Cleaning company quote to booking, end to end
Quote request to booked job for a cleaning company: instant acknowledgement, the questions that let you price without a site visit, and the follow-up ladder.
What you get from this guide
- The full path from quote request to booked job, with the decision points named
- An instant acknowledgement that qualifies instead of just saying thanks
- The six questions that price most domestic cleans without a site visit
- Explicit rules for when a human has to price it, so the automation never guesses
- A three-touch follow-up ladder for quotes that go quiet
The shape of the funnel
A domestic cleaning enquiry has a short half-life. Somebody decides on a Tuesday lunchtime that they want a cleaner, fills in three or four forms in ten minutes, and books whichever company gives them a price they can act on. Nothing else in this funnel matters as much as being that company.
That makes the cleaning quote-to-booking problem unusually tractable, because the whole path is short and most of it is rules rather than judgement:
| Stage | What happens | Automated or human |
|---|---|---|
| Capture | Enquiry arrives from a form, phone, Facebook, Google or WhatsApp and lands in one queue. | Automated. Every channel, one record shape. |
| Acknowledge | Reply within a minute naming the job, setting expectations and asking the first question. | Automated. Speed is the point. |
| Qualify | Six questions that determine whether and how you can price it. | Automated, on the channel they contacted you. |
| Route | Standard job to instant price, unusual job to a person. | Automated decision, explicit rules. |
| Price | A number, what is included, what is not. | Automated for standard work, human otherwise. |
| Book | Real availability, job written into the schedule, confirmation sent. | Automated. |
| Follow up | Three touches over ten days on anything unaccepted. | Automated, stops on any reply. |
Worth being clear about what this is not. It is not a chatbot that talks to customers about cleaning, and it is not a replacement for the owner's judgement on unusual work. It is a pipeline with two automated ends and a deliberately narrow band of automated decision in the middle, and the value is concentrated in the speed of the first reply and the fact that nothing is ever dropped. Those are boring properties. They are also the ones that show up in the booking numbers.
The two places this breaks are the ends. At the top, enquiries arrive in five different inboxes and nobody owns them, so the clock starts late. At the bottom, quotes that go quiet are never chased because the owner is on a job. Both are exactly what automation is for. Sector context and the other workflows worth building are on the cleaning company automation page.
The instant acknowledgement
The first message is not a receipt. "Thanks, we will be in touch" is a wasted send: it does not move the enquiry forward and it does not distinguish you from the auto-responder the other three companies also have. A good acknowledgement does three jobs at once.
- Names the job in the customer's own words, so it reads as a person who looked rather than a machine that fired.
- Sets a specific expectation - a price in the next few minutes if you can price it, a callback window if you cannot.
- Asks the first qualifying question, so the conversation starts immediately instead of waiting for a form.
Acknowledgement - SMS, sent within a minute
Hi {first_name}, {name} from {business} here -
thanks for the enquiry about a {clean_type}
in {area}.
I can usually get you a price in a couple of
minutes. Quick one to start: how many bedrooms
and bathrooms is the property?Reply on the channel they used. Someone who messaged your Facebook page at 9pm does not want an email, and someone who called wants a text back on the number they called from. That last case is its own build - see missed call text-back with n8n and the missed call text-back entry - and for a cleaning company it is frequently the highest-value single automation available, because a missed call at 4pm on a weekday is a client who is already ringing the next company on the list.
Speed is the mechanism, not a growth hack. The reason a fast reply wins is simply that the enquiry was sent to several companies and the person is still holding their phone. Once they have a price from someone else, you are no longer answering an enquiry, you are asking them to reconsider a decision. The lead response time and speed to lead entries cover the concept properly. If you send automated SMS from a US ten-digit number, register under A2P 10DLC first or the messages will be filtered rather than delivered - the A2P 10DLC guide covers it.
Questions that let you price remotely
Here is the practical core of this page. Six answers price the large majority of domestic cleaning work, and every question beyond them costs you completed enquiries.
| Question | Why you need it | What it changes |
|---|---|---|
| Bedrooms and bathrooms | The standard proxy for property size across the industry. | Base duration. Bathrooms drive time more than bedrooms do. |
| Type of clean | Regular, one-off deep, end of tenancy, post-build, commercial. | Whether it can be priced automatically at all. |
| Frequency | Weekly, fortnightly, monthly or one-off. | Rate per visit and the lifetime value of the client, which may justify a keener first price. |
| Postcode or suburb | Service area and travel time. | Whether you quote at all, and any travel loading. |
| Condition or last cleaned | The single biggest source of underquoting in this industry. | A first-clean loading, or a route to a human quote. |
| Pets and access | Dogs, cats, keys, key safes, someone home, parking. | Duration, cleaner allocation, and whether the job can actually be delivered. |
Ask them one at a time in a conversation rather than as a ten-field form where the channel allows it. A back-and-forth over text has much better completion than a long form, it feels like talking to the company rather than filling in paperwork, and it lets you skip questions that stop being relevant - there is no point asking about frequency once someone has said end of tenancy.
The condition question is the one owners leave out and the one that costs them. "When was it last professionally cleaned?" is a gentler way of asking how bad it is, and the answer "never" or "a couple of years ago" should change your number or route the job to a person. A cleaning company that quotes a standard rate on a property in poor condition either loses money on the job or has an uncomfortable conversation on the doorstep, and both are worse than one extra question.
Let a model read, but not decide
What the machine can price
Your pricing model probably already exists, in the owner's head. The work is writing it down in a form something else can execute. For most domestic cleaning businesses it decomposes into four parts:
- A base duration from bedrooms and bathrooms - your own numbers, from your own completed jobs, not an industry table.
- Multipliers for clean type - a deep clean against a regular clean, and so on.
- Loadings for condition, pets, travel outside the core area, and a first-visit uplift on a regular contract.
- A frequency rate - what a weekly visit costs against a fortnightly or a one-off.
Derive the base durations from your actual job records rather than from a spreadsheet you built in year one. If your scheduling software logs start and finish times, you already know how long a three-bed two-bath regular clean really takes your team, and it is quite often not what you have been quoting. That analysis is frequently worth more than the automation on top of it.
Present the number with its assumptions attached, because an unqualified price is what produces a dispute on the day. What is included, what is not, roughly how long the team will be there, and one line making clear the price assumes the property is in the condition described. Then a booking link, which should open on real availability rather than a contact form - every step between the price and the calendar loses jobs.
Frequency deserves its own thought, because it is the one field that changes what the client is worth rather than what the job costs. A fortnightly regular is a recurring relationship worth many times a one-off deep clean, and plenty of cleaning companies price both from the same table without ever acknowledging that. You do not have to discount the recurring work - but you should know, when you set the rate, which of the two you are actually competing for, and your automation should be able to tell you afterwards how the split came out by channel.
The first-visit uplift on a regular contract is worth building in explicitly rather than leaving to the cleaner to discover. The first clean of a property that has not been done professionally takes longer than every subsequent visit, and a company that quotes the steady-state rate for visit one has effectively bought the contract with an hour of unpaid labour. Make it a named line in the price message - a one-off first-visit charge, with the ongoing rate stated underneath - and the customer understands it immediately, because it matches their own expectation of the state of the house.
One last thing on the price message: do not bury it. The number goes near the top. A customer who has to read four paragraphs to find out what it costs will scroll for the figure anyway, and making them hunt reads as evasive.
When a human must price it
Automating what should not be automated is more damaging in this vertical than in most, because an underquote is not just lost margin, it is a cleaner sent to a job that cannot be done in the time booked. Write the escalation rules down explicitly, put them in front of the automated pricing, and make them obvious enough that anyone in the business can explain them.
| Trigger | Why | What the automation does instead |
|---|---|---|
| End of tenancy or post-build | The price depends on the state of the property, which you cannot see. | Acknowledge, gather details and photos, book a call or a visit. |
| Condition described as very poor or never cleaned | Duration is unpredictable and often double a standard clean. | Ask for photos, then route to a person with them attached. |
| Commercial or multi-site | Different pricing model, contracts, insurance and access. | Route to the owner with a note. Never quote it automatically. |
| Property above your size threshold | Large homes scale non-linearly and often need a team. | Acknowledge, escalate, set expectation of a call today. |
| Outside the service area | Travel may make it uneconomic or undeliverable. | Say so honestly and quickly rather than going quiet. A fast honest no earns referrals. |
| Anything the extraction was unsure about | A confident price on a misread enquiry is worse than a question. | Ask the clarifying question. One extra message beats a wrong number. |
This is human-in-the-loop done properly: the human is not reviewing everything, they are handling the defined set of cases where judgement genuinely adds value. Everything routed to a person should still get an instant acknowledgement with a real time expectation, because the customer should never be able to tell which path they landed on from how long the first reply took.
Then review the exception queue every week. Each quote a human touched is either a rule you could add or a case that genuinely needs a person, and sorting them into those two piles is how the automated share climbs over time. Most cleaning companies find the automated share settles somewhere well short of everything, and that is the correct outcome rather than a failure.
The follow-up ladder
A quote that gets no reply is not a no. It is usually a person who got distracted, or who is waiting on the other two quotes they requested. Three touches over about ten days is right for domestic work - the decision cycle is short, and a ladder that runs for a month is chasing someone who booked a competitor two weeks ago.
| Touch | Timing | Job of the message |
|---|---|---|
| 1 | Next day | Check the price arrived and invite a question. Catches the quote that went to spam. |
| 2 | Day 4 | Answer the objection you get most often - usually what is included, or availability - without repeating the ask. |
| 3 | Day 10 | Close it out. Say you will stop, note how long the price holds, keep the door open. |
Touch 2 - day 4 (SMS)
Hi {first_name}, no pressure on the {clean_type}
quote - one thing people usually ask me:
that price covers {included_summary}, and the
team brings all products and equipment. Ovens
and windows are extra if you want them.
If {quoted_slot} does not suit, I have other
slots this week: {booking_link}Every touch stops on any reply, any booking, and any opt-out, and the check happens immediately before each send rather than only when the signal arrives - otherwise a customer who books online at 10:58 gets a chasing text at 11:00. The general quote follow-up guide covers the state table, the stop logic and the copy rules in full, and unsold estimate follow-up covers the same machinery for higher-value contractor work where the ladder runs for a month rather than ten days.
The build, step by step
Get every channel into one queue
Web form, phone, Facebook and Instagram messages, Google business messages, WhatsApp, email. Each one writes the same record: name, contact, channel, raw enquiry text, timestamp. Until this exists you have no funnel, you have five inboxes and a hope. Use a webhook where the channel offers one and polling where it does not.
Fire the acknowledgement immediately
Target under sixty seconds, on the channel they used, naming the job. Put a business-hours guard on it only if your out-of-hours message says something different and useful - a 10pm enquiry should still get an instant reply, it just should not promise a call in five minutes.
Extract what the enquiry already told you
Most enquiries contain half the answers. Run the free text through extraction into your six fields, mark each field as confident or not, and only ask about the ones you are missing. Asking someone their bedroom count when they wrote "3 bed semi" in the form is the fastest way to look automated.
Run the escalation rules before the pricing rules
Check the escalate list first, deliberately. If any rule matches, create the task, notify whoever prices those jobs, and send the customer a real expectation. The order matters: pricing first and escalating second is how a post-build clean gets quoted at a regular rate for two seconds before someone catches it.
Price from a rules table, not a prompt
Base duration, multipliers and loadings in a table you can read and edit without touching the workflow. A deterministic rules table is auditable, explainable to a customer, and it does not change its mind. Keep the language model to reading the enquiry and writing the sentence around the number.
Send the price with a booking link on real availability
Price near the top, inclusions and exclusions under it, duration estimate, and a link that opens straight into bookable slots. If the link shows availability you cannot honour, this whole build makes you look worse than a phone call did, so the calendar has to be the real one.
Turn acceptance into a scheduled job automatically
On booking, write the job into your scheduling system with address, access notes, duration and add-ons, block the cleaner's time, take the deposit or card if you use one, and send confirmation with the cancellation policy. Make this step idempotent so a double-submitted booking form cannot create two jobs, and wrap it in error handling that alerts a human if the write fails - a confirmed job that never reached the schedule is the worst outcome in this entire funnel.
Report on the funnel, not on the messages
Enquiries by channel, median time to first reply, share priced automatically, quote-to-booking conversion, and the exception queue. Messages sent is a vanity number. Median first-reply time and conversion are the two that tell you whether the build did anything, and cost per booked job is how to compare it against your advertising spend.
Where these builds go wrong
- Automating the reply but not the qualification, so you are fast and then silent for four hours.
- A fourteen-field quote form, most of whose answers the pricing never uses.
- No condition question, so standard rates get quoted on properties that need double the time.
- Automated prices on end-of-tenancy and post-build work, which is where underquoting hurts most.
- A booking link that does not show real availability, so you double-book or offer slots you cannot staff.
- Acceptance that lands in an inbox instead of the schedule.
- A follow-up ladder that runs for a month on a decision that takes three days.
- Stop checks only at event time, so someone who books at 10:58 gets chased at 11:00.
- Missed calls not covered, so half your enquiries never enter the funnel at all.
- Reporting messages sent rather than first-reply time and quote-to-booking conversion.
Cleaning is one of the better verticals for this kind of build, because the work is repeatable, the pricing is genuinely rule-shaped, and the decision cycle is short enough that speed converts directly. If you want the sector view of what else is worth automating - rota changes, recurring-clean reminders, cleaner no-shows - start with AI automation for cleaning companies. For the money side see AI automation cost, for how we scope and hand over see automation services, and if you would like it looked at properly, the free automation audit is the place to start.
FAQ
- Can a cleaning company really quote without visiting?
- For standard domestic work, usually yes, provided you ask the right questions and are explicit about what the price assumes. Regular and one-off cleans for a normal home price reliably from bedrooms, bathrooms, clean type, frequency and condition. What does not price remotely is anything where the variable is the state of the property rather than its size - end of tenancy on a place that has been neglected, post-build, hoarding, or heavy commercial. The trick is not pricing everything remotely. It is knowing exactly which enquiries you can and routing the rest to a person quickly.
- What questions should the quote form ask?
- Bedrooms, bathrooms, type of clean, frequency, postcode, and condition or last-cleaned. Then pets and access as follow-ups. That is enough to price the large majority of domestic work. Resist adding fields: every extra question costs you completions, and questions your pricing does not actually use are pure friction. If you find yourself wanting a tenth field, that is usually a sign the enquiry should be routed to a human instead.
- Should I show prices on my website?
- For standard domestic work, showing a price or a tight range is usually the higher-converting choice, because a client requesting quotes from several companies will engage with the one that answered. The reason many owners resist is real though: a published price is a ceiling in negotiation and it invites comparison on price alone. The workable middle is a self-serve instant quote for the standard jobs, with the unusual ones routed to a conversation - so you compete on speed where you can, and on judgement where the job deserves it.
- How fast does the first reply need to be?
- Fast enough that you are the first company to respond, which in practice means within a minute or two rather than within the hour. Domestic cleaning enquiries are almost always sent to several companies in one sitting, and the response that arrives while the person is still on their phone is in a different conversation from the one that arrives that evening. Lead response time is the single lever in this funnel with the clearest mechanism behind it.
- Should the price be automated or should a human always confirm it?
- Automate the ordinary and escalate the unusual, with the boundary written down as explicit rules rather than left to judgement. The failure mode people fear is an automated price that is too low on a filthy property, and the fix is not human review of everything - it is a rule that says properties described as heavily soiled, or above a size threshold, or of certain clean types, never receive an automatic number. Reviewing every quote by hand simply reintroduces the delay you were trying to remove.
- What should happen when the customer accepts?
- The job should be in your schedule before the confirmation message goes out, not after someone gets round to it. Write the job with address, access notes, duration and any add-ons, block the cleaner's time, then confirm with what is included, what to expect on the day, and your cancellation terms. A quote that is accepted and then sits in an inbox for six hours is where cleaning companies double-book themselves.
- How many follow-ups should an unaccepted cleaning quote get?
- Three across roughly ten days is enough for domestic work, because the decision cycle is short. A check that the price arrived, one message that answers the objection you expect most often, and a close-out. Anything longer is chasing a client who has already booked someone else, and the close-out message earns its place by producing replies from people who felt awkward about going quiet.
- What does this cost to build?
- The cost depends almost entirely on how many places enquiries currently arrive and whether your scheduling software has an API you can write bookings into. A single web form into a system with a decent API is a small build. Five channels, no API and a pricing model that lives in the owner's head is a project, and most of it is the pricing conversation rather than the code. We scope it in a free consult and quote a fixed price in writing rather than publishing a package price that would only be a guess.
Done for you
Want every cleaning enquiry answered in under a minute and priced while they are still deciding?
We build the whole path: every channel into one queue, instant qualification, rules-based pricing with explicit escalation, real-availability booking, and the follow-up ladder. The audit comes first, and it is free.
The consult is free, and you keep whatever we build.
Keep building
The rest of the cluster
Automate Quote Follow-Up (Build Guide + Templates)
The four-touch cadence, copy-paste templates, stop-on-reply logic, and the n8n or Make build.
Read the guideAutomate Unsold Estimate Follow-Up for Contractors
The estimate-state trigger, five touches of real copy, stop-on-reply and stop-on-acceptance detected from the system itself.
Read the guideMissed Call Text-Back with n8n (Free Build Guide)
Twilio trigger, business-hours guard, AI qualification, calendar booking, and the A2P 10DLC registration step.
Read the guideFree AI Automation Audit Template
Score every recurring task and rank the top three worth automating before you build anything.
