Every cleaning company eventually has the same morning. A crew is parked outside a house at 8:05, the client is already at work, and nobody on the truck knows whether the key is on the ring, under the mat, or in a lockbox whose code was texted to somebody who quit in March. The visit does not happen. The client gets a phone call they did not want. Two hours of paid labor go somewhere other than a billable job, and the crew's whole route slides right.
Nobody plans this. It happens because access is the one operational asset in a cleaning business that almost never gets systematized. Parts get counted. Hours get tracked. Invoices get sent. But the keys, the lockbox codes, the gate fobs, and the alarm procedure live in a group chat, a shoebox in the office, and one supervisor's memory. That works fine right up until somebody quits with a ring of client keys and you cannot produce a list of which doors those keys open.
As of August 2026, most cleaning and janitorial operations still run access informally, and the cost shows up in places they do not connect back to it: lost visits, angry clients, awkward rekey conversations, and a turnover process that eats a weekend. This guide covers where the access record actually belongs, what goes on it, how to treat physical keys like the tooling they are, and how to make an employee departure a checklist instead of a crisis.
The access record belongs on the property record
The fix is unglamorous: every piece of access information belongs on the customer's property record, inside the same system that schedules the visit and produces the invoice. Not a group text. Not a binder. Not a spreadsheet somebody maintains out of goodwill.
A group text fails for three specific reasons, and they are worth naming because each one bites at a different moment. It has no owner, so nobody is responsible for it being current. It has no history, so you cannot tell when a code changed or who was told. And it is unqueryable, so when the question is "which properties can this person get into," the only answer is scrolling.
A property record fixes all three by attaching access to the thing it actually belongs to: the address. When the record lives with the customer, it inherits everything else you already track about that customer, including who has been assigned there, when the last visit happened, and what was billed. That connection is what turns access from tribal knowledge into something you can look up, hand off, and audit.
There is a related discipline worth being explicit about. Access details are internal operating records, not client-facing ones. A customer portal is the right home for receipts, service history, and online payment, because those are things a client benefits from self-serving. Entry instructions and key custody are not that. Keeping the line clean is part of why clients hand you a key in the first place.
What goes on the record, and what should not be casually shared
A good access record answers the questions a cleaner standing on the porch would otherwise have to call about. In practice that means:
- Entry method. Front door key on the office ring, lockbox on the side gate, client home on Tuesdays, building fob for the elevator. Be specific about which door.
- Parking. Where the crew is supposed to put a van in a townhouse complex or a downtown block, and whether a permit or a visitor spot is involved.
- Pets. Names, temperament, where they are usually kept, and whether the crew is expected to do anything about them. This is a real safety and lost-visit issue, not a nicety.
- Alarm procedure. What the crew is expected to do on entry and exit, and who to contact if the panel does not behave. Whether the code itself belongs in the record is a decision the owner and client make together.
- Supply location. Which closet holds the vacuum, whether the client supplies chemicals, where trash goes on collection day.
- Do-not-touch notes. The home office desk, the aquarium, the room the client uses for work calls, the antique floor that gets a specific product.
Now the part that matters more than the list. Codes and key locations are sensitive records, and the right posture is to limit them to the people who need them for an assigned visit rather than broadcasting them to everyone who works for you. A company-wide thread is the opposite of that: it persists forever, it is readable by every phone that ever joined it, and it walks out the door with each departure.
No software makes that decision for you. A system can hold the record and control who sees it, and that is genuinely better than a text thread, but the judgment about which staff should have which codes stays with the owner and the client. Treat the tooling as a place to keep the decision, not a substitute for making it.
Physical keys are tooling, so check them out like tooling
Cleaning companies already understand the concept when the item is a floor machine or a carpet extractor. Something valuable leaves the shop with a person, it is expected back, and somebody notices when it does not come back. Client keys deserve exactly the same treatment, and the same tool and equipment tracking habits apply almost unchanged.
Concretely, that means four things:
- Every key is an item tied to a property record. Not a loose ring in a drawer. If you cannot say which address a key opens without trying it, you do not have a key system.
- Every key is checked out to a person with a date. The custody question is "who has it right now," and it should have a one-word answer.
- Keys come back on a schedule. Daily for most operations, weekly at minimum, and always at the end of an assignment.
- Tags do not carry addresses. A key tag that pairs an address with a working key is the single worst object your company can lose. Use a neutral identifier that only means something inside your own records.
The weekly reconcile is where this pays off. Ten minutes comparing what should be in the cabinet against what is there catches a missing key while you still know who had it last, instead of six weeks later when the answer is "one of four people."
Turnover: the day somebody leaves
This is the scenario the whole system exists for. A cleaner gives notice, or does not, and within the hour you need an accurate list of every property that person could physically enter.
If access lives on property records with assignment history and key check-outs, that list is a query. Pull every property they were assigned to, every key checked out in their name and not returned, and every code they were given. You then work it down as a checklist: keys back or replaced, codes changed where the client wants them changed, clients notified where notification matters, and the record updated so the next person to look does not inherit a ghost.
If access lives in a group chat, that same list is a weekend of scrolling, guessing, and calling supervisors. And the honest problem is that you will never be certain you got all of it, which means you either overreact and rekey everything or underreact and hope. Neither is a good position to be in with a client who trusted you with a house key.
Assignment history matters here beyond the security angle. The same data that tells you which properties a departing cleaner could enter is the data that drives team performance tracking, because both questions come from the same source: who was actually at which job, when. Companies that capture one usually get the other for free.
Handing a property to a substitute without a phone call
Call-outs happen. The test of an access system is whether a substitute can cover a route they have never seen without a chain of 6:40 a.m. phone calls to whoever knows that house.
When the property record is complete and readable on a phone, the substitute reads it in the driveway: this door, this lockbox, park on the street side, the dog is friendly but stays in the laundry room, supplies are in the hall closet, do not touch the desk. That is the entire briefing, and it happens without waking a supervisor.
Photo documentation makes it better. A picture of the supply closet, the alarm panel location, or the specific gate latch removes the ambiguity that text descriptions leave behind. It also makes the record self-maintaining, because the person who solves a confusing entry once can leave the answer behind for everyone who comes after.
Access failures are a lost-visit cause you can actually count
Most cleaning owners feel the access problem as generalized friction. It becomes manageable the moment you turn it into a number.
A failed visit is not a neutral event. The crew was paid, the van burned fuel, and the revenue for that slot is gone, which is exactly the kind of quiet operating leak the SBA's guidance on managing business finances warns owners to track continuously rather than absorb. The way to make it visible is to give it a reason code.
Code the reason a visit failed as a distinct outcome: no access because the code changed, key not on the ring, gate locked, pet loose, client not home when the record said they would be. Then count them monthly by property and by cause. What almost always emerges is concentration, where a handful of addresses generate most of the failures, and a single cause dominates. That is not bad luck. That is either a stale record you can fix in five minutes or a client conversation you have been avoiding.
The billing side follows from the same data. If your agreement includes a lockout or trip fee, having the failed visit logged with a reason and a timestamp is what makes charging it a routine line item instead of an argument. Wiring that into your cleaning business invoicing means the crew's time is captured whether or not the door opened, which is the honest accounting of what the visit actually cost you.
Here is how the three common approaches compare in practice:
| Group text and key shoebox | Shared spreadsheet | Access on the property record | |
|---|---|---|---|
| Who can see a code | Everyone who ever joined the thread | Anyone with the file link | Staff assigned to that property |
| History of changes | None | Whatever someone remembered to note | Attached to the property record |
| Turnover list | Reconstructed by scrolling | Manual filter, if it is current | Pull by person, in minutes |
| Key custody | A drawer and a guess | A column that goes stale | Checked out to a person with a date |
| Substitute handoff | Phone call to a supervisor | Find the file, find the row | Read it on the phone at the door |
| Lost visits | Felt, never counted | Untracked | Coded reason, countable monthly |
Rolling it out without stopping work
You do not need a project plan. You need four weeks of small habits.
- Inventory the keys you physically hold. Every ring, every fob, matched to an address. Whatever cannot be matched gets retired, because an unidentified key is a liability with no upside.
- Fill in one record per property as you visit it. Do not try to backfill 200 accounts in an afternoon. The crew that is standing in the house today is the cheapest source of accurate information you will ever have.
- Turn on check-out for physical keys. Start the custody log the day you start, and accept that history begins now.
- Add the failed-visit reason codes. One month of data is enough to see the pattern.
- Write the turnover checklist before you need it. Draft the exact steps you will follow the day somebody leaves, while nobody is leaving.
Operators comparing platforms for this kind of work should look at how the property record and assignment history actually behave, not just the calendar. It is worth seeing where IntelliDrive OS lands against Jobber and Housecall Pro, including the cases where a competitor is the better fit for how you run.
The bottom line
Client access is a real operational asset and most cleaning companies manage it like a rumor. Putting it on the property record costs nothing except the discipline of writing it down once, and it converts three recurring problems into routine work: a substitute can cover an unfamiliar house without a phone call, a departure becomes a checklist you can finish before lunch, and lost visits stop being a vague feeling and start being a number with a cause next to it.
The companies that handle this well are not the ones with the most sophisticated software. They are the ones where the answer to "who can get into 14 Oak Street" takes ten seconds and nobody has to ask a supervisor.
Related reading: Cleaning business team performance tracking · Cleaning business invoicing software · Tool and equipment tracking for service businesses. For a complete machine-readable feature and pricing reference, see our LLM reference page.