The feedback form is not usually the problem. Most hotels have one, guests fill it in, and the answers land somewhere. The problem appears on the morning a score of three out of ten arrives with a comment about the air conditioning in room 412, and the form has no way of getting that in front of the person who can fix room 412 before the next guest checks in.
By the time the sheet is reviewed at the end of the month, the guest has gone, the room has turned over four more times, and the same complaint has arrived twice more. The score is recorded accurately and nothing has changed.
A guest feedback form is worth the effort only if two things are true: the questions produce answers specific enough to act on, and a low answer reaches a named person the same day. Everything below is about those two.
Two forms, not one
Asking once, after checkout, is the arrangement most properties end up with, and it is the one that guarantees every problem is reported too late to fix.
A short form during the stay exists to catch things while the guest is still in the building. One or two questions, reachable from a QR code on the desk or in the lift, or sent by message on the morning of the second night. The purpose is not measurement. The purpose is a chance to put something right while it still counts, which is the only version of feedback that converts a bad stay into a good review.
A longer form after checkout exists to measure. It can ask about the whole stay, the booking process, the breakfast and the value for money, because the guest is no longer standing in a corridor with a suitcase. Sent within about a day of departure, while the stay is still distinct in memory and before the next trip overwrites it.
The two forms want different lengths, different tones and different routing. Trying to serve both with one form produces something too long to fill in during a stay and too vague to be useful after it.
| When | Length | What it is for | Who reads it |
|---|---|---|---|
| During the stay | One or two questions | Fixing something before the guest leaves | Duty manager, same day |
| At checkout, at the desk | One question | Catching what nobody says out loud | Front desk supervisor |
| One day after departure | Five to eight questions | Measuring, and asking for a public review | Whoever owns quality, weekly |
| After a cancelled or shortened stay | Two questions, open text | The reasons the other forms never surface | Management, individually |
What to ask a guest who has thirty seconds
Guests fill in short forms. The evidence is in the shape of the answers: a form with eighteen rating rows gets abandoned halfway or answered in a straight line down the middle, which produces a clean looking dataset with no information in it.
A workable after stay form is five to eight questions, structured like this.
One overall rating. The number that goes on the trend line. Keep the scale the same forever, because changing it makes every comparison with last year meaningless.
Three or four ratings that map to a department. Room and cleanliness, staff, food, and check in or check out. The point of splitting them is not detail for its own sake. It is that each one has an owner, so a low score can be routed rather than discussed.
One open text question, phrased to produce specifics. "What is the one thing that would have made this stay better" outperforms "any comments" by a wide margin, because it asks for a single item and gives permission to name it. This field is where almost all of the actionable material lives.
One question about whether they would return or recommend. Not because the number is magic, but because the gap between a high overall rating and a low intention to return is the most reliable early warning a property gets.
Anything beyond that is asking for a favour. If a question exists because a head office report has a column for it, and nobody acts differently based on the answer, it is costing completion rate for nothing.
Route it by department, or it dies in a shared inbox
The fastest way to waste a working form is to send every submission to one address that four people can open and nobody owns.
What replaces that is unglamorous. Each submission gets an owner and a state. The owner is a person, not a department, because departments do not read email. The state says whether it has been looked at, whether something is being done, and whether it is finished. A board where every response sits under a stage makes the pile legible at a glance, which is the difference between a manager who can see three unresolved complaints and a manager who believes there are none.
Two more properties earn their place quickly. Tags, so that "air conditioning" can be counted across a quarter rather than remembered anecdotally. And internal fields, meaning the columns the team fills in after the fact: the room number if the guest did not give it, whether a refund or a credit was offered, and which shift was on. Those are not questions for the guest. They are the record of what was done, and they belong with the response rather than in a separate spreadsheet that goes out of date within a week. Tools built for handling responses after they arrive treat all of this as part of the submission, which is what keeps the two halves from drifting apart.
The routing rule itself can be simple. Anything with an overall rating below a set threshold goes to the duty manager immediately, by whatever channel that person actually watches, and everything else waits for the weekly read.
When a score comes back low
Speed matters more than wording, and wording matters more than compensation.
Same day, if the guest is still in the house. A phone call or a knock is worth more than a message. The complaint is often smaller than it looked in writing, and a problem solved during a stay rarely appears in a public review afterwards.
Within a day, if the guest has left. A reply from a named person, not an address that cannot be answered. Three things need to be in it: what was understood, what has been done or checked, and an acknowledgement that it should not have happened. A guest who receives that tends to update their view of the property, and a substantial share of them say so.
What not to do is well documented by the results. Do not open with an explanation of why the thing happened, because it reads as a defence. Do not offer a discount before addressing the issue, because it converts a complaint about a room into a transaction. Do not reply with a template that could have been sent to anyone, because guests recognise them and it confirms nobody read the message.
Then separate the individual case from the pattern. One low score about a noisy corridor is a case to answer. Five in a month naming the same floor is a maintenance job with a date on it, and the only way to tell the two apart is to have tagged the first four.
The public review sites, and what the form is for
The form and the review platforms do different work, and confusing them creates two familiar mistakes.
The first is using the private form as a filter, asking for a rating and then only inviting the happy guests to review publicly. Platforms have explicit rules about selective solicitation, and the practice tends to get found out. The second is skipping the private form entirely on the grounds that the public sites already show the score, which means the only place a complaint can land is in public, permanently, where it can no longer be resolved quietly.
The defensible arrangement is straightforward. Invite every departing guest to the private form. Ask for a public review in the thank you message that follows, with the same wording for everyone regardless of what they said. Low scores get a reply and a fix. High scores get a reason to write one publicly. Both are honest, and the second is where most of a property's new bookings come from.
Languages, and the phone in the guest's hand
Almost all of these forms are opened on a phone, often on hotel wifi, sometimes by someone who does not read the language the property is written in.
That makes two things practical requirements rather than refinements. One question per screen, because a long scrolling page of rating grids is what abandonment looks like on a small display. And the same form available in the languages of the property's actual guests, which is worth checking against the booking data rather than assumed. A rating scale translates cleanly. An open text question does not, so expect answers in several languages and make sure somebody can read them before promising a same day reply.
A QR code on the desk works if it goes to the short form. Pointed at an eight question survey it gets scanned, glanced at, and closed.
Counting across rooms, shifts and seasons
A single month of scores says very little. The comparisons that produce decisions are the ones that hold something constant.
By room and by floor, which surfaces maintenance rather than service. By shift, which surfaces training. By channel, since guests who booked direct and guests who arrived through an agent often rate the same stay differently and expect different things. By season, because a resort in August and the same resort in February are close to different businesses.
None of that works without two habits: a consistent tag set, and an export that carries the internal fields alongside the guest's answers. If the tool can only export the raw questions, the analysis stops at the overall score, which is the one number that never tells anyone what to do next.
What to change first
Put one short form in front of guests during the stay, with a route that puts anything low in front of the duty manager the same day. Leave the longer after stay form as it is for now, because the timing of the fix matters more than the quality of the questionnaire. Then decide who owns the weekly read and give every submission an owner and a state, which is where a form stops being a record and starts being a process. Seeing what that looks like in practice takes a few minutes on a working setup, and the use cases for enquiry handling apply almost unchanged to a front desk.
Q1. When is the best time to send a hotel feedback form?
Use two moments. A very short form during the stay, reachable by QR code or sent on the morning of the second night, so problems can still be fixed while the guest is present. Then a longer form about a day after departure, while the stay is distinct in memory. Waiting a week loses both the detail and the chance to put anything right.
Q2. How many questions should a guest feedback form have?
Five to eight after the stay, and one or two during it. One overall rating, three or four ratings that each map to a department, one open question asking for the single thing that would have improved the stay, and one about returning or recommending. Longer forms get abandoned or answered in a straight line, which looks like data and is not.
Q3. How should a hotel respond to a low score?
Same day if the guest is still in the building, within a day if they have left, from a named person rather than a no reply address. Say what was understood, what has been done or checked, and that it should not have happened. Leave compensation out of the first message, because offering it before addressing the problem turns a complaint into a negotiation.
Q4. Is it acceptable to ask only happy guests for a public review?
No. Review platforms treat selective solicitation as a breach of their terms, and properties that do it tend to be found out. Invite every departing guest to the private form, then ask everyone for a public review in the same wording. The private form is where complaints get resolved, not where they get filtered.
Q5. What should be tracked alongside the guest's answers?
The things the guest cannot supply: room number if it was not given, shift on duty, whether anything was offered, and which channel the booking came through. Recorded as internal fields against the submission rather than in a separate spreadsheet, they are what make it possible to tell a one off complaint from a maintenance problem on a specific floor.