Guest messaging for a 50-room hotel: what changes below 100 rooms

At 50 rooms the hard part of guest messaging is not the volume. It is that nobody is awake at 2am, and most software is not sold to fix that.

Almost everything written about guest messaging assumes a property with a night team, a revenue manager and somebody whose job is the technology. Below 100 rooms you have none of those. The advice does not scale down cleanly, and in places it inverts. Here is what is genuinely different at your size, including where our own category stops earning its fee.

What changes at 50 rooms, and what does not

Channel mix does not change. Across the 190+ properties vGuest works with, 97% of guest conversations arrive on WhatsApp (vGuest portfolio data, August 2026). That holds at a 40-room townhouse as much as at a 400-room resort, because guests message on whatever is already on their phone. So the answer to "which channels do I need" is the same at every size: the one your guests already use, plus email, because that is where the booking confirmation lives.

What changes is the shape of the day. A 200-room inbox never empties; there is a base load all day and a queue to work through. A 50-room inbox does empty. You get long quiet stretches, then four messages in ten minutes because a flight landed or a coach arrived.

That difference decides staffing. At 200 rooms you can staff for average load, because the average is a real number that someone is always working through. At 50 rooms you cannot staff for the average, because the average is close to nothing and the peaks still have to be answered within a few minutes. You are not buying capacity. You are buying cover.

So do not accept a per-room estimate of your message volume from anyone, including us. Count it yourself; the total will be lower than a vendor's calculator suggests, and the share arriving outside your staffed hours will be higher than you expect.

Who owns the inbox when there is no night team

This is the real constraint below 100 rooms, and most advice skips it because at enterprise size it does not exist. Three honest answers.

A named person on their own phone, with stated hours. Cheapest, and it works. It fails on holiday, on their day off, and on the day they leave, because the history and the habits leave with them. If this is your model, write the hours down and publish them.

A rota across the whole team. Fair, and it fails on ownership. When everyone can see the inbox, nobody is accountable for the message that sat unread for six hours, or for keeping the standard answers current.

An assistant answers the repeatable questions, a person takes the rest at stated hours. Across our portfolio around 70% of the replies guests receive are sent by the assistant rather than a person (vGuest portfolio data, August 2026). That is the number the model lives or dies on, and it is only ever as good as what the assistant has been told.

Whichever you pick, the same three things have to be written down: your real staffed hours, one escalation route for a genuine emergency, and the list of requests that must never be handled automatically. Group bookings, accessibility, and any complaint with money attached belong on that last list at any size.

The failure to avoid is a silent inbox that looks staffed. An automated reply that says "we will get back to you shortly" at 1am, followed by silence until nine, is worse than no channel. A first reply that says when a person will answer, and answers the three questions most guests are actually asking, buys you the night.

What a PMS integration is actually worth below 100 rooms

An integration earns its keep by removing a lookup a person would otherwise do repeatedly, and by making the answer right. Arrival time, balance outstanding, late checkout. Those three are most of it. If a member of your team opens the PMS a dozen or more times a day to answer guests, the integration pays. If it happens twice a day, it does not.

The second condition matters more at your size: is it configuration, or a project? vGuest connects to Optima, EzGo and MiniHotel, and to Mews, Cloudbeds, Guesty and SimpleBooking. If your system is already on a vendor's supported list, the connection is a form and a test booking. If it is not, someone is writing code against your PMS, and below 100 rooms that is usually the wrong first purchase. Buy the messaging, run it for a year without the integration, and revisit when you know what your team is actually looking up.

Three cases where it is not worth the setup at all. Your PMS has no usable API, or a read-only one, so the assistant can answer but never write back. Your availability is effectively binary in season, full or nearly full, so "have you got a room" is answered by a person in one glance. Or the integration is quoted as a build fee plus professional services exceeding a year of the platform itself.

Put the specifics in writing. The PMS questions to ask before any AI purchase is the list, and a vendor who cannot answer it within a day is telling you something.

Where a specialist platform stops being worth it

Sometimes the right answer is a shared inbox and good templates, and pretending otherwise would cost you money.

Keep the shared inbox if your whole message volume fits inside one person's staffed day with room to spare and you have no out-of-hours expectation to meet. Keep it if your questions are genuinely varied, with no repeatable top ten — rare, but it happens at properties with an unusual guest mix. Keep it above all if nobody will own the configuration. An assistant with a knowledge base nobody has updated since launch is worse than a human typing slowly, because it is confidently wrong in public.

A specialist platform starts to pay on four triggers, and rarely on just one. The same few questions arriving dozens of times a day. Messages landing outside your staffed hours. Guests writing in languages nobody on shift reads. And more than one person needing the same inbox, with the same history, so the second person does not ask the guest to start again.

That last one is the quiet threshold. The moment two people answer the same number and neither can see what the other said, you have outgrown a phone, regardless of room count.

The fortnight test

Before you buy anything, log every inbound guest message for two weeks: channel, time, question, who answered, how long it took. Then mark each one twice. Could a competent stranger have answered this from a single written page? And did it arrive inside your real staffed hours?

A high count on the first question means automation pays. A high count on the second means your case is cover, not cost, and you should buy against response hours rather than against seat licences. Low counts on both mean you should keep the shared inbox and spend the money on your website instead.

That last outcome is legitimate, and a vendor who will not say it out loud is selling rather than advising. For the wider set of tools, see how to pick the right tech stack for a small hotel or boutique property; our guest messaging guide collects everything we have written on the channel.

Start with the log. Two weeks of your own data is a better brief than any vendor will ever be given, and it is what tells you which of the answers above is yours.