Verify Cleaner Location at Start of Shift: 7 Tests to Run

A property manager calls Thursday morning about the medical suite you clean four nights a week. The exam rooms were skipped Tuesday, she says, and her alarm panel shows the suite was disarmed at 7:52 p.m., not the 6:00 p.m. start your timesheet claims.

You have a paper sheet with a scribbled time and a crew member who insists he was there early. That argument costs you a credit, a month of unpaid supervision, and sometimes the account.

Verify cleaner location at the start of a shift by requiring a geofenced GPS clock-in from the worker's own phone, inside a radius sized for that specific property, usually 100 to 150 meters, and then reconciling the stored coordinates against one independent signal: an alarm disarm log, a badge swipe, or a timestamped arrival photo.

That is the whole answer in one sentence. The rest of this article is how you find out whether the method you are about to buy, or the one you already run, actually does it. Below are seven tests, ordered so the cheap ones eliminate bad options before you waste a night in the field.

What a "verified" shift start actually proves

A geofenced clock-in proves one thing: a device carrying credentials for that employee reported coordinates inside a defined circle at a specific second. That is it.

It does not prove the person entered the building. It does not prove they stayed. It does not prove the work happened. Operators who oversell GPS clock-in to a client are setting up a bigger fight later, when the client assumes you have a live map of everyone.

What it does do is shift a dispute from "he says, she says" to "here is the record, here is the alarm log, they agree." That is worth real money at renewal time, and it is the standard you should test against.

Test 1: Can it produce raw coordinates for a single shift?

Run this one from your desk in five minutes, before any demo call. Ask the vendor, or open your current system, and pull the record for one shift last week.

Pass: the record contains latitude and longitude, a timestamp with time zone, the reported accuracy in meters, and which geofence the punch was matched against.

Fail: you get a time and a name. A timestamp with no coordinates is a punch clock with extra steps. Plenty of tools marketed as "mobile time tracking" store nothing more than that, and you will not discover it until you need the record.

What a fail means: eliminate it now. Every remaining test depends on the coordinate existing.

Test 2: How big should the geofence radius be for this building?

Most systems ship a default radius, commonly around 150 meters (about 492 feet). A default is not a decision. Before you accept one, do the arithmetic on your hardest account.

Take Meridian Medical Plaza, a two-story, 42,000 sq ft building on a suburban pad. Footprint is roughly 21,000 sq ft, so about 145 ft on a side if it were square. The diagonal is about 205 ft, meaning any interior point sits within roughly 103 ft (31 m) of the center.

Now add the parking. The crew parks at the far end of a lot that runs about 250 ft from the building, so the real farthest legitimate clock-in point is around 350 ft (107 m) from your pin. A 100 m radius would reject a cleaner standing next to her van. A 150 m radius covers it with margin.

Pass: radius is configurable per location, and you can save a different value for the 8,000 sq ft dental office and the 300,000 sq ft distribution center.

Fail: one global radius for the whole company. You will end up widening it until it swallows the strip mall next door, because that is easier than fielding calls about rejected punches.

Test 3: Does a phone actually get a fix where your crew clocks in?

This is the field test, and it is the one nobody runs before signing. Pick your worst building: basement, metal roof, downtown, or all three. Bring the phone a real cleaner uses, not your new one.

1. Clock in from the parking spot the crew actually uses. Note how many seconds until the punch is accepted.
2. Clock in from the janitor closet or the interior corridor where the shift really begins.
3. Clock in from the sidewalk across the street, or from the neighboring tenant's lot.
4. Repeat step 2 with the phone in low-power mode, because that is how half your crew's phones live.

Pass: steps 1 and 2 accept within about 20 seconds and step 3 is rejected. That is a radius doing its job.

Fail: step 3 is accepted. Your circle is too wide for that site, and the record will not survive a serious dispute. Tighten it and retest.

Also a fail: step 2 hangs for a minute or times out. Do not fix that by widening the radius. Fix it by moving the clock-in point to the entry vestibule or the lot, and documenting that in the site instructions.

Test 4: What happens when GPS fails in a basement?

It will fail. Freight elevators, loading docks below grade, and interior stairwells in older concrete buildings all kill a fix. What matters is the behavior your system chooses, and whether you chose it or inherited it.

There are three possible behaviors. Ask which one you get, then verify it by putting the phone in airplane mode and trying to punch.

• Hard block: no fix, no clock-in. Clean data, but it creates off-the-clock work and a text message to the supervisor at 6:03 p.m. every night.
• Accept with a flag: the punch records, marked as low accuracy or out of range, and lands in an exception list a supervisor clears the next morning. This is usually the right answer for cleaning.
• Silent accept: the punch records with no coordinate and no flag. This is the worst outcome, because your report looks complete and is not.

Pass: failures are visible as exceptions you can review and resolve, with a note field for why.

Fail: silent accept, or a hard block with no documented fallback. If you go with a hard block, you still owe the employee for hours worked, and you need a written manual-entry procedure. Federal recordkeeping rules require accurate records of hours worked regardless of what your app did.

Test 5: Can you corroborate the clock-in with a second signal?

One record wins nothing. Two independent records that agree end the conversation. Before you finish evaluating any method, decide which second signal you can get at each account and how hard it is to pull.

Pass: for every account over a threshold you set, say anything billing more than a few thousand dollars a month, you have at least one second signal identified and a way to get it within a day.

Fail: your only record is the app. That is fine for a small strip-mall route. It is not fine for the hospital wing where a disputed start time becomes a compliance issue.

Test 6: Can one person clock in the whole crew?

Buddy punching does not disappear because GPS is involved. It just moves. Test it directly, using your own admin access and a real crew member's login, from your office 12 miles away.

Pass: the punch is rejected for being outside the geofence, and the attempt is visible somewhere you can see it.

Fail: it goes through. Now ask the harder question: does one crew lead clock in three people on one phone from inside the geofence? If your process allows a lead to punch the crew, your location record only proves the lead's phone was there.

Be honest with yourself about spoofing too. Mock-location tools exist on Android, and no consumer clock-in product reliably detects every one of them. Many operators find that the practical control is not detection but consequence: a written policy, plus the second signal from Test 5, plus the fact that a spoofed punch and an alarm log that disagrees is a very short disciplinary conversation.

Test 7: Can you rebuild a shift from 60 days ago in 10 minutes?

Disputes are never about last night. They arrive on the fifteenth of the month, attached to an invoice, about a Tuesday six weeks back. Run this test before you commit, using historical data if you have it or the vendor's demo data if you do not.

1. Pick one date roughly 60 days back and one location.
2. Pull every punch for that location and date, with coordinates and accuracy.
3. Export it to CSV without asking anyone for help.
4. Time yourself.

Pass: under 10 minutes, self-serve, in a format you can attach to an email or hand to the client's facilities director.

Fail: screenshots only, no export, or a data retention window shorter than your billing dispute window. Ask directly how long punch records are retained. If it is 30 days and your clients pay on net 30, you are structurally unable to defend an invoice.

Do you need employee consent to track location at clock-in?

Treat this as a policy question you settle in writing, not a technical detail. Several states regulate electronic tracking of individuals or devices and have notice requirements, and rules differ enough that a national template is a bad idea. Have counsel review your language once, then reuse it.

Two operational rules keep you out of most trouble regardless of jurisdiction. First, capture location at the punch only, and be able to say that plainly to a crew that will absolutely ask whether you are following them home. Second, if a crew member uses a personal phone, address data usage and reimbursement in your policy rather than leaving it to a rumor in the break room.

The conversation goes better than owners expect when you frame it correctly. Cleaners get accused of skipping buildings constantly. A record that proves they showed up at 5:58 p.m. protects them first.

Frequently asked questions

How accurate is phone GPS when a cleaner clocks in inside a building?

Under open sky, GPS.gov reports smartphones are typically accurate to within about a 4.9 m radius, and accuracy degrades near buildings and other obstructions. Inside a structure, expect much larger error, and in a basement often no usable fix at all. Design the geofence around a punch point with sky visibility, such as the entry door or the lot.

Can an employee fake a GPS clock-in location?

Yes. Mock-location tools exist on Android, and no consumer time-tracking product reliably detects all of them. The practical defense is not detection, it is corroboration: pair the punch with an alarm disarm log or badge swipe, and make falsifying a punch a written terminable offense. When two independent records disagree, you have your answer.

What geofence radius should I use in a shared office park?

Start with the tightest radius that still covers the parking your crew legitimately uses, often 100 to 150 m, then check what else falls inside the circle. If the neighboring tenant's building is captured, either tighten the radius and move the punch point to the entry door, or accept that the geofence alone will not distinguish the two buildings and lean on a second signal.

Do I have to pay a cleaner who could not clock in because GPS failed?

Yes. Hours worked are hours worked, and FLSA recordkeeping rules (29 CFR Part 516) require accurate records regardless of whether your app cooperated. Build a manual-entry path with a supervisor approval and a reason code. A hard block with no fallback quietly pushes people into unpaid time, which is a far more expensive problem than a missing coordinate.

Should I show clients the clock-in records?

Selectively, and on your terms. Sharing verified arrival times in a monthly report is a strong differentiator in bids against low-price competitors. Do not promise live tracking or a map of movement between punches, because location captured at clock-in and clock-out is not a breadcrumb trail, and a client who expects one will be disappointed.

Where CleanTrack360 fits

If your tests point toward a system rather than a paper sheet, CleanTrack360 includes geofenced GPS clock-in and clock-out with a default radius of 150 m that you can configure per location, which is exactly what Test 2 asks for. It runs in the phone browser, so there is nothing for a crew to install: the native mobile app is still in development. Location is captured at clock-in and clock-out only, not as a trail between them, and there is no GPS spoofing detection, which is why Test 5 and Test 6 matter no matter what software you use.

For Test 7, reports export to CSV, and the browser-based client dashboard shows schedules and inspection reports if you choose to share them. Plans are Starter at $99 per month for up to 5 team members, Pro at $199 per month for up to 20, and Business at $249 per month for up to 50, priced per plan rather than per user. There is a 14-day free trial with no credit card, which is enough time to run all seven tests on your worst building before you decide.