Case study · Blissful Escape
Ten properties. No front desk. No data entry.
Two people run ten short-term rentals across four countries, from Miami to Milan. HostEarnings Professional Services built the automation that drafts every guest reply and wakes the right person at 2am. The HostEarnings platform files every booking, payout, and fee straight into the bookkeeping, so nobody keys in a transaction by hand anymore.
82%
Occupancy across the portfolio
Up from 71%
11 hrs
Of manual data entry gone each month
Bookings, payments, and statements file themselves
67%
Fewer nights someone gets woken
About 9 a month down to 3
The portfolio.
Ten units across four countries. One in Miami, seven on the Riviera Maya, and one each in Toronto and Milan.
The problem
Every extra country meant one more way to miss a message.
A ten-property portfolio can't justify a front desk, but guests expect one anyway. Messages arrived on Airbnb, WhatsApp, email, and the direct booking form, at all hours. Someone had to read each one, work out who it was for, and forward it. Some nights nobody did, and a first reply typically took two hours and forty minutes. Airbnb counts that against your search ranking.
A burst pipe in Miami in the middle of the night needed the same response as a lost key in Milan at lunchtime, and nothing in the setup knew the difference. Whoever was still awake dealt with it, and it kept being the same two people.
Check-in ran on a text chain: confirm the flight, send the door code, hope the guest read it. Check-out worked the same way in reverse, and units sat running the air conditioning long after everyone had left.
Bookings lived in Airbnb, VRBO, and a direct-booking spreadsheet. Building an owner report meant pulling numbers out of all three by hand, every month. Statements went out late, and the questions they raised came back even later.
What HostEarnings changed.
This came in two pieces. Professional Services built the automation, and the HostEarnings platform took over the bookkeeping underneath it.
Professional Services
The automation was shaped around how they already work.
The rules for who gets which message, and who gets called at night, are specific to this team.
The AI drafts, a person approves
Every message gets a drafted reply the moment it lands, routed to whoever it concerns: property manager, cleaner, or maintenance. A person approves from their phone before anything reaches a guest.
Emergencies reach a real person, fast
An urgent message skips the queue and triggers the emergency system, including a bedside Bluetooth speaker that wakes the on-call manager or cleaning lead. Anything that can wait until morning is held until morning.
Every channel in one place, including your own
Airbnb, VRBO, Booking.com, and the direct booking form all land in one inbox. Direct reservations carry no platform commission, so every booking the team moves across is worth more to them.
Booking to check-out, on autopilot
A confirmed booking sends its own pre-arrival messages and door code, and the smart lock lets guests in and out on schedule. Once the lock shows the unit is clear, the check-out message goes out on its own.
Air conditioning that knows when the room is empty
Sensors tell the system when nobody is in the unit, so the air conditioning shuts off instead of cooling an empty room. It comes back on before the next arrival.
The HostEarnings platform
Bookkeeping stopped being a monthly rebuild.
Automation only helps if the money it moves ends up somewhere useful. Everything the automation touches lands in HostEarnings, ready to read by the time anyone opens it.
Every reservation lands in the ledger on its own
Payouts, platform fees, cleaning costs, and direct bookings flow in from every channel without anyone keying them, categorized as they arrive.
The numbers are already pulled together
Collecting figures from every platform used to be the whole job. Now the only thing left is checking the handful of rows that need a human eye.
Owners read the report and reply inside it
Owners open their statement, comment on any line item, and get the answer on that line instead of in an email thread.
The engagement
The four-week build.
Four weeks from the first conversation to an automated monthly close. The build shipped one piece at a time, so the team was never learning a whole new system at once.
- Days 1 to 3
Sat in with the team
Three days alongside them, ninety days of message history, and their last close. That produced the map everything else was built from: who to call for what, which unit has which lock, and every place a number was being retyped.
- Week 1
After-hours dispatch, and nothing else
The one thing they asked for first, live before the end of week one. Rules for what counts as urgent, based on their real messages, an on-call schedule for each country, and the bedside speaker in each manager's room. Nothing else changed until this was working.
- Week 2
Messaging, check-in, and the units
Reply drafting tuned to their voice, every booking channel into one inbox, smart-lock schedules per property, and the air-conditioning shutoff in the units with sensors fitted.
- Weeks 3 and 4
Bookkeeping and month-end
Every kind of charge mapped to the right category, payments matched automatically as they arrive, and the owner report rebuilt in the format their owners already knew how to read. The first close on the new setup ran at the end of week four.
- Since then
Small changes on request
Onboarding for a new property, a revised escalation rule after one slow night, and a second owner report layout. Most land within a day.
Systems integrated during the build
- Airbnb
- VRBO
- Booking.com
- Direct booking form
- Smart locks
- Presence sensors
- Bluetooth speakers
- SMS and email routing
- HostEarnings ledger
All of it is configuration, so the same build runs on a different portfolio once the rules are rewritten for that team.
Before and after.
First reply to a guest
Typically 2h 40m
9 minutes, and 73% of drafts send unedited
Urgent issue at 2am
Whoever was still awake
The right person, woken in 90 seconds
Check-in
Text chain, door code typed out each time
Smart lock, no staff call on 94% of arrivals
Empty unit, AC running
Ran until somebody noticed
Powers down on its own, about $380 a month saved
Direct bookings
8% of reservations
31% of reservations
Entering transactions
Keyed in from every platform
None. They arrive categorized
Matching payments across currencies
Converted and checked one by one
Done automatically as they arrive
Month-end close
A day and a half
About 40 minutes
Owner statements
Out around the 12th
Out on the 3rd
If any of this sounds like your week, we should talk.
Talk to usStart to finish
One night in Miami.
This is a single water leak, from the guest's first message to the line item an owner reads in their next statement. Nobody on the team was awake when it started.
02:14
Guest messages on Airbnb: there is water coming out from under the bathroom door.
02:14
The system reads it as urgent and attaches the property, the unit, the guest's checkout date, and the shutoff valve location before anyone has seen it.
02:15
The Bluetooth speaker in the on-call manager's room wakes them. The same alert texts the Miami plumber on the after-hours list.
02:18
A reply to the guest is already drafted and waiting. One tap sends it: someone's on the way, and here's where to find the valve.
02:41
Plumber on site. Mains off.
03:20
Leak stopped. The bathroom is out of use until morning, so the guest is offered a partial refund for the night before anyone has to ask for one.
Month-end
The incident, the plumber's invoice, and the partial refund are already categorized in the ledger. The owner reads all three in their report and leaves a comment on the line item.
Before, the same night would have ended with a guest waiting until morning, a one-star review, and an invoice nobody could match to a property six weeks later. And nobody staffs a night shift to get there.
In their words.
"We used to lose sleep over what might be going wrong at a property we couldn't see. Now the reply is already written by the time I pick up my phone, and I'm approving it instead of writing it at midnight."
"I get the statement on the 3rd and I can ask about a line item right there in the report. I used to wait until the 12th, then send an email nobody answered until the following month."
How these numbers were measured
- Figures compare the three months before rollout with the three most recent months, across all ten properties. Response and close times are medians.
- Response times and occupancy come from the booking platforms' own host dashboards rather than internal estimates.
- Electricity is metered per unit and weather-normalized against the same months a year earlier.
- Amounts are converted to USD at the average rate for the period, across the four currencies the portfolio earns in.
Have a portfolio that outgrew your inbox?
Tell us what's breaking first, whether that's guest messages, emergency response, direct bookings, or month-end. We'll show you what this setup looks like on your properties.
Talk to usMore case studies on request. Tell us what your setup looks like and we'll send the closest match.