ConceptSelf-initiated, built on public research. Not client work.See real case studies
Self-initiated work Concept 01 · Hospitality

Listen first. Then make it personal.

A Voice of the Customer programme for a five-star beach resort, and where I’d take it next, depending on what the guests tell us.

  • Self-initiated
  • Built on public research
  • Fictional resort
  • A proposal, not delivered work
The problem

Guests speak in six places. The survey arrives after they’ve left.

Part 1 · The VoC programme

Hear every guest, everywhere, and fix it before checkout.

Then · Let results decide

Fix the basics, add a human fallback, or personalise with consent.

How it’s measured

Issues fixed in-stay, repeat stays, zero privacy complaints.

The scenario

Nothing is broken. That’s the risk.

Full all year, short stays

Guests from all over the world stay a few nights. A problem not fixed fast leaves with them.

New rivals every year

New luxury hotels open nearby, and rivals launch digital check-in and AI concierges.

Feedback everywhere, owned by no one

Reviews, surveys, messages, front desk, social media and the team, each in a different department.

The General Manager’s brief“Keep us the resort guests remember and come back to. Hear every guest while they’re still here, use new technology where it genuinely helps them without losing the human touch, and prove it in repeat stays and ratings within 12 months.”
The anti-brief

What this programme is not

The traps I’d rule out on day one, so nobody mistakes activity for progress.

Part 1The Voice of the Customer programme

Hear every guest, while they’re still here

Guests already speak in many places. The programme connects them, gives each one an owner, and acts before checkout.

Listening posts across the stay

  1. Booking
    • Booking details
    • Special requests
    • Optional: what brings you here?
    Owner: Reservations
  2. Before arrival
    • Replies to the pre-arrival message
    Owner: Guest Relations
  3. Arrival
    • Front desk conversations
    • Waiting times
    Owner: Front Office
  4. During the stay
    • A day-one “How’s everything so far?”
    • Messages and the app
    • What the team notices
    Owner: Guest Relations, F&B, Beach & Pool
  5. Departure
    • “One thing we could do better?”
    Owner: Front Office
  6. After the stay
    • Post-stay survey
    • Online reviews
    • Social media
    Owner: Guest Relations, Marketing

Two loops: one guest, and the whole resort

Same day fix it for this guest Every week fix it for every guest
Inner loop · same dayFix it for this guest
  1. The signal reaches its owner
  2. The owner fixes it
  3. Someone follows up before the guest leaves
Outer loop · every weekFix it for every guest
  1. A weekly forum on root causes, with every department
  2. A monthly review with the General Manager
  3. Changes to standards, training and processes

Listen for the why, not only the complaints

01Why this destination

A celebration, a family holiday, a wellness break, a stopover.

02Why this resort

The beach, the privacy, the kids’ club, a restaurant, the brand.

03What they enjoy

The moments guests praise, not only the ones they complain about.

Make sense of it, in every language

AI reads at scale

Reviews, surveys and messages grouped into themes, whatever the language.

People check and decide

The team validates each theme against what they see every day.

Every theme gets an owner

And a date. Reviews get substantive replies, not templates.*

* A Cornell study of 80 hotels found revenue rises with review response rate up to about 40%; hotels replying to more than 85% earned less than hotels that replied to none.

What the first 90 days will test

Tested against a small sample of real public reviews, then the resort’s own data. Nothing will be named.

Where I’d take it from there

Let the results decide the next step

The programme isn’t the end. What guests say decides what I’d do next. Choose a finding:

What the VoC shows

Waits at arrival, a crowded pool, rooms that need maintenance.

What I’d do next

Fix the basics first. Personalising a stay that doesn’t work is decoration. I’d run an operations sprint with owners, service standards and a clear before-and-after measure.

Why: current research

In one study of hotels with the most negative reviews, 83% of complaints were about tangibles: facilities, cleanliness and room condition.

Most likely it’s a mix. Then the order matters: fix the basics, make sure a person is always there, and only then personalise.

Part 2If guests want it personal

Personalisation guests control

When the voice of the customer says guests want to be known, the next challenge is doing it without making them feel watched.

The paradox

Remembered, not watched

Some guests want to be recognised, others want to stay invisible. Five-star service has to tell which is which.

70%worried about privacy in hotel apps
71%preferred a person to a machine
74%felt cared for when preferences were remembered

2026 study of 290 guests at one luxury hotel (preprint). A signal, not proof.

The personalisation dial

The guest decides how much we remember

Turn the dial and watch the same stay change. Anyone who doesn’t choose gets Essentials: only what the stay needs.

The default for anyone who doesn’t choose: only what this stay needs, nothing kept without a yes.

  1. 01At booking

    Two optional questions: what brings you here, and how you’d like to be looked after.

  2. 02Before arrival

    One reminder to share allergies or accessibility needs, if any.

  3. 03Arrival

    Needs confirmed discreetly, never quizzed.

  4. 04During the stay

    Only what this stay needs is noted.

  5. 05Departure

    Kept only if the guest says yes.

  6. 06Next visit

    The essentials are ready. Nothing more.

Where the line is

Personalise the service, never profile the person

May be remembered, with consent
  • The occasion: a celebration, a family trip, a rest
  • Dietary needs and allergies
  • Preferred language
  • Room temperature and pillow type
  • Accessibility needs
  • A favourite table or sunbed area
Never recorded
  • Personal belongings
  • Private habits
  • Who guests spend their time with
  • Anything about their body or appearance
  1. 01Learn at booking

    Two optional questions and the dial. Travel-site guests get them in the confirmation.

  2. 02Read the cues

    The team learns who wants recognition and who wants privacy.

  3. 03AI with consent

    AI spots patterns across stays. People decide what to do.

How I’d know it’s working

A few measures, linked to the business

Fixed in-stayIssues resolved before checkout
EffortSatisfaction and effort during the stay
Themes closedRoot causes fixed, not just reported
ReturnRepeat stays and direct rebookings
Dial setShare of guests who choose a level (Part 2)
TrustPrivacy complaints: zero
What I’d validate first

Do the themes match the floor?

Check what reviews say against what frontline teams see every day, and test the booking questions with a few guests before rolling anything out.

Where I’ve done this for real

Hotel Paris, before the internet

A 24/7 feedback loop, run by people, and more than half of each season’s bookings from returning guests.

Read the case
Sources
  • Luxury Group by Marriott International, The Intentional Traveler (1,750 affluent travellers, 2025), via TTG Asia: ttgasia.com
  • Hotels.com survey of 450 hotels (2025): expedia.com newsroom
  • Guest expectations and data sharing (Data Axle 2025; The Drum 2024), via Lighthouse: mylighthouse.com
  • Hotel performance attributes and consumer complaints (SERVQUAL analysis of 35 hotels, 2025): reference-global.com
  • The digital readiness paradox in luxury hospitality (Research Square preprint, 2026): researchsquare.com
  • Cornell Center for Hospitality Research, review responses and hotel revenue (80 hotels): business.cornell.edu

Building a customer experience team? Let's talk.