Shift Handover in Hospitality: The Fifteen Minutes That Decide the Night

Ask a hotel duty manager or a venue operations lead where their worst service failures come from and a surprising number will point at the same moment: the shift change. Not the busy period, not the sold-out night. The twenty minutes where one team stops being responsible and another starts.

It makes sense once you look at what is happening. Every open loop in the building has to survive a change of custody, carried out by people under time pressure, mostly by talking. Nothing about that process is designed. In most venues, handover is a habit rather than a system, and it works right up until the night it does not.

This guide sets out a handover process that holds together on a busy floor, and a template short enough that people will actually use it.

What Handover Is Actually Transferring

There are four distinct things moving across a shift change, and they carry very different levels of risk. Most handovers spend their time on the least dangerous one.

The ordering matters because handovers get cut short constantly. If your running order starts with standing context, the open work gets dropped when someone needs the supervisor at the gate. Front-load risk.

Why Verbal-Only Handover Degrades

A verbal handover is a copy operation with loss. The outgoing supervisor holds twenty items in their head, says twelve of them, and the incoming supervisor retains eight. That is not incompetence, it is normal human working memory under time pressure with a radio going.

The loss compounds across shifts. By the third handover, an item raised on the morning shift has either been written down somewhere or it no longer exists. In hotels running three shifts, and venues running split shifts around an event, this is the mechanism by which a guest request raised at 2pm gets rediscovered at 9pm by the guest.

The fix is not to write more. It is to make the record a by-product of doing the work rather than a separate documentation task. If every request already carries an owner, a status and a timestamp, the handover log exists whether anyone maintained it or not. That is the real operational argument for a request system over a radio: not speed, but memory.

The Four-Block Handover

Here is the structure we see working. It is deliberately short and deliberately ordered by risk.

1. Open items - Read out everything unfinished, and assign each one to a named person on the incoming shift before moving on. If your system shows pending and in-progress requests, this block is reading a screen rather than reciting from memory.
2. Exceptions - What is not normal tonight. Closures, outages, staffing gaps, anything a guest might ask about that has a non-standard answer.
3. Commitments and VIPs - Promises made, comps authorised, holds in place, guests who need attention and why. Include who made the promise, because that is the person the guest will name.
4. Equipment and stock - What is broken, what is low, what has a contractor coming. This is the block most often skipped and most often the cause of the next shift's biggest problem.

Then one closing question from the incoming supervisor: what are you worried about? The answer is frequently something that fits none of the four blocks and matters more than all of them. Judgement does not fit in a template, which is why the verbal exchange still has a job to do.

A Template You Can Use Tonight

Keep it to one page. A handover form that takes ten minutes to fill in does not get filled in.

• Shift, date, outgoing and incoming supervisor. Names, not initials.
• Open items. One line each: what, where, who owns it now, promised by when.
• Exceptions. Closures, outages, staffing gaps, with an expected fix time where one exists.
• Commitments. Guest or group, what was promised, who authorised it.
• Equipment. Item, location, status, contractor and ETA if applicable.
• Stock. Anything at or below par that will not survive the next service period.
• Incoming supervisor sign-off. A signature is not bureaucracy here. It is the moment custody actually changes.

That last line does more work than it looks like it does. An unsigned handover has no moment of transfer, which means both supervisors can reasonably believe the other still holds an item. A signature closes that ambiguity.

Handover at the Individual Level

Supervisor handover gets attention. Staff-level handover almost never does, and it is where a large share of guest-visible failures actually originate.

A suite attendant going off shift has three or four live relationships in progress: a table waiting on a second round, a guest who asked about a dietary substitution, a party that wants the check at the end of the third period. If she hands those off by telling whoever is nearest, they are gone.

Two mechanisms fix this without adding process. First, if requests are assigned rather than broadcast, the outgoing attendant can reassign her open items in the system before she leaves, and the incoming attendant sees them in a queue rather than hearing about them. Listo supports dynamic staff assignment for exactly this, and mass staff assignment when a whole block of locations needs to move to different coverage at once.

Second, make the last action of every shift explicit: clear or reassign your queue before you clock out. Thirty seconds, and it removes the single most common cause of a guest being told nobody knows about their request.

Hotels: Three Shifts, One Set of Rooms

Hotels have the hardest version of this problem, because the same physical inventory passes through three teams a day and guest requests routinely span shifts by design. A late checkout agreed at 8am has to survive two handovers to mean anything at 2pm.

The departments that most often lose things between each other are front desk, housekeeping and engineering, and the reason is usually that each has its own log. A maintenance issue reported by a room attendant lives in housekeeping's notes; engineering hears about it when someone walks over. The guest hears about it when they check in.

Running those requests through one system rather than three logs removes the handover entirely for cross-department work, because there is nothing to transfer. The request already sits in engineering's queue with a status the front desk can see. We cover the mechanics of that in more detail in our writing on transforming the hotel guest experience through technology.

The Overlap Question

Almost every venue eventually argues about whether shifts should overlap, and the argument is usually framed as a cost question. It is really a risk question.

With no overlap, handover happens at a boundary where one team is trying to leave and the other is trying to start. Both are distracted, and the outgoing supervisor has every incentive to compress. With a fifteen-minute paid overlap, the outgoing supervisor is still on the clock while the incoming one takes custody, which changes the quality of the conversation entirely.

The honest counterargument is that overlap costs real money. Fifteen minutes for two supervisors across three daily changeovers is a meaningful annual number in a large building. The way to settle it is to price the other side: take your duplicate-request count in the thirty minutes after a shift change, multiply by what a service failure costs you, and compare. At Ford Field, Levy reports each service request carries more than 100 dollars of food and beverage revenue, which tends to make the overlap look cheap in premium areas even if it does not elsewhere.

A middle path that works in venues with tight labour budgets: overlap only the changeover nearest to doors, and only for supervisors. That is where the loss concentrates, and it is the one handover where being rushed does the most damage.

What Changes When the Record Is Automatic

The operators who fix handover properly tend to describe the same shift in what the conversation is about. It stops being a recital and becomes a discussion of the two or three things that need judgement, because the routine open items are visible on a screen and do not need to be spoken.

At TD Garden, Delaware North runs Listo across 90 premium suites and completed 1,472 guest requests over six months with average response under five minutes. In a building with that volume, a verbal handover of open items is not realistic at all. The only way an incoming supervisor can know what is outstanding is to look at a live queue of pending and in-progress requests, which is what makes the handover short rather than long.

There is a second effect worth naming. Once handover is based on a shared record rather than one person's recollection, disagreements about what was handed over stop happening. That sounds like a small thing. Anyone who has spent a Monday morning reconstructing who knew what on Saturday night will recognise it as anything but.

Measuring Whether Handover Is Working

Handover quality is measurable, which surprises most operators. You are looking for evidence of rediscovery: work the incoming shift had to find out about the hard way.

1. Requests created within thirty minutes of a shift change that duplicate an earlier request. A duplicate is a guest asking twice, which means the first one did not transfer.
2. Age of the oldest open request at handover. If items regularly cross two boundaries, they are not being owned, they are being mentioned.
3. Requests reassigned versus abandoned at shift end. Abandoned queues are the leading indicator of tomorrow's complaint.
4. Escalations in the first hour of a shift. A spike here almost always means the incoming shift started blind.

Read these by shift and by day of week. In venues the pattern is usually concentrated: one particular changeover, often the one that lands closest to doors, produces most of the loss. That is a fixable staffing and timing problem once you can see it, and effectively invisible if handover leaves no trace.

Where to Start

Pick your riskiest changeover, which in a venue is normally the one nearest to doors and in a hotel is normally the overnight to morning transition. Run the four-block structure there for two weeks with a one-page written record and a sign-off. Do not roll it out everywhere first, because the version that survives contact with your building will not be the version you designed.

Then look at what the written records tell you. Almost every operator who does this finds the same thing: the items that kept dropping were never the complicated ones. They were ordinary requests that simply had no owner for twenty minutes. If you want to see how assignment, reassignment and a live request history work in practice, our platform page walks through it, and the intelligent venue management guide puts it in the context of the wider operation.