Hotel KPIs: When Good Results Hide Bigger Problems

I remember one morning when I reviewed the hotel’s key performance indicators and, apparently, there were not too many reasons for concern. Occupancy was performing reasonably well, average rate was holding steady, reputation had not seen any significant movement, and labour costs were close to budget. On paper, the hotel was operating correctly. However, when I started walking the operation, I found a far less reassuring reality: rooms reaching Reception too late, certain incidents beginning to recur, team members accumulating additional hours, and guests describing, in different words, virtually the same problem. The dashboard looked fine. The hotel was beginning not to be. That experience reinforced a conviction that has grown over the years: an indicator can tell the truth and, at the same time, conceal an important part of it from us. A reputation score of 9.0 may be entirely accurate while concealing a recurring issue at breakfast. Labour costs within budget may coexist with excessive overtime, turnover or burnout. Excellent occupancy may have been built on an unprofitable business mix. Growing RevPAR may make us happy while costs absorb an ever-larger share of the result. Numbers do not necessarily lie. The problem arises when we ask them questions they cannot answer. In Hospitality we have a particularly intense relationship with indicators. We need them. We operate complex businesses, with perishable inventory, changing demand, multiple departments and hundreds of daily guest interactions. It would be irresponsible to manage them through intuition alone. Occupancy, ADR, RevPAR, GOPPAR, labour cost, reputation, productivity, average spend, conversion or satisfaction are essential tools. But a tool does not replace judgement. Measuring more does not necessarily mean understanding better. Moreover, the current context makes this distinction between revenue, activity and profitability even more important. In August 2026, STR and Tourism Economics projected RevPAR growth of only 1.2% in 2026 across the European markets they analysed, driven mainly by ADR, while warning that occupancy growth is becoming increasingly difficult to achieve. At the same time, STR insists that metrics such as GOPPAR provide a better understanding of revenue conversion into profit because they incorporate the relationship between revenue and operating expenses, something RevPAR alone cannot explain. That is why I am increasingly less concerned with having many indicators and more concerned with having indicators capable of prompting good questions. A dashboard should help us identify deviations before they become negative results, rather than simply confirm at month-end what has already happened. Good hotel management is not merely about looking backwards and checking the result. It is about recognising the weak signals that emerge long before a problem reaches the P&L, reputation or guest experience.

When the average starts to mislead us

Averages are extraordinarily convenient. They reduce hundreds or thousands of different situations to a manageable number. That is precisely their advantage. And also their danger. Imagine we have an average satisfaction score of 9.1 out of 10. Excellent. But we may be facing entirely different scenarios. Perhaps virtually all our guests are satisfied. Or perhaps a significant share are absolutely delighted while another segment is beginning to experience a specific problem. The average may be identical. The operational reality is not. That is why, when I analyse reputation, I try never to focus solely on the score. I am far more interested in other questions: What topics recur repeatedly? Which departments receive the most negative mentions? On which days? At what times? With which room types? With which segments? After what type of stay? Are we seeing isolated issues or patterns? There is an enormous difference between one bad review and five different reviews describing the same flawed process. The first may be a one-off circumstance. The second is beginning to become management information. And here we encounter one of the most frequent errors when working with KPIs: confusing stability with the absence of problems. A score can remain stable while its underlying composition changes completely. We may lose exceptionally satisfied guests and replace them with merely satisfied guests without immediately seeing a major drop in the average. We may have a growing issue in one department that is statistically offset by improvement in another. We may maintain the same overall result while variability increases significantly. The average reassures. The distribution explains. Something similar happens with occupancy. Few figures create such an immediate sense of success as seeing a full hotel. Occupancy of 90%, 95% or 100% generates activity, energy and an obvious perception of business. But I have learned to ask myself another question: Full of what? Because two hotels with identical occupancy can deliver entirely different results. One may have protected rate, selected segments, optimised distribution and generated additional revenue. Another may have filled through discounts, expensive channels, highly flexible terms and low-contribution business. Same occupancy. Different businesses. Even within the same hotel, the same 90% can mean entirely different things depending on how we got there. That is why I believe occupancy should always be interpreted alongside rate, mix, acquisition cost and contribution. ADR cannot be analysed in isolation either. We may increase average rate simply because we sold fewer low-priced rooms. That does not necessarily mean we have improved our pricing capability. We may increase ADR and reduce total revenue. We may increase it thanks to an extraordinary event that is difficult to repeat. We may have an excellent ADR and excessive dependence on a particular segment. The number is correct. The interpretation may not be. The same applies to RevPAR. It is probably one of the most useful indicators we have because it combines occupancy and rate. But it does not answer every question either. RevPAR essentially measures rooms revenue performance against available inventory. It does not tell us how much it costs to generate that revenue. It does not explain the contribution from food and beverage. It does not measure commercial costs. It does not sufficiently capture differences in operational structure. It does not tell us how much money we ultimately retain. This is why indicators such as TRevPAR and GOPPAR are so important when we want to get closer to a more complete view of the business. This distinction becomes particularly important when costs rise. A hotel can improve RevPAR and worsen its margin. There is no contradiction. We are simply observing two different dimensions. Generating more revenue does not guarantee earning more profit. Another KPI that deserves particular care is labour cost. Imagine our target is to keep it at a certain percentage of revenue. We finish the month exactly on budget. Excellent. But we do not necessarily yet know whether we have managed well. Perhaps we achieved that result by appropriately reducing hours, improving productivity, planning better or adapting shifts to demand. But we may also have achieved it by not filling vacancies, overloading certain roles, reducing training, accumulating rest days, generating additional hours that will appear later, or accepting a silent deterioration in service. Same KPI. Radically different realities. This is where people indicators begin to gain importance. And current data should make us think. Gallup places employee engagement in Spain at just 10% in 2025, compared with 12% in Europe and 20% globally. Seventy-six per cent of Spanish employees are classified as not engaged, while another 15% are actively disengaged. It is not a Hospitality-specific statistic, but it would be naive to think our sector operates outside this phenomenon, particularly as it is such a people-intensive industry. A sound labour cost does not automatically offset a disengaged team. We may save today and pay tomorrow through turnover, absenteeism, lower productivity, poorer guest experience or loss of knowledge. Financial indicators usually measure very well what has already happened. People indicators can warn us about what is likely to happen next. There is another metric I always view with caution: productivity. Rooms per housekeeper. Guests per waiter. Revenue per employee. Cost per occupied room. All can be extraordinarily useful. The problem arises when we try to maximise productivity without first determining the level of service we want to protect. A housekeeper can clean more rooms. The question is how far. A receptionist can process more arrivals. The question is at what quality. A waiter can cover more tables. The question is how much service time changes. Productivity should not measure only how much we do with a resource, but how much value we can create without undermining what we aim to deliver. Otherwise, we can improve a KPI remarkably while deteriorating the product. I am particularly concerned when an improvement in productivity coincides with more rework, more complaints, more overtime, more incidents, more errors, more turnover or lower satisfaction. At that point, the indicator is not wrong. Our concept of productivity is too narrow. Over the years, I have developed a kind of healthy mistrust of some indicators when they appear too perfect. Not because I think they are false. But because I always want to know what lies beneath. There are several situations I try to review systematically:

Stable reputation with recurring comments. If the average does not change but the same friction appears repeatedly, I prefer to act before the problem becomes large enough to move the score.
High occupancy with weak profitability. I review ADR, channels, commissions, discounts, variable costs, segments and displacement. A full hotel can conceal unprofitable business.
Growing ADR with the loss of certain segments. Raising rates can be an excellent decision, but we need to check who stops buying from us and whether that change forms part of the strategy.
Controlled labour cost with increasing overtime or turnover. It may indicate that we are shifting the problem into future months or financing savings through unsustainable effort.
Excellent overall satisfaction with one clearly weak touchpoint. Guests may rate the overall experience very highly and yet there may still be a significant operational opportunity in check-in, breakfast, housekeeping, maintenance or departure.
Low number of complaints. It does not always mean happy guests. It may also mean guests do not consider it worthwhile to complain and simply do not return.
Many incidents resolved quickly. It may seem like an efficiency indicator. If they are always the same, perhaps we are measuring our ability to put out fires rather than our ability to prevent them.
Budget achieved. Meeting a budget does not automatically make a result excellent. Perhaps the budget was wrong, the market offered a greater opportunity, or we achieved the target by sacrificing a variable we are not measuring.
Growing conversion. It may result from a better purchase experience, but also from prices that are too low or excessively permissive terms. We need to understand the quality of that conversion.
Growing revenue. It is always worth celebrating, but only after checking how much incremental margin we are achieving and what additional resources were required to secure it.

This is probably the central idea I try to convey to my teams: a KPI should never close a conversation; it should open one. If occupancy is 94%, the question does not end there. Is it the right occupancy? At what ADR? With what mix? At what acquisition cost? With what profitability? With what operational impact? And what demand have we turned away? If reputation is 9.2, it does not end there either. What is improving? What is worsening? Which segment gives the lowest score? Which mentions recur? Which problems are still too small to affect the average? That last type of question particularly interests me. Major problems rarely begin as major problems. Before that, they were signals. One room delivered late. Then three. One noise complaint. Then several. One voluntary resignation. Then turnover. One breakfast complaint. Then a pattern. A small cost deviation. Then lost margin. The work of anyone in a leadership role consists, to a large extent, of recognising those signals while they are still small.

From dashboard to early-warning system

Over the years, I have stopped thinking of the dashboard simply as a collection of KPIs. I prefer to see it as an early-warning system. There are indicators that tell us what we have achieved. And others that can anticipate what we will achieve if we continue doing exactly the same thing. The former are outcome indicators. The latter are process or leading indicators. We need both. An outcome indicator might be monthly reputation. A process indicator might be rooms not available at the scheduled time. An outcome: labour cost. A leading indicator: overtime, absenteeism or unfilled vacancies. An outcome: food and beverage revenue. A leading indicator: covers, average spend, table time, sales mix or cancellations. An outcome: RevPAR. A leading indicator: pickup, booking pace, rejected demand, changes in booking window or performance by segment. When we combine both types of information, something very interesting happens: we can intervene before the negative outcome appears. That completely changes the way we manage. We no longer wait for reputation to fall before discovering a problem. We observe the causes that could probably make it fall. We do not wait for labour costs to soar. We monitor shifts, overtime, productivity and absenteeism. We do not wait for Revenue to lose momentum. We analyse pickup and demand. We do not wait for a department to start receiving complaints. We monitor incidents and rework. The EFQM approach has long emphasised the relationship between direction, execution and results. I find it especially useful in Hospitality because it reminds us of something fundamental: results do not emerge spontaneously. They are the consequence of systems, processes, people and decisions. That is why one of the most useful exercises we can undertake with every KPI is asking ourselves: What behaviours and processes produce this result? If we want to improve reputation, perhaps we should also measure waiting time, first-contact resolution, incidents per room, rework, rooms ready, response times or recurrence of certain reasons. If we want to improve rooms profitability, perhaps we should complement RevPAR with acquisition cost, contribution by channel, net RevPAR, TRevPAR, GOPPAR or margin by segment. If we want to improve our people outcomes, we should monitor turnover, absenteeism, overtime, vacancies, internal promotion, training, management stability or engagement. This is not about filling a screen with figures. That is another common mistake. I have seen dashboards so comprehensive that they were virtually impossible to manage. When everything is a KPI, nothing is a priority. The objective should be to select a small set of indicators capable of answering the hotel’s strategic questions. What do we want to improve? What might prevent it? What signals would appear first? What data would allow us to intervene? That is where a useful dashboard begins. It is also important to work with segmentation. Averages begin to reveal far more when we break them down. I do not want to know satisfaction alone. I want to see it by market, channel, room type, length of stay, arrival day, board basis, segment, nationality, time band or any relevant dimension. I do not want ADR alone. I want to know which segments build it. I do not want labour cost alone. I want to know it by department, time band and activity volume. I do not want incidents alone. I want recurrence, severity, department and cause. Segmentation turns a number into a story. And stories are what enable decisions. I also always try to differentiate between normal variation and a real signal. Not every movement merits intervention. A hotel is a living organism. There is statistical noise. Better days. Worse days. Exceptionally critical guests. Unrepeatable situations. The danger lies in overreacting to every change. We can end up managing normal fluctuations as if they were trends. That is why I prefer to look for recurrence, trend, concentration and relationships between indicators. One negative review does not define a trend. Five mentions of the same issue begin to do so. One poor day of productivity may be explained by a circumstance. Four weeks of deterioration require attention. One person leaving is part of the life of any organisation. A series of departures in the same department tells a story. This is where another tool I consider very powerful comes in: cross-referencing indicators. If occupancy rises, what happens to reputation? If ADR rises, what happens to conversion? If we reduce staffing, what happens to overtime? If housekeeping productivity increases, what happens to rework? If we increase F&B revenue, what happens to margin? If we reduce check-in time, what happens to satisfaction? An isolated KPI describes. Two or three related KPIs begin to explain. Sometimes, we will discover uncomfortable relationships. We may observe that our highest-occupancy days systematically coincide with poorer service scores. That tells us something. Perhaps our commercial capacity exceeds our operational capacity. We may see that a certain level of productivity coincides with an increase in incidents. That tells us something. We may find that a promotion increases conversion but reduces ADR so much that it barely generates incremental contribution. That also tells us something. This is precisely what I expect from an information system: that it challenges our intuitions. Not that it confirms them. And there is still one source of information that no dashboard should replace: the real hotel. Walking. Observing. Talking. Listening. Reading full reviews. Sitting for a while at Reception. Going into the laundry. Watching breakfast service. Joining the housekeeping team. Asking Maintenance. Listening to Reservations. Talking to guests. Indicators show patterns. The operation provides context. When both tell different stories, it is worth investigating. I have identified major problems simply by asking: “What makes you lose the most time every day?” It is an excellent question. The answers usually identify slow systems, absurd processes, unnecessary approvals, duplication, communication failures or tasks nobody has questioned for years. Another equally powerful question is: “What problem are we solving continuously?” That is where major opportunities are usually hidden. Because a problem we solve every day may no longer be an incident. It is part of the process. And it has probably become time to change the process. There is also a human dimension that data can conceal. An organisation can operate correctly for months thanks to the extraordinary effort of certain people. The KPIs will look good. The operation will keep running. Guests will be satisfied. But we will have built up a debt. A debt of fatigue. Of outstanding holidays. Of professional development. Of recognition. Of patience. Of motivation. Sooner or later, it emerges. And it often does so in an apparently sudden way: a sick leave, a departure, a conflict, a drop in productivity or a chain of errors. That is why I find it significant that, in Gallup’s latest data, only 12% of European employees are engaged at work, compared with 20% globally. We should not use this figure to diagnose our hotels automatically, but it should remind us that people metrics deserve a genuine place on the dashboard, not merely an annual review by Human Resources. Guest experience and employee experience are not separate worlds either. A large part of our product is created precisely through that interaction. The guest does not know whether the hotel is meeting its budget. They know whether someone looked them in the eye. They do not know our cost per occupied room. They know whether their room was clean. They do not know our F&B productivity. They know how long dinner took. They do not know the turnover rate. But they will experience its consequences. That is why I increasingly believe in an integrated view of Hospitality: guest, people, operations and financial performance are parts of the same system. Optimising one while permanently ignoring the others ultimately creates imbalances. And imbalances usually appear first in the details. My first piece of advice would be to review your dashboard tomorrow and ask an uncomfortable question about each indicator: what could be happening beneath this number without my seeing it? If you have excellent labour costs, look at additional hours, absenteeism and turnover. If you have outstanding reputation, look for recurring negative reasons. If you have high occupancy, study mix, margin and acquisition cost. If you have good RevPAR, check how much ultimately converts into profit. Do not look for problems artificially. Look for explanations. The second would be to select three or four processes that are especially important for your hotel and add leading indicators to them. Do not wait for reputation to fall; measure the frictions that could make it fall. Do not wait for a labour variance; observe the signals that precede it. Do not wait to lose profitability; monitor contribution. A good KPI is not merely one that accurately describes the past. It is one that still gives us time to change the future. And the third would be to preserve something no system should eliminate: managerial curiosity. When an indicator appears too good, ask why. When it is poor, avoid immediately looking for someone to blame and first look for the process producing it. Talk to those closest to the work, test data against reality and distrust any figure that claims to explain, on its own, a business as complex as a hotel. After many years of looking at dashboards, I still believe numbers are indispensable. But I have also learned something equally important: a hotel is not managed by looking at indicators; it is managed by understanding what lies behind them.

Explore All Articles & Resources

The Hotel KPIs That Reassure You May Be Hiding Bigger Problems

This article ends here. The archive does not.

What would you like to explore next?

The Tyranny of Urgency: Why Many Hotels Are Busy but Poorly Managed

Turn reading into actionable knowledge

Latest from Lead Hospitality

The Smart Hotel: Why I Wrote This Book