The following day, I returned to the hotel with a rather strange intention: to try to forget everything I already knew about it. It was not really possible. After so many conversations, I knew some of its decisions too well, I had read enough reviews to recognise certain patterns, and we had just found a formulation that seemed to explain much of its way of welcoming guests: first understand the person; then decide how to look after them.
That was precisely why I needed to impose a restriction on myself. If I wanted to know whether that identity could be understood from the outside, I could no longer use information that someone who had never been there would never have access to.
I closed the notes from the previous conversations and opened the hotel’s website again. This time, I tried to approach it as someone who had just come across the property for the first time, perhaps after a search or a recommendation among several alternatives. I did not know the team, had never walked through reception, did not know how they responded when something departed from the expected, and had no access to the reasons behind many of their decisions. All that remained in front of me was what the property had chosen to make visible and what other people had written about it.
At first, it was difficult to let go of a question I had used for years. Whenever I looked at a hotel’s website, I almost automatically tended to ask what it was saying. This time, I tried to replace that question with a considerably more demanding one: what could I reasonably say about this property based on what I found?
The difference seemed minor, but it completely changed the way I read the information. It was no longer enough to recognise what the hotel wanted to communicate; I needed to discover what someone could understand with nothing more than what was in front of them.
The website provided a great deal of information. I could learn the location, look at the rooms, explore the services, get a sense of the gastronomy, recognise the landscape and form a fairly precise idea of the level of comfort it offered. The photographs made it possible to imagine a stay, while the descriptions completed what the images could not show. If I had needed to explain what the hotel was physically like, I would have had enough material to do so. But we were looking for something else: whether someone who had never been there could discover that particular way of understanding the guest that we had found by looking at the property from within.
I returned to some of the texts with that single question in mind and found familiar expressions: personalised attention, hospitality, experiences designed around each guest, quality service. Read quickly, they seemed to confirm exactly what we were looking for. The problem appeared when I tried to determine what I had actually learned from them. A hotel claiming to offer personalised attention was telling me how it wanted to be understood, but it was not necessarily showing me why I should understand it that way.
There was nothing wrong with those words. Some of them even described what we had observed quite accurately. The difficulty was that they could be true and still offer very little to someone who needed to distinguish that property from another using the same expressions.
The difference began to become visible when I found much simpler information. One part of the website explained what happened when a guest arrived outside the usual hours: with advance notice, the team could coordinate reception according to the guest’s arrival time.
I stopped there for longer than such a small piece of information seemed to justify. We had spent several minutes reading far more ambitious claims without making much progress and yet an almost operational explanation was beginning to show us something concrete. If someone arrived late, the hotel had provided a way to adapt to that circumstance. That did not by itself demonstrate that the entire property was organised around an individual understanding of each guest, but it constituted a different kind of signal from a general claim. We were no longer dealing only with what the hotel said it valued; here was a situation in which that value could be seen at work.
I continued through the information with that distinction in mind and began to notice something that until then had gone almost unnoticed. For years, I had read hotel policies as an administrative part of the experience, a necessary but secondary area of communication. Check-in and check-out times, cancellations, breakfast, children, pets, special requests, booking conditions. They seemed to belong to a different language from the one we used when talking about identity, but now they were beginning to reveal just how much they too could describe the way a property understood its guests.
If a hotel defined itself as flexible while every one of its policies was built around the impossibility of departing from procedure, the two pieces ultimately described different properties. If it claimed to understand individual needs but offered no visible way to communicate an exceptional arrival, request an adaptation or raise a different circumstance, the promise began to lose some of its strength. By contrast, when a policy allowed for reasonable alternatives, explained how to request them and showed that certain exceptions formed part of the normal way of working, that apparently administrative information also began to become information about hospitality.
That did not mean that a good hotel should accept every request or that personalisation meant removing boundaries. There were operational, financial and safety constraints that no property could ignore. But even those boundaries could say something about the relationship the hotel proposed with the person on the other side. Two properties could have exactly the same breakfast hours and yet, through the way they explained what happened when a different need arose, convey two completely different ways of understanding service.
I moved on to the photographs. Until then, we had used them mainly to observe the image of the property they constructed: architecture, landscape, rooms, gastronomy, design, common areas. They were good photographs and fulfilled that function reasonably well. But when I tried to find the logic we were looking for within them, the result was less obvious.
An impeccably prepared room could show care, a well-presented table suggest attention to detail and a welcoming reception area convey a certain sense of closeness; yet many of those interpretations depended too heavily on what I already knew. If I removed that prior knowledge, the images made it reasonably safe to say that the hotel was comfortable, cared about its aesthetics and stood in an attractive setting, but it was much harder to conclude that it tried to understand each person before deciding how to look after them.
Those photographs were not wrong; they simply showed something else.
I began to distinguish more clearly between a piece that described the hotel accurately and one that helped demonstrate what made it distinctive. They were not always the same. A photograph could be excellent at showing a room and contribute very little to understanding a service culture, while an apparently secondary policy could do exactly the opposite.
When I moved on to the reviews, the reading changed again because scenes began to appear. One guest described arriving later than expected and someone remaining attentive to their arrival; another remembered an adaptation related to breakfast; in different reviews, members of the team were mentioned by name, not as part of generic praise but in connection with specific situations. There were comments about requests resolved, recommendations made after a conversation and small gestures that, viewed in isolation, might simply have seemed like examples of good service.
We had already encountered many of them during the previous conversations. The difference was that now I could not use them as pieces of a story we already knew. I had to ask what conclusion someone might have drawn if they had reached those reviews without any prior explanation.
When I mentally placed them alongside what I had just observed on the website, certain continuities began to appear. A late arrival corresponded with a policy that allowed for that possibility; the adaptation of a particular need resembled other situations described by people with no connection to one another; names appeared repeatedly where the hotel spoke, much more abstractly, of personalised attention.
No single source contained the complete conclusion or used exactly the same words, and perhaps that was precisely what made it interesting. The website could formulate a promise, certain policies could show that the operation was prepared to sustain it, and reviews could provide situations in which someone appeared to have experienced it. A photograph could provide context without needing to explain anything. When pieces from different places began to point in a similar direction, the characteristic acquired a consistency that none of them could have built on its own.
Identity was beginning to appear not only in each signal, but in the relationship between them.
Until then, I had thought that our problem consisted mainly of making visible what the hotel knew about itself. Now I was beginning to suspect an additional difficulty: leaving traces was not enough; those traces had to allow a reasonably compatible reading when someone brought them together.
I wanted to see how far that idea would hold, so I opened an artificial intelligence system. I tried to formulate as neutral a question as possible. I did not explain what we were looking for or give it the conclusion we had reached. I used only information available from the outside and asked what seemed to distinguish that property from other similar hotels.
The answer spoke of the setting, the design, the tranquillity, the gastronomy and certain attributes related to service. It was not wrong. Several of those characteristics were clearly supported by what we had just reviewed. Yet the idea we considered most important appeared weakened, almost lost among other qualities that were much easier to identify.
For a few moments, I thought the system had failed to understand the hotel properly, until a considerably more uncomfortable question emerged:
What reasons had it found to understand the hotel any differently?
We had needed to go through reviews, compare situations, observe decisions and devote several conversations to discovering a logic the property had never expressed particularly clearly. Now we expected someone who had not made that journey—human or artificial—to arrive immediately at the same conclusion using a set of signals in which some of the most visible characteristics were still the landscape, the rooms, the architecture or the gastronomy.
Perhaps the problem was not that artificial intelligence still did not understand the hotel well enough, but that the hotel had not yet left enough for it to be understood in that way.
That possibility changed the discussion considerably. Over the previous few years, I had heard more and more questions about how to make a property appear in the answers provided by new systems, what information should be published, which platform mattered most, or what needed to be done for an artificial intelligence to recommend a hotel. All were reasonable questions, but they were beginning to seem secondary to something far more basic: before asking how to make a system recommend us, perhaps we needed to know what it could reasonably understand about us from the information it found.
A recommendation, after all, could not be built solely on our intention to be recommended.
So I ran the test again, but changed the question. This time, I did not ask for a description of the property or try to discover what differentiated it in general terms. I introduced a person.
I imagined someone travelling with their elderly mother, likely to arrive late because of the flight schedule, who particularly valued tranquillity and needed some flexibility around breakfast. Then I asked whether, based on the information available, that hotel seemed like a good option for the trip.
The answer was different. The landscape remained relevant and the room still mattered, but some of the signals that had previously seemed secondary acquired a different value. The possibility of coordinating a late arrival ceased to be merely an operational condition and became a concrete reason why the hotel might be a good fit. Comments about breakfast adaptations began to address a specific need, while reviews recalling the team’s availability and attentiveness provided reassurance in a situation in which the traveller would probably value precisely those qualities.
We had not added a service, changed the hotel or improved its description. We had changed the person in relation to whom we were trying to understand the hotel and, in doing so, information that had previously seemed marginal began to participate directly in the decision.
That brought me back to the first conversation in this book. We had begun by asking what happened when travellers stopped searching for a list of hotels and started explaining their situation. Many chapters later, we were beginning to see the other half of the problem. To respond well to that person, it was not enough to know that the properties existed. It was necessary to understand what each could solve, for whom, under what circumstances, and with what degree of confidence that conclusion could be sustained.
A hotel could be extraordinarily well suited to one traveller and not to another. The difference that mattered, therefore, was not always a universally superior characteristic; it could be a capability that was particularly valuable in relation to a specific need.
That also changed the way I thought about communication. For a long time, we had tried to emphasise attributes capable of appealing to as many people as possible: better location, better service, greater comfort, unique experiences. Now I was beginning to wonder whether part of the work consisted precisely in doing the opposite: making certain characteristics understandable enough for someone to recognise when we were particularly well suited to them—and also when perhaps we were not.
Precision was beginning to seem more valuable than breadth.
I continued testing situations, changing the traveller’s profile, priorities and some of the circumstances of the trip. In each case, the same pieces acquired different weights. A decisive characteristic for one person could be completely irrelevant to another. An adapted breakfast ceased to be a curiosity when there was a dietary need; flexibility around a late arrival mattered little to someone planning to arrive in the middle of the afternoon; tranquillity could become the main reason for choosing the hotel for certain guests and remain a secondary characteristic for those seeking a different kind of experience.
I began to understand that what we called identity did more than explain who the hotel was. It could also help determine for whom it might be particularly valuable, and that relationship required more than good declarations.
If, in relation to a specific need, we found a promise from the hotel, a compatible policy, a visible operational decision and guest experiences that confirmed it, the recommendation began to rest on a reasonably solid foundation. If all we found was an advertising line, the conclusion might still be true, but we had fewer reasons to trust it.
The difference seemed obvious once seen and yet, for years, we had built much of hotel communication without paying much attention to it. We had devoted enormous effort to deciding what to say and far less to asking what could demonstrate what we were saying.
When I returned to the pieces of the hotel, a new difficulty emerged: not every signal seemed to carry the same weight. A statement made by the property itself did not produce exactly the same effect as an experience spontaneously described by different guests. A photograph could provide fairly strong evidence of the size of a room while being insufficient to demonstrate the quality of the service. A published policy could confirm an operational possibility but not necessarily show how it was carried out when someone actually needed it. Even a review capable of describing an extraordinarily specific experience was still, ultimately, the experience of one person.
The more I tried to organise those differences, the clearer it became that we had reached the limit of the question with which the conversation had begun. Someone could begin to discover the hotel without us explaining it; we ourselves had just done so in part by restricting ourselves to what could be found from the outside. But we had also established that the possibility of understanding it depended on the clarity of the signals available and on the relationship they managed to establish with one another. Some merely suggested a characteristic, others made it possible to observe it, and several, when they converged without needing to repeat the same words, began to support it.
I still did not know how to determine when that collection of signals was sufficient. Nor did I know whether we should place the same trust in a characteristic mentioned ten times by the hotel itself as in another spontaneously described by three guests, or how much it mattered for a signal to be recent, specific or independent of the others.
We had begun the conversation trying to establish whether an identity could be reconstructed from the outside and ended up facing a problem we had not previously needed to formulate: if future decisions were increasingly going to depend on systems capable of gathering, comparing and interpreting scattered information, it was no longer enough to ask what signals we were leaving. We needed to begin understanding which of them actually deserved our trust.
Management Perspective
A characteristic does not become understandable simply because the hotel declares it. It begins to acquire consistency when it can be recognised in decisions, services, policies, images and experiences coming from different places. The less the identity depends on the property’s own explanation of itself, and the more it can be reconstructed from what the hotel actually does and from what others experience and remember, the more understandable it becomes to someone who needs to decide whether that hotel is suitable before recommending it.
The Case Study
In the hotel we have been observing, we temporarily set aside what we had learned from the inside and tried to reconstruct it using only information available to someone who had never been there. Some characteristics could be recognised relatively easily, while others depended too heavily on general claims or scattered signals. When we introduced specific traveller needs, certain decisions, policies and experiences that had previously seemed secondary began to become reasons for choosing the hotel. The next step will be to understand why some signals allow a conclusion to be sustained with greater confidence than others.
A Question Before Moving On
If tomorrow your hotel could no longer explain who it was and could only show what it does, what it allows and what its guests remember, would someone reach the same conclusion as you about what makes it different?