aimpl
← All work

Dalton's Paddock

Luxury farm-stay · Manjimup WA

  • Scheduling
  • Staff phone apps
  • Pricing engine
  • Smoobu and Xero

The farm-stay that runs itself more.

Dalton's Paddock is three luxury stays on 40 acres in Manjimup. aimpl built one operations platform that reads every booking and schedules the work around it.

  • Cleans schedule themselvesBookings create, move and cancel cleans; cleaners are told automatically
  • Guests get answeredMessaging, guides and pricing run from the same calendar
  • Works without signalThe cleaner app queues photos and sign-offs and syncs later
daltonspaddock.comThe Dalton’s Paddock website
The Dalton’s Paddock websiteThree stays on 40 acres
BusinessThree stays on 40 acres, Manjimup WA
Built305 commits in 19 days, October 2026
In daily useSince 6 October 2026
Core stackReact, Supabase, Netlify, Smoobu, Claude, Xero
The story
The situation

Four systems, lined up around guests by hand

Three stays on 40 acres. Cleaning ran through Turno and an earlier cleaner app, guest messages through Airbnb and Smoobu, nightly pricing through a separate scheduled job, and grounds work through contractors by text. None of them could see the bookings the others were working from.

Every change of booking meant someone re-checking the cleaner, the contractor and the message schedule. The owner was the integration.

01 · One calendar

Every booking, clean and task on one view

BeforeBookings lived in Smoobu, cleans in a cleaning app, tasks in text messages.

NowBookings, money, rates and messages arrive from Smoobu by webhook and an hourly reconcile. Bars are drawn to each booking’s true check-out and check-in times, with short gaps hatched. The calendar is the current picture, and everything else is scheduled from it.

daltonspaddock.comBookings, cleans and tasks on one calendar
Bookings, cleans and tasks on one calendar
02 · Cleaning

A clean created for every check-out, before anyone asks

BeforeEach clean was created and offered by hand, and moved by hand when a booking changed.

NowSmoobu webhooks and an hourly reconcile create, move or cancel cleans. The cleaner is emailed about every change. Same-day turnovers are flagged, and past cleans that have not been signed off are listed for attention.

daltonspaddock.comUpcoming cleans, same-day turnovers flagged
Upcoming cleans, same-day turnovers flagged
03 · The cleaner’s phone

Checklists, photos and sign-off that work without signal

BeforeThe old cleaner app needed signal and the owner to chase sign-offs.

NowThe cleaner app shipped on day one with everything the old one did. Every tick, photo and sign-off queues on the device and syncs when signal returns. Checklist items and required photos are set per property in Setup.

Cleans on the phone
Cleans on the phone
04 · Grounds work

Tasks planned around guests, not just dates

BeforeMowing and slashing were booked by text and sometimes ran next to a guest trying to sleep in.

NowThe task planner cuts each 7am to 6pm workday by actual check-out and check-in times and by fire bans, caps each person at 480 minutes a day, and keeps noisy work away from any stay within earshot. Each of the 31 work zones knows which stays it disturbs. The planner re-runs on every Smoobu sync and every 15 minutes.

Grounds tasks on the phone, planned around guests
Grounds tasks on the phone, planned around guests
05 · Early check-in

Guests told the moment their stay is ready

BeforeEarly check-in meant the owner checking with the cleaner, then messaging the guest.

NowWhen the cleaner signs off a turnover on arrival day with nothing flagged, the guest gets a message through their own channel saying they can check in early. If grounds work is planned near that property today, the message says so. One switch, one template, sent once per stay.

daltonspaddock.comThe early check-in rule: when it sends, when it holds back
The early check-in rule: when it sends, when it holds back
06 · Guest messages

Six timed messages, and a drafted reply to everything else

BeforeEvery guest message was written by hand, in three different inboxes.

NowSix messages go out from booking to after departure, through Smoobu into each guest’s own Airbnb, Booking.com or email thread, at most one a day and never overnight. For everything else, Claude drafts a reply from the property facts, the booking and similar past exchanges, and the owner approves. Nothing sends on its own.

daltonspaddock.comGuest threads grouped around today, replies drafted
Guest threads grouped around today, replies drafted
07 · The stay guide

A private guide per booking, and the numbers on how guests use it

BeforeWifi passwords, how-tos and local tips were repeated in messages.

NowA secret link unlocked by the booking surname: wifi, how-tos, weather, arrival time, late check-out, the shop. It works offline and alerts the owner when a guest sets an arrival time. The guide dashboard shows what guests opened and used, so the guide improves.

daltonspaddock.comStay guide: reach, when guests open it, what they use
Stay guide: reach, when guests open it, what they use
08 · Pricing

Every unbooked night repriced every two hours

BeforeNightly pricing ran as a separate job nobody could see into.

NowThe engine starts from a rate card, applies last-minute cuts, lifts prices when a month is selling faster than last year, eases minimum stays as dates get close and opens short gaps between bookings. Every night shows why it is priced the way it is. Only changed nights are pushed to Smoobu.

daltonspaddock.comThe next 30 unbooked nights, each with its rate and the reason
The next 30 unbooked nights, each with its rate and the reason
09 · Gap nights

Unsold nights offered to past guests, drafted and ready to approve

BeforeGap nights went unsold.

NowThe planner spots gaps worth filling, drafts a text and an email to past guests who stayed in that property or at that time of year, and queues them for approval. Availability is re-checked at send time.

daltonspaddock.comGap-night campaigns, drafted by the planner, approved by the owner
Gap-night campaigns, drafted by the planner, approved by the owner
10 · Insight

How the operation is running, not just what it earned

BeforeTurnovers, cleaner hours and issues were not measured.

NowOperations metrics cover turnovers, same-day turnovers, cleaner hours, issues per 100 stays and unsold gap nights, for any period and property, alongside profit, pace, seasonality and channel views compared with the same dates last year. Costs come from Xero.

daltonspaddock.comOperations metrics for the year to date
Operations metrics for the year to date
What changed

From first commit to daily use in 19 days

It started as a spec and prototype: a scheduling engine, a 13-table schema and a zone map over an aerial photo.

The live platform replaced the earlier admin and cleaner app, and the cleaner app shipped on day one with everything the old one did, so staff switched straight over. Features were added almost daily. On 18 October a full logic sweep of every function and screen fixed 46 bugs. In daily use since 6 October 2026: 305 commits, 101 tables, 17 scheduled automations.

18 functions, each with the screen it runs on
daltonspaddock.comEvery booking, clean and task on one calendar
01 · Systems

Every booking, clean and task on one calendar

A scrolling view of every stay across the three properties, with price, clean and task layers.

Bars are drawn to each booking’s true check-out and check-in times, with short gaps hatched. Bookings, money, rates and messages arrive from Smoobu by webhook and an hourly reconcile, so the calendar is always the current picture.

Upcoming cleans, same-day turnovers and replies, on the phone
02 · Operations

A clean for every check-out, created before anyone asks

A turnover clean is created for every check-out and offered to the cleaner.

Smoobu webhooks and an hourly reconcile create, move or cancel cleans, and the cleaner is emailed about every change. Same-day turnovers are flagged.

The cleaner app works offline: every tick, photo and sign-off queues on the device and syncs when signal returns.

Grounds tasks planned around guests, on the phone
03 · Operations

Grounds work planned around guests

Recurring and one-off grounds and maintenance tasks, each given a day and a time slot automatically.

The planner works within check-out and check-in times, fire bans and daily load caps, and knows which stays each of the 31 work zones disturbs.

The detail

Grounds work scheduled around guests, not just dates

The task planner cuts each 7am to 6pm workday by actual check-out and check-in times and by fire bans, caps each person at 480 minutes a day (a clean counts 90 or 180), and keeps mowing or slashing away from any stay within earshot.

Each of the property’s 31 work zones knows which stays it disturbs. The planner re-runs on every Smoobu sync and every 15 minutes, so a new booking can move a task, and the person is emailed when their day changes.

daltonspaddock.comThe next 30 unbooked nights, each with its rate and the reason for it
04 · Systems

Every unbooked night priced, every two hours

Sets each unbooked night’s rate and minimum stay, and pushes the changes to Smoobu.

The engine starts from the rate card, then applies last-minute discounts, lifts prices when a month is selling faster than last year, eases minimum stays as dates get close and opens short gaps between bookings. Every night shows why it is priced the way it is, and only changed nights are sent.

daltonspaddock.comGuest threads grouped around today’s arrivals and departures (names changed)
05 · Guests

Guest threads around today, with replies drafted

All guest threads from every channel, grouped around today’s arrivals, departures and stays.

Quick replies matched to each stay’s phase pre-fill names and times. Claude drafts a reply to every new guest message from the property facts, the booking and similar past exchanges, and nothing sends without the owner.

daltonspaddock.comOperations metrics for the year to date
06 · Insight

How the operation is running, not just what it earned

Turnovers, same-day turnovers, cleaner hours, issues per 100 stays and unsold gap nights.

Calculated from cleans, issues and calendar gaps for any period and property, alongside profit, pace, seasonality and channel views compared with the same dates last year.

Also built

7 more functions, grouped by the part of the business they serve.

Operations
Supplies and issues
Low stock, issues raised during cleans and damage claim packs.
How it works

"Out" emails the owner immediately. Damage claims carry the booking platform’s 14-day deadline.

Guests
Scheduled guest messages
Six messages from booking to after departure.
How it works

Sent through Smoobu into each guest’s own Airbnb, Booking.com or email thread. At most one a day, never overnight.

Guests
Early check-in invite
Tells arriving guests their stay is ready early.
How it works

Fires only once the turnover clean is signed off with no open issues, and mentions any noisy work that day.

Guests
Personal stay guide
A private guide per booking: wifi, how-tos, weather and arrival time.
How it works

A secret link unlocked by the booking surname. Works offline and alerts the owner when a guest sets an arrival time.

Guests
Guest shop
Guests pay for items from the in-stay shelf.
How it works

A QR code opens the shop; picking the stay and surname ties each sale to a booking.

Systems
Roles and sign-in
Separate apps for owner, cleaners, vineyard manager and neighbouring winery.
How it works

Email link or 4-digit PIN. Row-level security limits each role to its own data.

AI
Local guide knowledge
Checked local places fed into reply drafts.
How it works

A search of the guide database is added to each draft, with unchecked entries labelled unverified.

Built and ready to switch on

Built and tested, waiting on an account or a final check before use.

Gap-night campaigns
Offers unsold nights to past guests by SMS and email, re-checking availability at send time.
Journal drafting
Claude drafts website articles with gaps only the owner can fill before publishing.
Aerial zone map
Clickable work zones drawn over an aerial photo of the property. Prototype stage.

Under the hood

101database tables
17scheduled automations
306commits
67 serverless functions108 row-level security policies31 work zones6 automated guest message types

How it was built

It started as a spec and prototype: a scheduling engine, a 13-table schema and a zone map over an aerial photo. The live platform replaced an earlier admin and cleaner app, and the cleaner app shipped on day one with everything the old one did, so staff switched straight over. Features were added almost daily after that, and on 18 October a full logic sweep of every function and screen fixed 46 bugs.

What changed after real use:

  • Returning guests are matched by phone number, because Smoobu issues a new guest ID on every booking.
  • Check-out moved to 10am at all three stays, with older bookings kept at their original time.
  • Cleaners are told when an accepted clean is moved, cancelled or reassigned.
  • Calendar clean markers now show only when the owner needs to act.
  • Mid-stay messages go out only with real news; rain alerts were too frequent.
  • The stay guide’s "On our way" button became an arrival-time picker that alerts the owner.