywhere as standard practice, in the languages your crew reads.
Biometrics are a separate legal category. If you are considering fingerprint or facial recognition for punch verification, understand that Illinois' Biometric Information Privacy Act (740 ILCS 14) and the Texas CUBI statute impose consent and retention obligations with real teeth. A plain selfie photo attached to a punch is generally a lighter footprint than a biometric template, but confirm with counsel.
Pricing models: run the math before you sign
Per-user pricing sounds cheap at 6 cleaners and stops sounding cheap at 30. Turnover makes it worse, because you are constantly adding and deactivating seats and every vendor handles mid-cycle changes differently.
Use this illustrative comparison. Assume a hypothetical vendor charging a $20 base fee plus $6 per user per month, against a flat tier at $199 per month.
| Team size | Hypothetical per-user model | Flat tier at $199 |
|---|---|---|
| 6 people | $56 per month | $199 per month |
| 15 people | $110 per month | $199 per month |
| 25 people | $170 per month | $199 per month |
| 40 people | $260 per month | $199 per month (if the tier covers 40) |
The figures above are arithmetic on invented rates, used only to show the crossover point. Do this calculation with real quotes at your headcount today and at your headcount if you land the next two accounts.
Two questions decide the winner. First: does the seat count include office staff, supervisors and inspectors, or only field crew? Second: when a cleaner quits on the 3rd of the month, do you get that seat back immediately?
A 30-day rollout that does not blow up
The technology is the easy part. Getting 24 people to change a nightly habit is the hard part. Sequence it.
Week-by-week implementation plan
- Days 1 to 3: Write the policy in one page: when to punch, where to punch, what happens if you forget, who to text. Translate it.
- Days 4 to 7: Load your locations with correct service-entrance coordinates. Use bulk import if the platform offers it, and verify every pin on a map.
- Days 8 to 10: Set the geofence radius per site using the table above. Do not accept one global default.
- Days 11 to 14: Pilot with two sites and your most cooperative supervisor. Run GPS punches and paper in parallel. Compare them nightly.
- Days 15 to 18: Fix what the pilot exposed. Usually: one site with a bad pin, one building with dead signal at the door, one person whose phone denies location permission.
- Days 19 to 25: Roll out to all crews, one supervisor's territory per night so you can support each group live during their shift.
- Days 26 to 30: Retire paper. Announce the cutoff date in writing, then hold it. Parallel systems run forever if you let them.
The four numbers to watch after go-live
A GPS clock-in system that nobody audits becomes a very expensive digital clipboard. Track these four metrics weekly for the first 60 days, then monthly.
| Metric | Formula | What it tells you |
|---|---|---|
| Punch compliance rate | (Shifts with valid geofenced clock-in and clock-out / scheduled shifts) x 100 | Adoption. Low numbers after 60 days mean a training or configuration problem, not a people problem |
| Manual edit rate | (Manually edited punches / total punches) x 100 | A persistently high rate usually points to one bad geofence or one dead-signal building |
| Missed clock-out rate | (Open punches auto-closed or corrected / total shifts) x 100 | Habit formation. This should fall sharply after week three |
| Scheduled vs actual hours variance | Actual on-site hours minus scheduled hours, per site per week | Your real labor cost by account. This is the number that fixes underpriced contracts |
Many operators find the variance number is the one that pays for the software. When you learn that a building you bid at 3.0 hours per night consistently takes 3.7, you either raise the price at renewal or change the scope. Both beat absorbing it.
Common mistakes to avoid
- Setting the radius too tight to "catch" people. A 25 meter fence on a high-rise fails honest employees and trains everyone to text you instead of using the app. Start wide, tighten with data.
- Geocoding by street address only. The pin lands on the front door or the middle of the parcel. Your crew enters at the dock. Verify every location manually.
- Treating a valid punch as proof of quality. Presence is not performance. Pair punch data with a checklist inspection and photo evidence, or you have only automated the honor system.
- Skipping written notice and consent. Five minutes of paperwork prevents a genuinely expensive problem, especially in states with electronic monitoring notice laws.
- Assuming spoofing is impossible. Mock location apps exist. Unless a vendor explicitly documents spoof detection, assume it does not have it, and rely on secondary signals: inspection photos with timestamps, client sign-off, supervisor spot checks.
- Running paper and digital in parallel indefinitely. Crews will default to whichever system is less work. Set a hard cutoff date and enforce it.
- Buying a standalone time clock. If punches do not connect to schedules, inspections and invoicing, you have created a new data silo to reconcile by hand every Friday.
How often to review your setup
| Cadence | What to review |
|---|---|
| Daily (first 30 days) | Failed punches and open shifts from last night. Fix root causes immediately |
| Weekly | Manual edit log, missed clock-outs, late starts by crew and by site |
| Per pay period | Punch detail export reconciled against payroll before you submit it |
| Monthly | Scheduled vs actual hours variance by account, feeding your margin review |
| Quarterly | Geofence radius tuning on the five sites generating the most exceptions |
| Annually | Written policy, notice and consent documents, retention practice against 29 CFR Part 516, and pricing at your current headcount |
How CleanTrack360 handles this
CleanTrack360 includes geofenced GPS clock-in and clock-out that runs in the phone browser, so there is nothing for a new hire to install on their personal device. The default radius is 150 m and is configurable per location, which lets you tighten a strip-mall unit and widen a downtown tower independently. Location is captured at clock-in and clock-out only, not continuously between punches, and there is no spoofing detection, so treat the punch log as one input alongside inspections rather than the last word.
Because punches attach to drag-and-drop schedules and recurring shifts, you can compare scheduled hours against actual hours by site, then export punch detail to CSV for payroll and record retention. Quality verification lives in the same system through custom inspection checklists with photo evidence and automatic scoring, and clients can see schedules and inspection reports in a browser-based dashboard. Plans are flat rate rather than per user: Starter at $99 per month for up to 5 team members, Pro at $199 for up to 20, and Business at $249 for up to 50, with a 14-day free trial and no credit card required. The employee-facing interface is available in English, Spanish and Portuguese.