Domestic cleaning · 3 staff, 85 contractors · a system first, then the incentives
Scheduling, routing and incentives
Open a part, then hover a box for detail or a button to see it working.
AutomatedAIPersonYes or noSourceDataResult
The whole week, from enquiry to bonus · hover anything, or click it to open that part
A · Package offer
Word of mouth, and later a paid Instagram funnel, bring enquiries into a CRM.
A person prices a package of promises in place of an hourly job:
Weekly, at a set time
Defined outcomes for the house
All materials, and on time
A free return clean if the client is unhappy for any reason
B · Match + route
Each job is booked against a property type and a time band, then matched to a cleaner on geography and rate, with travel time inside the model.
The week holds hours for every cleaner and standing cover together, shown on a map.
C · Calendar sync
The schedule lands in each cleaner's own phone calendar, and the changes they make there sync back.
No cleaner had to adopt new software.
D · Clean + survey
A clean to defined outcomes at a fixed price, then a 1 to 10 survey after every job.
8 or above: a good clean, and a review request by email
Under 8: kept internal
E · Bonus settled
Time per clean and quality pay together, with a bonus for more than 3 jobs a day.
A return clean is paid, and it eats into the bonus.
F · Cover called
More than 24 hours notice of illness earns an incentive, and the slot refills from the backup pool held in the week.
85 contractors
85 self-employed cleaners, working alone inside houses, with no line management over any of them.
The schedule reaches each one through the calendar app already on their phone, and their edits sync back.
3 staff
The founder and two staff run the dashboard, the schedule, the map and the bookings.
They are the only people who touch the system.
Return clean?
If the client is unhappy for any reason, a free return clean, with no arbitration step.
It is paid to the cleaner, and it eats into their bonus.
Word of mouth
Word of mouth stayed the main channel.
Word-of-mouth revenue increased because of the offer that was created.
The packageWeekly, at a set timeDefined outcomesAll materialsOn timeFree return clean
Per cleanerGeographyRateTravel timeHours per weekStanding cover
No: kept internal
Settled fromTime per cleanQuality scoreJobs a dayReturn cleans
No: outside the pool
Yes: netted off the bonus
No: a morning call
APackage offerAn hourly job against the package, as a client would see it. Illustrative
The same clean, sold two waysillustrative quote · the five promises are the real package
BeforeAn hourly job
What the client buys
Hours of cleaning
A longer clean
A bigger bill
A faster clean
A cheaper clean, so the business bids against itself
The deliverable
A count of hours
▲Revenue+10–15%weekly sales, from the new package offer
▲QualityGrowing againclient base, quality no longer the bottleneck
▶
AfterThe packageWeekly clean · 3-bed house · Tuesdays, 11:00
✓Weekly, at a set timeThe same slot every week. Here, Tuesdays at 11:00.
✓Defined outcomesThe state of the house is the deliverable, room by room.
✓All materials broughtNothing for the client to buy or leave out.
✓On timeArriving when the booking says.
Not happy, for any reason?A free return clean.
Example outcomesKitchen surfaces, hob and sinkBathrooms and toiletsFloors hoovered and moppedDusted, top to bottom
Price per visitFixedset by a person, for this house
The move from hours to outcomes is what makes the rest possible. Once the client buys a result at a fixed price, speed becomes the company’s gain, and the quality bar is what stands between speed and a poor clean. The two tiles are the real results.
BMatch + routeThe cleaners, a job being matched, and Tuesday on the map. Synthetic data
The dashboard, one weeksynthetic · 10 of the 85, codes in place of names
Cleaner
Patch
Rate
Wants
Booked this week
Cover
C-03
N1
R2
25h
22h
Tue
C-07
N2
R1
20h
20h
C-11
N2
R3
30h
27h
Fri
C-14
N2
R2
40h
40h
C-18
E1
R1
25h
25h
C-21
E1
R2
15h
13h
Tue
C-26
S1
R2
30h
28h
C-30
S1
R3
20h
17h
Thu
C-35
W1
R1
25h
22h
Mon
C-42
W1
R2
20h
20h
Wed
Matching a new job
Tue 11:00 · 3-bed housePatch N2 · 3 hour time band
C-14In patch N2Free at 11:008 min from the 08:00 cleanRate R2Matched
C-11In patch N2Free at 11:0021 min awayRate R3Next in line
C-07In patch N2Booked until 12:30Busy
C-03Patch N1On cover TuesdayHeld as cover
Geography says who can get there. Rate says what the job costs to staff. The journey from the previous job counts in.
Each cleaner carries a patch and a rate band. The bars show hours booked against the hours each one wants, and the cover column is the standing cover held in the week. Rate bands stand in for real rates.
Tuesday on the mapillustrative · an invented town, synthetic routes
Routes render on a map, so coverage and gaps read by geography. Travel sits inside the model, and the cleaners on cover are held free within reach. Hover a route or a cover marker for detail.
CCalendar syncOne cleaner's Friday on their phone, and an edit syncing back. Illustrative
One cleaner’s Friday, both waysillustrative · one synthetic cleaner
DashboardC-14 · this week
Day
Jobs
Booked
From the phone
Mon
3 cleans
8h
Tue
3 cleans
8h
Wed
4 cleans
9h
Thu
4 cleans
9h
Fri
2 cleans, 1 return
6h
Unavailable 15:30 to 17:00synced back
The 3 staff see the week as the cleaner sees it, edits included.
The schedule lands in the calendar app the cleaner already carries, so nobody had to adopt new software. A change made on the phone, here an afternoon marked unavailable, syncs back to the dashboard.
DClean + surveyThe 1 to 10 survey after a clean, and where each score goes. Synthetic scores
One question after every cleanillustrative survey · synthetic scores
After today’s cleanHow would you rate today’s clean?
12345678910
110
Send
One question, sent after every clean.
C-14’s scores this week
A short survey in the shape of an NPS question. The threshold is published, so every cleaner knows exactly where the bar sits. A good score also asks the client for a review, which is Peter’s own eight-and-above pattern.
EBonus settledOne cleaner's week, settled, and why the pair holds. Synthetic data
One week, settledsynthetic · one cleaner, counts only
Mon
in band · 9in band · 8over · 9
3 jobs
Tue
in band · 9in band · 10in band · 8
3 jobs
Wed
in band · 9in band · 8in band · 10in band · 7
4 jobsvolume day
Thu
in band · 8in band · 9in band · 9in band · 6
4 jobsvolume day
Fri
in band · 10in band · 9return clean
2 jobs
in band · 8+ qualifies over ran over the band under 8 outside the pool return clean netted off
Settlement · C-14 · this week
Cleans16
Qualifying, in band and 8 or above13
Volume days, more than 3 jobs2Wed and Thu
Return cleans, paid and netted off1
Thursday’s fourth job scored 6 and the client was unhappy. Friday carried the free return clean, paid to C-14 and taken back out of the volume bonus that job earned.
The bonus settles in counts here, because no amounts are on record. The rules are the real ones: time and quality together, a bonus above 3 jobs a day, and a return clean that eats into it.
Why the pair holdsillustrative · synthetic cleans
Each dot is one clean. Time per clean and quality are held as one measure, so a clean counts only in the green box: inside its time band and scored 8 or above.
FCover calledA day's notice against a morning call, for the same booked clean. Illustrative
The same booked clean, two waysillustrative · one synthetic sick day
Penalising absence backfires three ways at once: people clean while ill, people stop calling and simply fail to turn up, and across 85 contractors it starts a fight the business can do without. So the design pays for the information, because early notice is what the business needs.
Enquiry in
Enquiries arrive two ways:
Word of mouth, the main channel
Paid Instagram advertising, added later, with a funnel behind it
Word of mouth is the best lead and the least controllable. The funnel is a tap the business can turn on when the week looks light.
Into the CRM
Each enquiry lands in a CRM, and automation carries it from enquiry through to booking.
Package priced
A person settles what the package is worth to that client and quotes it.
Pricing stayed a human judgement. The system carries the rate per cleaner and the time band per job.
Package quoted
The old shape was an hourly job. The new shape is a set of promises:
Turning up weekly, at a set time
Defined outcomes for the house
All materials brought
Arriving on time
A free return clean if the client is unhappy for any reason
While a client buys hours, a faster clean is a cheaper clean. Once they buy a result at a fixed price, speed becomes the company's gain.
Sales up
Weekly sales up 10 to 15% after the move from hourly jobs to a fixed package.
Word-of-mouth revenue increased because of the offer that was created.
Word of mouth
Word of mouth stayed the main channel.
Word-of-mouth revenue increased because of the offer that was created.
Job booked
The accepted job enters the schedule against a property type and a time band.
Cleans run 2 to 4 hours, on 2 and 3 bedroom houses, top to bottom.
Cleaner record
Every cleaner on one screen, carrying the two fields that make a schedule solvable:
Geography: who can physically get there
Rate: what the job costs to staff
Google Sheets
Cleaner matched
The dashboard matches a cleaner on geography and rate, with the journey from their previous job counted in.
Google Maps
Travel sits inside the model, which keeps the hours-per-week number honest.
Week built
Two aims held at once, and they pull against each other:
Hours per week: a full week for every cleaner who wants one, inside their own patch and at their own rate
Standing cover: somebody within reach of every booked slot is free to take it
A week built only for hours has no slack in it. A week with comfortable slack wastes hours. This one holds both.
Read on a map
Routes render on a map, so the founder and two staff read coverage and gaps by geography, patch by patch.
Google Maps
It replaced a spreadsheet and manual admin.
The week
The built week on the dashboard: every job, its address, its time and its cleaner.
Pushed out
Each job is pushed into the cleaner's own calendar, on their own phone, in the app they already use.
Google CalendarZapier
Cleaner's phone
85 self-employed cleaners, working alone inside houses, with no line management over any of them.
No cleaner had to adopt new software. The thing they use is the calendar they already had.
Cleaner edits
A change the cleaner makes in their own calendar syncs back into the system.
Edits sync back
Changes made on the phone sync back into the dashboard, so the 3 staff see the week as the cleaners see it.
Google Calendar
Clean delivered
The cleaner works to a defined set of outcomes for the house at a fixed price.
The deliverable is the state of the house, where it used to be a count of hours.
Survey out
A short survey follows every clean: one question, rated 1 to 10, in the shape of an NPS question, on a form on the company's own website.
Quality has a number, it comes from the client, and it arrives after every single job.
8 or above?
The threshold is published, so every cleaner knows exactly where the bar sits:
8 to 10: a good clean
1 to 7: stays internal, with whoever owns the client relationship
Review request
A good score triggers a review request by email, at the moment the client is most willing to write one.
The eight-and-above review request is Peter's own pattern, run at more than one client. An unhappy client is never asked for a review.
Quality logged
Each score sits against the clean and the cleaner, as the quality half of the bonus.
Time + score
Each clean carries two numbers:
Time: did it land inside its time band?
Quality: the client's score, 1 to 10
Both held?
A clean qualifies only when it lands inside the time band and the quality holds.
Pay on speed alone and quality falls: skirting boards get skipped
Pay on quality alone and jobs overrun: the next client waits
The two failure modes point in opposite directions, so there is no direction left to cheat in.
Volume bonus
A bonus for more than 3 jobs a day.
The business asks for throughput with money, in place of pressure.
Return clean?
If the client is unhappy for any reason, the return clean happens, free to the client, with no arbitration step.
It is paid, so nobody redoes a house for free
It eats into the bonus, so the cost lands where the decision was made
A rushed fourth job collects the volume bonus, and the return clean takes it back.
Bonus paid
Time and quality pay together, with the volume bonus and any return clean netted off.
The incentive to do more jobs is real, and it cannot be funded by doing them badly.
Cleaner ill
A cleaner needs to call in ill.
Penalising absence backfires: people clean while ill, or stop calling and simply fail to turn up.
24h notice?
More than 24 hours notice earns an incentive.
The business needs to know early, so early notice is what gets paid.
Backup pool
Cover is a standing requirement, settled when the week is built, so somebody within reach of every booked slot is free to take it.
Slot refilled
The slot refills from the backup pool, and the cover job reaches that cleaner's own calendar like every other job.
Google Calendar
A day's warning is a slot refilled, a client who never finds out, and a cleaner who still gets hours somewhere else in the week.
Fewer empty slots
Empty calendar time roughly halved, from about 17% of cleaner hours to 8%.
Clients still booked a year later rose from about 6 in 10 to 7 in 10, after overruns and missed slots had been driving them away.
Sell · Deliver · Train
Every engagement has the same shape:
Sell: finding the real problem, and winning the work
Deliver: building what fixes it
Train: handing it over so the team runs it
When a client gets the value they want, they're more inclined to do more.
Once they see what is possible, they bring their own ideas to the table, ready to be implemented.