Hotel guest satisfaction survey: the questions, the timing and the follow-up

A hotel guest satisfaction survey is a short private questionnaire sent to a known guest: about ten closed questions by stage of the stay plus one open question, a brief check during the stay, the full survey 24 to 72 hours after checkout, a KVKK notice at collection, and a named person who contacts anyone reporting a problem.

The survey goes out, forty guests answer, the average looks fine, and the same complaint about the breakfast coffee appears in three public reviews a month later. The problem is rarely the guests. It is that nobody owned the question, the timing was wrong, and the one guest who named the problem was never called back.

This article gives you the full routine: the question set, the wording, the two send timings, the privacy notice, and the follow-up that closes the loop.

What does a guest satisfaction survey tell you that public reviews do not?#

A survey tells you what a known guest thought at a moment you chose, in a form you designed. A public review tells you what a self-selected guest chose to publish, usually after the stay and for an audience.

Reviews come from guests who feel strongly. A survey reaches everyone you ask, including the quiet guests who never write anything. That difference matters most exactly where problems are small: the lukewarm shower, the slow lift, the coffee.

You also control the question. A review says the staff were rude; a survey separates front desk, housekeeping and restaurant, so the answer points at a department and a shift. And you control the timing: a review usually lands days or weeks after checkout, while a survey can arrive while the guest is still in the building and the problem can still be fixed.

A survey gives you a denominator. You know how many guests you asked and how many answered, so a movement in the numbers means something. A review count does not.

One caution: a survey is self-reported and often polite. Guests under-report small irritations to a hotel they liked, so read the open question and the lowest scores, not the average. Public reviews still decide the next guest's booking, so keep the two routines separate: the survey is an operations tool, the review reply is a shop window. For that side, see a weekly routine for hotel online reputation management.

Which questions belong in a hotel guest satisfaction survey?#

About ten closed questions grouped by the stage of the stay they belong to, plus one open question at the end. Every question must map to a department that can act on it.

Booking questions belong to reservations: was the booking easy, and was the confirmation clear about what was included. Arrival questions belong to front office: speed of check-in, whether the room was ready, and whether the desk explained what the guest needed for the stay.

Split the room into separate items: cleanliness, quiet, bed comfort, temperature and bathroom. A single question about the room cannot be acted on, because a low answer tells you nothing about what to fix. Ask about breakfast always, and about restaurant and bar only if a filter question shows the guest used them.

For staff, ask whether a request was handled and how quickly, not whether staff were friendly. Response is measurable; personality is not. At departure, ask about the speed of checkout and the clarity of the bill. These two carry more weight than hotels expect, because they are the last thing the guest remembers.

The open question is one line and specific: what is the one thing we should fix before your next stay. Do not ask what you already hold in the system, such as dates, room type or rate; prefill it or leave it out.

That is fourteen items; trim to ten by dropping the items your hotel cannot act on.

How should you word the questions and the scale?#

One idea per question, the guest's own language, and the same scale every time, so that this month can be compared with last month.

One idea per question is the hardest rule. "The room was clean and quiet" cannot be answered if it was clean and noisy, and the answer you get will be about whichever half the guest cared about. Avoid questions that assume the answer, such as "how much did you enjoy our breakfast", and words that mean different things to different guests, such as value, comfort or experience.

Use a five-point scale with every point labelled, and keep the labels identical across languages. An unlabelled 1 to 10 invites each guest to invent a meaning. Keep the direction the same everywhere: always the top of the scale is best. Mixed directions produce answers that look like data errors.

Translate natively, not with a machine, and have a native speaker read the finished screen. A clumsy translation is itself a signal about the hotel. Net Promoter Score is one question with a fixed calculation; if you use it, use the definition and do not blend it with your own scale. For how the two compare, see guest satisfaction metrics explained: review scores, NPS and complaint rates. Test the survey on three colleagues before launch. If one of them hesitates, a guest will abandon the page.

When should you send a hotel survey?#

A short check during the stay and the full survey after checkout. They answer different questions and need different owners.

The in-stay check is two or three questions, sent at the midpoint of a stay of three nights or more. Its purpose is recovery, not measurement. Send it only if someone can act on the answer within a couple of hours; otherwise it teaches the guest that reporting a problem changes nothing.

The post-stay survey goes out after departure. Too soon and the last hour, the bill or the taxi dominates the answers; too late and the detail is gone and the response falls. A working range is 24 to 72 hours after checkout. Send one reminder, then stop: a second reminder converts goodwill into annoyance and costs you the next survey.

Use one channel and one link, with no account to create. A login wall is the fastest way to lose the guest who has the most to say. Keep the invitation free of offers: a survey email that also carries a room discount is a marketing email wearing a survey's clothes. Never send a post-stay survey to a guest who is still in the building, and never send the same survey twice for the same stay.

The survey is part of a wider sequence of guest messages; for the earlier steps, see hotel pre-arrival emails: what to send guests and when.

What must the survey's privacy notice say?#

A short notice at the point where the guest gives their answers, covering what KVKK Article 10 requires, and a question set that collects nothing beyond the purpose.

KVKK requires that guests are informed about who is processing their data, why, and what their rights are. Make sure your notice covers these elements, and consult your legal adviser for the exact wording your case requires. The same law asks that data be relevant, limited and proportionate to its purpose. A satisfaction survey does not need an ID number, a date of birth or a passport copy.

Prefill what the hotel already holds, such as stay dates and room type, instead of asking again. It is shorter for the guest and more proportionate for the hotel.

On the invitation itself: the Law on the Regulation of Electronic Commerce requires prior consent for commercial electronic messages and lets the recipient refuse at any time without giving a reason (Law No. 6563, Articles 6 and 8, mevzuat.gov.tr (opens in a new tab)). Whether a survey invitation counts as commercial depends on its content and framing, so confirm with counsel. Give a refusal route in every message and honour it for the survey as well as for marketing; a guest who opts out of marketing has not automatically opted out of the survey, and the reverse is also true.

Decide how long you keep answers and say so in the notice. An answer kept forever is a liability, not an asset. This is a description of the duty, not legal advice: confirm the wording, the retention period and your own case with your adviser.

How do you read the answers?#

Read the open question first, then the lowest scores by department, and compare against your own past months rather than an industry figure you cannot source.

Assign an owner to every question before launch. A question with no owner is a question to delete. Read counts as well as averages: with a small number of responses a single answer moves the average, so a drop may be one guest having a bad morning.

Compare month over month, and then by source market, room type and length of stay. A fall in one segment against a flat total is a specific problem, not a general one. Read the open question for repeated words: three guests naming the same lift or the same coffee is a pattern; one guest naming it is a note. Separate the score from the volume of comment, because a low score with no comment is a puzzle to investigate, not a verdict. Do not benchmark against a number you cannot trace to a named source; track your own line and let it become the benchmark.

Who contacts a guest who reported a problem, and how fast?#

A named person, within two hours for an in-stay report and within one working day after checkout. The contact is a conversation, not a template.

Write the trigger down: any answer at the bottom of the scale, any open comment that names a problem, or any guest who asks to be contacted. Name the owner by shift: guest relations during the day, the duty manager at night. A shared inbox is not an owner. Call if the guest left a number and the problem was serious; otherwise send a short personal email that names the problem back to them in their own words.

Say what changed, not only that you are sorry. "The corridor light is on a nightly check now" closes the loop; "we value your feedback" does not. Record the outcome on the guest profile so the next stay starts with the information rather than without it. Feed recurring problems into the weekly review, so the same complaint is not closed twice by two different people. For the face-to-face version of this skill, see handling hotel guest complaints at the front desk: a recovery routine.

Never offer a discount, an upgrade or a prize in exchange for a better score or a public review. It turns feedback into a purchase and misleads the next guest.

What to do next#

Choose ten questions, write an owner beside each one, set the two timings, and put the information notice on the first screen. Run it unchanged for three months before you add or remove anything, because a survey you keep rewriting never produces a line you can read.