A house cleaning cancellation policy should state the notice window (24 to 48 hours is the range to weigh), the late-change fee, what counts as a lockout, how to cancel, and that the customer agreed at booking. AI can draft the wording and reminder texts, but you set the numbers and check local consumer rules first.
The wording is the easy part, and a dozen template pages will give you a version. What they rarely show is where the numbers come from and how to remind customers so late cancellations become rare. Two decisions do more than any wording: set each fee from what a late change actually costs you, and time the key reminder before the free-change cut-off rather than just before the visit. None of this is legal advice; fees and messaging rules differ between places, so read the checks near the end before anything goes live.
What cleaning-software guides agree on, and what they leave to you
ZenMaid's cancellation policy guide for cleaning businesses is a fair summary of the common view. It calls 24 to 48 hours' notice the industry standard for residential cleaning, with 48 hours as the sweet spot, and describes the most common fee structure as no fee outside the notice window, 50 per cent of the service price for a late cancellation, 100 per cent for a same-day one, and the full charge plus travel for a no-show. It treats a lockout (missing keys, a wrong entry code, a locked gate) the same as a no-show, says a voicemail left with an individual cleaner should not count as a valid cancellation, and recommends a card on file taken at booking. To make a fee enforceable, it says the policy must be disclosed in writing before the customer books and agreed, through a signed service agreement, a checkbox at online checkout or a confirming email, with the record kept. For regulars it suggests releasing the slot after two skipped visits in a row and a deposit after three cancellations in 90 days.
That is sound advice, and the policy below keeps its structure: a free window, tiered fees, a lockout rule, a card on file and a recorded agreement. The gaps are the ones that matter in practice. A guide cannot tell you whether 50 per cent is right for your business, so the next section derives each fee from costs, and they come out lower than the guide's. It cannot tell you when the reminders should go out. And it does not test the texts, which is where AI is useful. The wording and the numbers below are an illustration for a four-cleaner residential company; replace the figures with your own.
Price the late cancellation from your own numbers
Start with a typical visit: two cleaners for two and a half hours, five cleaner-hours, billed at $42 an hour, so $210. The costs behind it (all illustrative) are wages of $18 an hour plus 15 per cent on-costs, or $103.50 for the five hours; supplies of $7.50; travel and drive time of $15; and a $21.50 share of overheads such as software, insurance and admin. Total cost is $147.50, leaving a margin of $62.50, about 30 per cent.
Now see what a lost slot costs. If the cleaners are paid for the shift anyway and nobody takes the slot, you lose the $210, less the $22.50 of supplies and travel you did not spend, so $187.50. If the team has already driven to a locked door, the travel is spent too, and the loss is $202.50. What you actually lose also depends on how often you can refill a cancelled slot, and that depends on notice. Export your last 60 or so cancellations (cancel time, visit time, refilled or not) and ask an assistant to do the sorting.
This CSV lists my last 64 cancelled cleaning visits: cancelled_at, visit_start,
refilled (yes/no). Group them by notice given: 48+ hours, 24 to 48 hours,
under 24 hours, and no access on arrival. For each group give the count, how
many were refilled, and the refill rate. Show the working. Do not estimate
anything the data does not contain.
A sample result, with the expected loss worked out against the $187.50 and $202.50 figures:
| Notice given | Slots refilled | Expected loss per cancellation | Fee that stays below it |
|---|---|---|---|
| 48 hours or more | 19 of 20 (95%) | $9.38 | None |
| 24 to 48 hours | 14 of 20 (70%) | $56.25 | 25% of the visit: $52.50 |
| Under 24 hours | 8 of 20 (40%) | $112.50 | 50% of the visit: $105 |
| Lockout, no access | 0 of 4 (0%) | $202.50 | 75% of the visit: $157.50 |
Two things about that table. First, each fee sits just under the expected loss, which is the point: a fee that looks like a genuine estimate of what the late change costs you is easier to defend than a round number picked for effect. Second, the refill data supports the window. If you refill 95 per cent of slots with 48 hours' notice, a 48-hour free window costs you almost nothing and customers will find it generous. If your own data says you refill only 40 per cent inside 24 hours, a 24-hour window is the shortest you can afford. Run the sort again every six months, because refill rates change with the season.

The Complete Excel Course with AI Integration
From formulas and pivot tables to Power Query, DAX, VBA and AI with Copilot, ChatGPT and Claude — hands-on from the first lesson.
By Sagnik Bhattacharya and Virginia Thorn, Primetime Emmy Award winner
Buy nowThe cancellation policy wording, clause by clause
Here is the policy that those numbers produce. Keep it to one screen, put a plain-English version of the first two lines beside the booking button, and make the tick box say "I have read the cancellation policy" with a link, not a tangle of small print.
CANCELLATION AND RESCHEDULING POLICY
We reserve your time slot for you alone, and late changes cost us work
we cannot always refill.
1. Free changes. Cancel or reschedule at least 48 hours before your arrival
window starts and there is no fee.
2. Late changes. Between 24 and 48 hours, your first change in any 90 days
is free; after that the fee is 25% of the visit price. Under 24 hours, the
fee is 50% of the visit price.
3. No access. If our team cannot get in within 15 minutes of arriving, we
treat the visit as cancelled and charge 75% of the visit price.
4. How to change. Use your booking link, email [address] or text [number].
Messages left with a cleaner do not count. The time we receive your
message is the time that counts.
5. If we cancel. If we have to cancel (illness, vehicle failure, unsafe
weather) there is no fee, and we will offer the first available slot.
6. Exceptions. We may waive a fee for illness, bereavement or an emergency.
Tell us as soon as you can.
7. Payment. Fees are charged to the card on file and itemised on your receipt.
8. Regular visits. Skipping more than two visits in 90 days means we may
release your regular slot.
By booking, you confirm you have read this policy.
Each clause answers a dispute you will otherwise have in a worse mood. Clause 1 uses hours before the arrival window begins, not "by the day before", because "the day before" means different things to different customers. Clause 2 gives the first late change in 90 days a free pass: it costs you little, and it turns a first offence into goodwill instead of a complaint. Clause 3 puts a number on waiting time, so the cleaner is not left guessing at a door. Clause 4 makes the booking system the official channel and says what does not count, because a text sent to a cleaner's personal phone leaves you no record. Clause 5 shows the policy cuts both ways, which matters for fairness and for how it reads to a customer.
Clause 6 is deliberately loose: a list that tries to name every acceptable excuse invites arguments about the ones it leaves out, so the clause keeps the decision with you. Clause 7 depends on a card on file, which is the single biggest help in collecting a fee without a confrontation; check that your booking or payment software can store the card securely. Clause 8 is for regulars: a weekly customer who skips a third visit in 90 days is telling you the slot is not valuable to them, and you should be free to offer it elsewhere. If you want a clause about deposits for repeat cancellers, ZenMaid's guide suggests one after three cancellations in any 90 days; in this policy the trigger would be a third late change in 90 days.
Remember that the policy is a document the customer sees next to your service agreement. Can AI write a cleaning service agreement deals with scope, access and liability, and the two should not contradict each other on access, keys or lockouts.
Lockouts: wait times, evidence and a fee that survives an argument
A lockout is the cancellation customers dispute most, because they were not told it was happening. The team arrives, no one answers, the key does not turn or the door code has changed. Three things make it manageable: a fixed routine for the team, a record of what happened, and a fee tied to a number you can explain.
The routine is the same every time. At the door, the cleaner calls the customer at once. At five minutes they text. At ten they call again. At fifteen they send the final message and leave. Every step is logged in your booking software with the time, so the record shows the team was there, tried to reach the customer, and waited. Keep the log in the booking record rather than on a cleaner's phone, and do not rely on photographs of the door or the house, which raise their own privacy questions.
Think about whose fault it is before charging. If the customer gave a code and it was correct when you confirmed it, the lockout is theirs. If your own team lost a key, or arrived at the wrong time, it is yours, and a fee would turn a small mistake into a lost customer. A short checklist, run by the person who handles the complaint, settles it: was the access method confirmed in the 72-hour text, did the customer reply, did the team arrive in the window, and was the 15-minute routine followed? If any answer is no, waive the fee and apologise.
On the amount, the sum earlier showed a real loss of about $202.50 on a $210 visit, and ZenMaid's guide charges a lockout like a no-show, at the full service price plus travel. The policy above charges 75 per cent, $157.50, which stays under the loss and is easier to defend. If lockouts become a recurring problem, you can move closer to the full price, but check the rules on fees where you operate before you do.
Drafting the policy and the texts with AI without losing your numbers
An assistant is useful for turning your decisions into clean wording and friendly texts, and risky for anything that carries a number. The rule is simple: you supply every figure, the assistant supplies the phrasing, and you verify the figures afterwards.
I run a residential cleaning company. Draft a cancellation policy of no more
than 250 words in plain English, as numbered clauses.
Use exactly these terms and do not change, round or soften any of them:
- Free changes with 48 hours or more notice (before the arrival window)
- 24 to 48 hours: first change in any 90 days free, then 25% of the visit price
- Under 24 hours: 50% of the visit price
- No access within 15 minutes of arrival: 75% of the visit price
- Our cancellations: no fee, first available slot offered
- Fees charged to the card on file; skipping more than two regular visits in 90
days means we may release the slot
After the policy, output a table of every number and condition you used so I
can compare it with this list.
The table at the end is the control. In one realistic run, the clause for late changes came back saying fees "may apply" under 24 hours instead of "the fee is 50%". That softening is a classic assistant habit, and it would have left the policy impossible to enforce. A second run changed "the first change in any 90 days" to "the first change", which is more generous than intended and which nobody would notice without the check table. Compare each row with your list, fix the draft, and only then move on to the texts.
For the texts, you give the assistant the policy, the tone you want (warm, short, never threatening) and the constraints that apply to text messages. Each message should be 160 characters or fewer, use plain characters only, address the customer by [first name], and contain no emoji. The next section explains why those limits matter.
The reminder sequence: nine texts and when each one goes
The best cancellation policy is one that rarely has to be used, and reminders do most of that work. The sequence below is built around the cut-off, not the visit: the key reminder arrives while the customer can still change the booking for free. It uses a Thursday 9:00 visit as the example, so the free-change cut-off is Tuesday at 9:00 and the 50 per cent fee starts on Wednesday at 9:00.
| # | When it goes | Purpose | Length with typical values |
|---|---|---|---|
| 1 | At booking | Confirm the slot and the free-change deadline | 126 characters |
| 2 | 72 hours before (Monday morning) | Check how the team will get in | 139 |
| 3 | Monday afternoon, about 64 hours before | Free changes close tomorrow at 9:00 | 133 |
| 4 | Morning of the visit | Arrival window and access reminder | 139 |
| 5 | 20 to 30 minutes before arrival | "On the way" | 55 |
| 6 | Five minutes after arrival, if locked out | Ask the customer to call | 126 |
| 7 | Fifteen minutes after arrival, if still locked out | Say the team is leaving and what happens next | 154 |
| 8 | Same day as a late cancellation | Confirm the fee and offer a new slot | 145 |
| 9 | Same day, for a first late change you waive | Waive the fee and ask for notice next time | 143 |
Two timing rules hold the sequence together. Send reminders only in daytime hours; schedule text 3 for the afternoon before the cut-off, not the small hours. And count the cut-off from the arrival window's start, so the text is accurate for early and late visits alike. On a first visit, texts 2 and 3 land on the same day, so send 2 in the morning and 3 in the afternoon. With a weekly customer you can drop text 2 after the first visit, because access will not have changed, and keep texts 3, 4 and 5.
Here are the nine texts written out. Brackets mark what your software fills in. The lengths above were counted with a five-letter first name, a 12-character business name, short dates and times, a 20-character link and a 12-character phone number; text 7 is the one to watch, because a longer link pushes it past 160.
1. Hi [first name], your clean with [business] is booked for [day] [date], [window]. Free changes until [cut-off]: [link]
2. Hi [first name], [business] here. Cleaning [day] [window]. How will we get in? Reply 1 if you will be home, 2 for key on file, 3 for door code.
3. Hi [first name], free changes to [day]'s clean close tomorrow at [time]. Later changes may carry a fee. Change it here: [link]
4. Hi [first name], your [business] team arrives in your [window] window today. Please make sure we can get in. Call [number] with any questions.
5. Hi [first name], [cleaner] and the team are about 20 minutes away.
6. Hi [first name], we are at your door and cannot get in. Please call [number] in the next 5 minutes so we can still do your clean.
7. Hi [first name], we waited 15 minutes and could not get in, so we have to leave. Today's visit is charged under our policy: [link]. Call to rebook.
8. Hi [first name], we have your cancellation for [day]. It falls inside our 24-hour window, so the 50% fee ([fee]) applies. Rebook: [link]
9. Hi [first name], as this is your first late change we have waived the fee. Please give us 48 hours next time so we can offer the slot to someone else.
Several details in those texts are deliberate. Text 3 says a fee may apply, rather than naming one, because a customer's first change made 24 to 48 hours ahead is free; if your software can see whether that pass has been used, let it name the exact fee instead. All nine are under 160 characters and use only plain characters, and that matters for cost. Twilio's glossary page on SMS limits says a single text holds 160 characters in the standard encoding, and a longer message is split into segments of 153, each billed on its own. If a message contains an emoji or a curly quotation mark, the encoding changes and the limits drop to 70 characters for a single text and 67 per segment. A text that looks short can silently become three billed messages, which is why smart quotes pasted from a word processor are the usual culprit. Twilio publishes a segment calculator on that page; paste each text into it before you load the sequence.
Some providers add opt-out wording to the first message for you; if yours does not, about 22 more characters for "Reply STOP to opt out" still keeps text 1 under the limit. The first text is also where you ask for permission to send the rest, so build that into the booking form and keep the texts themselves about the booking. Do not slip a discount offer into a reminder; promotional messages are generally treated differently from booking messages, and the rules about consent differ between the two. Drafting texts with AI and then checking consent rules goes into the difference.
Text 2 invites a reply of 1, 2 or 3. That is for your benefit: a one-digit answer is easy for software to read, and reply 3 (door code) should trigger a person to call, not an automatic message. Do not ask customers to send door or alarm codes by text, and do not paste real ones into a chat assistant. Use a note field in the booking record that only your team can see.
Which software sends the texts, and what to ask before you pay
You do not need to send these by hand. Cleaning and field-service software can send them from the booking record. ZenMaid's website says it confirms appointment details with the client, sends reminders to clients and cleaners before the clean, and follows up afterwards, by email and text. Its pricing page lists three plans, Starter at $19 a month (up to 40 appointments a month and a limited set of message templates), Pro at $39 and Pro Max at $49 (marked as coming soon when checked), with more templates on higher plans and a note that SMS charges are not included. The page also has a price calculator by team size, so check your own figure. Jobber's pricing page lists appointment reminders and "on my way" text messages among its client communication features; which plan includes them is a point to confirm on the page for your plan. A side-by-side comparison of the two for a cleaning business is in ZenMaid versus Jobber for a cleaning business.
Whatever you choose, ask four things before paying. Does it keep a time-stamp for each message and each cancellation, so you hold the record described earlier? Can it hold a card on file securely? Can you edit the wording of each automated text, and does it let you set the send time relative to the visit? And what does each text cost on top of the plan? Then test it: book yourself in as a customer, trigger all nine messages (or have the vendor show them), and check what arrives, when it arrives and how long each one is.
Exceptions, waivers and the awkward replies
The policy will meet someone with a good reason, and how you reply decides whether they remain a customer. Decide the waiver rules in advance, so the person on the phone is not inventing them: the first change made 24 to 48 hours ahead in any 90 days is free; illness, bereavement and emergencies are waived with a short explanation; your own cancellations are never charged; and weather that makes the journey unsafe for your team is waived on both sides. After that, you use discretion, and you note each waiver in the booking record.
AI is useful for drafting the replies to customers who push back. Here is an example. A customer cancelled at 8 pm the evening before a 9 am visit and writes: "This is ridiculous, I had a work emergency, surely you can't charge me." The first draft from an assistant, given the policy and the facts, was:
Our policy clearly states that cancellations under 24 hours incur a 50% fee. As you agreed to this when booking, the fee of $105 will be charged to your card.
Technically correct, and likely to lose you the customer. After a second instruction ("keep the fee decision with me, be warm, offer a way forward") it produced something closer to what a person would write:
Hi [first name], I'm sorry to hear about the emergency at work. Because we only had about 13 hours' notice, the cancellation falls inside our 24-hour window, and the 50% fee would normally apply. As this is your first late change with us, I'm happy to waive it this once. If you would like to rebook, I can offer Thursday at 9 or Friday at 1.
Check one thing before sending: the facts. The draft says "about 13 hours' notice", so compare it with the time-stamp on the cancellation (8 pm to 9 am is 13 hours). The assistant should never choose to waive a fee; you decide that, and then you tell it what you decided. And if the same customer cancels late a third time in 90 days, the deposit rule applies and the reply should say so kindly and plainly.
Before it goes live: consumer rules, a read-through, a test booking
This is not legal advice, and the rules on cancellation fees and customer texts differ widely between places, so check the points below where you operate. Bring the policy, the cost table and a screenshot of the booking page to a local adviser so they can review the real flow rather than a loose description.
- How fees are judged. In many places a late fee has to reflect a reasonable estimate of your loss rather than work as a penalty. The cost table above is your evidence; keep it.
- Cooling-off rights. Some places give consumers a statutory right to cancel services booked remotely or at their door. That can override a notice window, especially for a booking made weeks ahead.
- Disclosure. Show the fees before the customer commits, in the booking flow, not only in terms and conditions.
- Texting. Check what consent and opt-out your text provider and local rules require, and honour a STOP reply immediately.
Then test the wording on real readers. Ask three people outside the business to read the policy and tell you, in their own words, what would happen if they cancelled with 20 hours' notice. If any answer differs from yours, the clause is unclear and needs rewriting. You can also ask an assistant to read the policy as a sceptical customer and list every point that is unclear or feels unfair; treat its list as points for you to weigh, not a ruling.
Finally, run a test booking from the tick box through all nine texts and check the records it leaves. Announce the policy to existing regulars with a month's notice, in the tone of a courtesy, using the approach in the cleaning company price increase letter tutorial, which covers segmenting customers and softening a change.
After a month, review the numbers. How many late changes came in, how many were waived, and what did you charge? Suppose reminders cut late changes in a 120-visit month from 11 to 6, roughly 9 per cent down to 5 per cent. Five fewer late changes, valued at the $112.50 expected loss of an under-24-hour cancellation, is up to about $560 a month you no longer lose, or less if some of them would have come 24 to 48 hours ahead, at $56.25 each. That comparison, more than the fees collected, is the measure of a good policy; cutting no-shows with reminders and automatic rebooking takes the measurement further. Loading the fee table and all nine texts into your own booking software is something my AI implementation consultation can help with.
Questions cleaners ask about cancellation policies
How much notice should a cleaning cancellation policy require?
ZenMaid's guide for cleaning businesses calls 24 to 48 hours the industry standard for residential cleaning. Pick yours from data: look at how often you refill a slot at each notice length. If you rarely refill inside 48 hours, a 48-hour window is justified; if you often refill within a day, 24 hours is fairer to customers.
What if a customer says they never saw the policy?
That dispute is why you ask for agreement at booking and keep the time-stamp. Send the policy again in the confirmation and the day before. If you cannot show that the customer saw and accepted it, waive the fee, apologise, and fix the booking flow so the next customer cannot proceed without ticking the box.
Should fees be charged to the card on file automatically?
Only if the agreement says so, and keep a human step before any charge: check the time-stamp, the notice given and whether a waiver applies. Send a receipt that states the reason and the policy clause. An automatic charge that was wrong costs far more in goodwill than the fee is worth.
Are automated reminder texts allowed without permission?
Rules on texting customers vary and are changing, so do not assume. Ask for permission to text at booking, keep reminders to the booking itself, include an opt-out, and honour it at once. Your text provider's help pages will say what it requires, and a local adviser can confirm the rest.
Further reads
- House Cleaning Estimate Template: Quote by Home Size With AI — The estimate that comes before the booking and the policy.
- Cleaning Invoice Template: Itemise Visits and Chase Payment With AI — Invoice the fee clearly and chase it politely.
- Can AI Help a Cleaning Company Rebook More Regular Clients? — Keep regular clients booked after a cancellation.
- Cleaning Business Management Software: Scheduling, Timesheets and AI — Scheduling and timesheet software that holds the records.
- Maid Service Scheduling Software: Can AI Handle Teams and Keys? — Teams, keys and recurring visits in one schedule.
- Dog Boarding Cancellation Policy: Peak-Season Wording With AI — See how another service handles peak-season cancellations.
- Riding Lesson Cancellation Policy: Weather, Welfare and Late Notice — Write the weather, horse-welfare and late-notice clauses for a riding school, plus credit rules and the messages you will actually send.
- How to Get House Cleaning Clients: A Local Plan Built With AI — A four-week plan to win house cleaning clients: Google profile, neighbourhood groups, referral wording, a costed first-visit offer, reviews and AI drafts.
- Price House Cleaning Jobs: Build a Calculator With ChatGPT — One cost engine, three quote styles: hourly, per-room and flat-rate house cleaning prices, with first-visit uplift, deep cleans and a margin check.
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: ZenMaid's guide to cancellation policies for cleaning businesses and its pricing and features pages; Jobber's pricing page; Twilio's glossary page on SMS character limits; all read in November 2026. Fee and refill figures are illustrative. This is not legal advice.