Front office KPIs: what a desk can measure and how to calculate it

A front office KPI is a number the desk controls, calculated the same way every time from a named source. Each one needs a formula, the report it comes from and a known trap. Occupancy is not a desk KPI, because the desk does not set rate or inventory. Check-in time, queue length, registration errors and void counts are.

The morning meeting opens with occupancy and average rate. The front office manager nods, though neither number was decided at the desk. The numbers the desk did decide, how long a guest waited and how many registrations had to be corrected afterwards, are usually not on the page at all. After reading this you will be able to build a one-page front office scorecard: measures the desk can actually move, each with a formula, a source and a known way to misread it.

What is a KPI in front office work?#

A front office KPI is a number that the desk's own actions move, calculated by a fixed formula from a named report, with one person responsible for it. The property system prints dozens of numbers every night; a KPI is one you have chosen to act on, with a written formula, an owner and a review day.

Apply the controllability test: if the number is bad on Monday, can the front office manager change it by Friday? If not, it belongs to revenue management, finance or the owner. Occupancy fails the test. Rooms sold come from rate, inventory and channel decisions made elsewhere; the desk can only convert the arrivals that reach it.

Write each entry in three parts: the formula, with numerator and denominator stated; the source, meaning which report, which timestamp and which system; and the trap, the specific way this number misleads. Then set a cadence per measure: speed and money-at-risk daily, accuracy and selling weekly, recovery cost monthly. A measure reviewed once a quarter is decoration.

Name one person per measure, not the team. A number owned by everyone is reviewed by no one.

For how this scorecard feeds the wider daily picture, see what a GM's morning page should contain.

How do you calculate check-in time and waiting time?#

Check-in time is the gap between the moment a guest reaches the desk and the moment the key is handed over; waiting time is the part of that gap spent standing in a queue. Measure them separately, and measure them in the peak hour.

Fix the two timestamps before calculating anything: arrival at the desk, from a queue ticket, a paper tally or the moment the guest is greeted, and key handover. The reservation record alone cannot give you the first one. Then take every arrival in the hour, subtract arrival time from handover time, and report the median and the 90th percentile. The median describes the typical guest; the 90th percentile describes the worst experience you are willing to defend.

Avoid the daily average. One group of thirty arriving at once moves the average and hides what happened to everyone else. Queue length is a count of arrivals waiting, taken at fixed intervals, for example every fifteen minutes, through the peak, reported as a peak figure rather than an average.

Phone answer time runs from ring to pickup in the phone system, and abandoned calls belong in the same report. A fast answer time with high abandonment means callers gave up before you picked up. Report walk-ins separately from reservation arrivals: they follow different steps, and mixing them makes both numbers unreadable.

The common trap is starting the clock when the agent opens the record. That measures typing speed and hides the wait, which is the part the guest remembers. To turn these numbers into staffing decisions, see how many desk agents a peak hour needs, from your own arrivals data.

How do you measure accuracy at the front desk?#

Count errors per hundred arrivals, not as a feeling: registration corrections, room status mismatches and folio corrections each get their own count and their own source.

Registration errors are guest records corrected after check-in, meaning name, identity number, dates or rate code, per 100 arrivals. The source is the change log in the property system, not a spot check by the manager. Room status mismatch is the number of times per day the desk's belief about a room disagrees with housekeeping's status at the moment of assigning it; count it at the moment, not at the end of the shift.

Folio corrections are charges moved, reversed or re-routed after posting, per 100 room nights. Corrections should be reversals with a reason attached, never deletions. Divide by volume in every case: raw counts rise with a busy month and make good staff look careless.

The quiet errors are the majority. Counting only what a guest complained about measures the guest's patience, not the desk's accuracy. Run a weekly sample as well: pull ten random folios and compare each against its registration record. The sample catches what the log misses.

How do you measure selling at the desk?#

Measure offers made and offers accepted as two separate numbers, then add walk-in conversion. The ratio between them tells you whether the problem is the offer or the ask.

Upgrades offered counts arrivals who were offered a paid upgrade; upgrades accepted counts those who took it. Never report a single conversion figure, because it cannot tell you which of the two to fix. A high acceptance rate with a low offer count usually means the desk only offers to guests who were going to buy anyway.

Walk-in conversion is walk-ins who became stays divided by walk-ins who asked. Record the ones who left and the reason, whether rate, availability or room type, or the number has no next step. Rate integrity, meaning stays sold below the published rate without an approved reason, is a selling measure, not a finance one, because the decision happens at the desk.

Timing matters. Offers work at booking, before arrival and at the desk; the desk only sees the last moment, so read its numbers as the end of a chain, not the whole of it. For the earlier moments, see matching each upsell offer to the right moment. Do not set a target on acceptance alone. It rewards discounting, which is the opposite of selling.

How do you measure money at the front desk?#

Money measures at the desk are about what leaves unrecorded: voids, unpaid balances at departure and deposit coverage at arrival.

Voids are the count and value of postings voided by desk staff, per 100 room nights, with the reason recorded. A correction should be a reversal, so the original stays visible. Deposit coverage is the share of arrivals that had the required deposit or pre-authorization in place before the key was handed over; this is the measure that prevents most end-of-stay disputes.

Unpaid balance at departure is the number of folios closed with a balance and their total value. Ageing the balance belongs to finance, but the desk creates it. Cash drawer differences should be counted and signed at every shift handover, per shift, not per month.

Separate the duties. The person who posts a charge should not be the person who voids it without a second check.

How do you measure service recovery at the desk?#

Count complaints closed within the shift, complaints that came back, and what recovery cost, and read the three together once a month.

Closure in shift is complaints logged and resolved before handover, divided by complaints logged. It measures the desk's ability to finish what it starts. Repeat issues are the same room or the same complaint category recurring within a month; a repeat is a process failure, not bad luck, and it belongs to the manager who owns that process.

Cost of recovery, meaning room moves, upgrades, complimentary items and allowances, should be tracked as one monthly figure rather than absorbed silently into the room rate. For the mechanics, see tracking allowances, upgrades and comps as a monthly number.

The trap: rewarding a low complaint count rewards not logging. Track complaints logged as well as complaints closed. Log at the moment it happens, in one place; a notebook under the desk is not a record and cannot be counted. Read recovery cost against complaint volume. Cost rising with volume is normal; cost rising while volume falls means the same problems are being bought off more expensively.

How do you pick five measures instead of fifteen?#

Take one measure from each of five groups, speed, accuracy, selling, money and service, write it down in full, and review the set on a fixed day each week.

A workable starter set: median check-in time in the peak hour, registration errors per 100 arrivals, upgrade offers and acceptances, deposit coverage, and complaints closed in shift. Write each line as formula, source, owner and review day. If a line cannot be filled in, the measure is not ready.

Run for four weeks before setting any target. The first month is a baseline, not a verdict, and a target set before the baseline is a guess. Retire a measure when it stops changing a decision: a number nobody acts on costs a line on the page and a minute in the meeting. Review the set once a quarter against the complaints guests actually made, not against what the desk finds easy to count.

What to do next#

Pick three measures from the scorecard this week, write the formula and the source next to each one, and run them for four weeks before anyone sets a target. Speed, accuracy, selling, money and service fit together as one picture: a fast desk that registers guests wrongly, or a careful desk that never offers an upgrade, is only half measured. The point is not to watch the desk more closely but to give it numbers it can actually move.