Room service delivery times: how they are measured and what is acceptable

The guest in 508 ordered a soup and a club sandwich at nine thirty at night. They arrived at ten twenty. In the morning meeting, the hotel manager says room service is too slow; the chef says the soup was out in twelve minutes; the server says he went up as soon as he was told. All three are right, and all three are talking about different clocks.

What delivery time actually is

In a hotel restaurant, a room service order passes through four moments that can be stamped with a time: the guest orders it, the ticket enters the kitchen, the dish is ready on the pass and the guest signs the slip at the door. Delivery time, the only one the guest cares about, runs from the first moment to the last. The other three are internal times that tell you where it got stuck.

The confusion in the meeting comes from each area measuring its own segment and defending it. The kitchen measures from ticket to dish ready. The server measures from being told to the signature. The person taking the order measures from message to ticket sent. Nobody measures the total, and the total is the promise the hotel made to the guest when it said “thirty minutes”.

Defining delivery time as order to signature changes the conversation. Nobody argues about whose fault it was; you look at which segment grew. And to see that you need all four time stamps on every order, not on a sample somebody writes by hand on a Tuesday.

The four clocks of an order

• Clock 1, order received: the moment the guest finished ordering, by phone or by message, and the order was confirmed.
• Clock 2, ticket sent: the moment that order reached the kitchen screen or printer. The difference from clock 1 is capture time.
• Clock 3, dish ready: the moment the kitchen marked the order as done and placed it on the pass. The difference from clock 2 is kitchen time.
• Clock 4, guest signature: the moment the server closed the delivery at the door. The difference from clock 3 is transit time, which includes waiting for someone to go up.

Why “acceptable” depends on the hour and on the dish

There is no universal number of minutes that is good for everything. Coffee and bread at seven in the morning and a steak at ten at night cannot be measured against the same target, and a hotel that tries to give them the same promise will fail on one of the two. The target is defined by dish type, and adjusted by time of day because the kitchen and the floor team do not have the same capacity at seven as at eleven.

There is a second reason to separate by hour: the morning concentrates advance orders that can be prepared in sequence, while the night receives loose orders that arrive when the dining room is also full. Thirty-five minutes at eight at night can be excellent, and the same time at seven in the morning, with simple breakfasts and a kitchen already producing, can be a sign that someone is not at their station.

That is why the question “what is acceptable?” has a two-level answer. First: acceptable is what you promised the guest when you took the order. Second: the promise has to come from the real capacity of your kitchen and your floor at that hour, not from a wish. If you promise 20 minutes and deliver in 35, the problem is not the kitchen; it is the promise.

How to measure it without surveys or stopwatches

The most common way to measure room service is not to measure it and rely on complaints. The second is to ask the server to write the time on the slip. Both fail for the same reason: they depend on someone remembering at the busiest moment. What works is having the ticket system record the four stamps as part of the normal flow, without anyone doing anything extra.

1. Clock 1 is stamped when the order is confirmed with the guest. If the order comes by messaging, it is the time of the guest’s affirmative reply.
2. Clock 2 is stamped when the ticket appears in the kitchen. If capture and sending are the same act, clocks 1 and 2 almost coincide, and that is the ideal.
3. Clock 3 is stamped when the kitchen taps “ready” on the screen or the expediter pulls the ticket. Without a kitchen screen, this is the first clock to get lost.
4. Clock 4 is stamped when the server closes the delivery from the tablet or phone on receiving the signature, which releases the charge to the guest’s folio at that same instant.

With that, every order carries four recorded times and three calculated segments. The kitchen page (Kitchen display) shows how the production screen stamps clock 3 without the cook stopping to cook, and the room service page (Room service) explains how the delivery is closed at the door from a device, which stamps clock 4 and ties the charge to the folio.

Average, median and the slowest order

Once you finally have the times, the temptation appears to summarize them in an average. The average lies in room service, because slow orders are few but very slow, and the average hides them. It is better to look at three numbers at once: the average, the median (the time of the order sitting in the middle of the sorted list) and the time of the slowest orders, for example the worst 10 % of the night.

The average tells you how the operation is doing in general. The median tells you what the typical guest experiences. The worst 10 % tells you how many guests you failed this week, and those are the ones who write the review. A hotel can have a decent average and still fail one in ten guests every night. That is the half of the problem the meeting does not see.

An illustrative example: a night of ten orders

The figures below are made up to show the calculation; they are not market data and not from any property. Imagine your hotel restaurant received ten room service orders between eight and eleven at night, and the system recorded the total time of each, from order to signature, in minutes.

The ten times add up to 330 minutes, so the average is 33. The median sits between the fifth and sixth order, between 30 and 31 minutes. So far the night looks reasonable if the promise was 30 minutes. But order number ten, a burger that in order four went out in 27 minutes, took 60. The average of 33 does not show it; the list does.

If you also look at the segments of order ten, the story becomes clear: 2 minutes of capture, 23 of kitchen and 35 of transit. The burger sat on the pass for half an hour because the only room service server was carrying two orders to the other building. It was not a kitchen problem. It was a floor problem, and it is fixed with a second server between nine and eleven or with a different delivery route.

The three bottlenecks and how to recognize them

With the segments separated, each bottleneck leaves a different footprint. Learning to read them is more useful than any target in minutes.

Slow capture

If the segment between order received and ticket sent grows, the problem is before the kitchen: someone takes the order in one place and copies it in another, or the person answering the phone is also running the dining room register. It is the easiest segment to eliminate, because when capture and sending are the same act, the segment becomes zero.

Saturated kitchen

If the kitchen segment grows, look at the hour. A long kitchen segment at nine at night, when the dining room is full, means room service is competing for the same station as the tables and losing. A clear rule helps: a room service ticket has the same priority as a table with the same entry time, neither before nor after, and the kitchen screen sorts them by time and not by origin.

Transit with no owner

If the segment between dish ready and signature grows, the dish waited on the pass. That happens when nobody is assigned to room service on that shift and “whoever is free” goes up. Nobody is free at nine at night. The segment drops when one server carries the named responsibility and the screen alerts them the moment an order is ready.

How the target is set by dish type

The target is not copied from anywhere; it is built from your own times. After two or three weeks of measuring, group the orders by dish type and look at the median of each group. That median, plus a safety margin you decide, is the promise you can make with confidence. If the breakfast median is 16 minutes, promising 20 is honest; promising 15 is setting yourself up to fail.

Then separate by time band. It is normal for the same burger to have one target at seven in the evening and another at ten at night. Telling the guest a different time depending on the hour is not a flaw; it is what any serious restaurant does when it says “right now we have a twenty-minute wait”. The difference is that in the hotel the guest cannot see the full dining room, so the promise has to be more precise, not less.

And the target gets reviewed. Every month, the report of times by dish and by hour is compared with the current promise. If a dish has been delivered faster than target for four weeks, the target comes down. If it has been failing for four weeks, the segment is investigated before the target goes up, because raising it is giving up. The reports page (Reports) describes which cuts should be ready for that review.

What to do this week

1. Define the four clocks with the team and who stamps each one: the person taking the order, the kitchen, the kitchen again when done and the server at the door.
2. Turn on automatic recording of those stamps in the ticket system; if it does not exist today, start with a sheet per shift, but with all four clocks, not only the total.
3. Name a room service owner per shift during the evening peak, with the screen that alerts them when an order is ready.
4. At the end of the week, sort every order by total time and read the ten slowest with their three segments; that is where your real bottleneck lives.
5. Write the first version of the target by dish type, using the measured median plus a margin, and adjust it in four weeks.

Inn Restaurant records the four clocks of every room service order as part of the normal flow: the ticket going to the kitchen, the “ready” on the screen and the guest’s signature at the door that releases the charge to the folio. In a 15-minute demo (contact) you can see the report of times by dish and by hour with a test order.