Food and Beverage Operations: An Operator's Playbook for High-Volume Venues

Venue food and beverage looks like hospitality and behaves like manufacturing. You have a fixed window, a known peak, finite capacity at each point of sale, and no ability to recover a lost hour. The buildings that consistently out-earn their comparables are not the ones with the best menus. They are the ones that lose the fewest minutes.

This playbook sets out the operating model that produces that, and the four numbers worth managing. It is written for F&B directors, concessions managers and the concessionaires running these operations under contract.

The Window Is the Whole Business

A restaurant trades for eight hours and can smooth demand across them. A venue trades for two to four hours with demand concentrated into three spikes: the period before doors and first pitch, each intermission or period break, and one dominant peak that varies by sport and by event type.

Two consequences follow, and almost every operating decision descends from them.

First, capacity has to be sized to the spike rather than the average. A stand configured for average demand will fail at intermission, and intermission is where the money is. Second, and less obvious: anything that takes a point of sale out of service during a spike is unrecoverable revenue, not deferred revenue. There is no later.

Per-Cap Has Exactly Two Levers

Per-cap gets discussed as though it were a strategy. It is arithmetic: revenue divided by attendance. You can raise transactions completed, or raise average basket. That is the entire space.

The reason this framing matters is that it disqualifies a lot of activity. A new menu item that neither speeds a line nor raises a basket is a merchandising change, not a revenue initiative. Being strict about this makes prioritisation much easier.

The Three Failures That Look Different and Are Not

Ask three departments where F&B revenue leaks and you will get three answers. Operations says queues. Warehouse says stockouts. Maintenance says equipment. They are describing the same event from different seats: a point of sale that cannot complete a transaction.

The closed register

A card reader fault, a POS crash or a power issue takes a till out of trading. The staff are there, the product is there, and nothing can be sold. This is the most expensive of the three per minute and the most likely to be reported late, because it gets radioed rather than logged.

The empty shelf

A fast-moving item runs out mid-peak. The till still works, but the highest-velocity product is gone, so basket size drops and some guests leave the line entirely. This is a par-level and replenishment problem, and it is almost always predictable from last season's data.

The line that stops moving

Nothing is broken and nothing is out. The line is simply slower than arrival rate, so a queue builds and guests abandon. This is a capacity and layout problem, and it is the one most amenable to mobile and pre-ordering.

The reason to see them as one problem is that the response is the same: somebody has to know within seconds, and somebody has to be dispatched. This is what Delaware North addressed at American Family Field, where deploying Listo across F&B cut downtime by 70 percent by keeping stands stocked and running rather than discovering problems after a stand had already stopped trading. The full case study covers the mechanics.

Stocking: Par Levels Beat Judgement

Replenishment is where the most recoverable revenue usually sits, because it is a planning problem masquerading as an execution one.

The pattern in most buildings is that stands are stocked to a level someone set years ago, and mid-event replenishment happens when a stand lead notices and radios the warehouse. By the time product arrives, the peak has passed.

Three changes move the needle without capital spend.

1. Set par levels per stand from actual velocity, not per format - Two stands with identical build will sell very different mixes depending on where they sit and who walks past them. Use last season's sales by item by stand.
2. Schedule replenishment against the event clock, not on request - A recurring task that says "top up stand 12 at end of first period" arrives before the shortage rather than after the radio call. Listo supports scheduled recurring task requests for exactly this pattern.
3. Make the stockout report a routed request, not a radio call - The stand lead raises it in seconds, it reaches the warehouse runner directly, and it produces a record you can read afterwards to fix the par level permanently.

That third point compounds. A season of stockout records by stand and by item is the dataset that lets you stop having the same shortage every year. Our writing on warehouse efficiency and inventory management covers the tooling side.

Premium Is a Different Business in the Same Building

Suites, clubs and boxes operate on inverted economics: a small share of attendance producing a disproportionate share of revenue, with service expectations closer to fine dining than to a concourse.

The operational difference is that premium revenue is service-limited rather than capacity-limited. A suite guest is not queuing. They are waiting for someone to come back. Every minute of that wait is a minute they are not ordering, which is why response time in premium areas is a revenue metric rather than a satisfaction one.

The numbers bear this out. At Ford Field, Levy reports each Listo service request generates more than 100 dollars in food and beverage revenue. At TD Garden, Delaware North runs 90 premium suites with average response under five minutes and completed 1,472 requests across six months. At Great Wolf Lodge Niagara, cabana revenue rose 30 percent and average guest spend rose 9 percent after deployment. Across its customer base Listo reports an average 15 to 20 percent increase in F&B revenue, which is an average rather than a promise about any particular building.

Mobile and Pre-Ordering: What They Actually Fix

Mobile ordering is often sold as a guest experience upgrade. Operationally it is a capacity tool, and it fixes two specific constraints.

It raises transactions per hour where the bottleneck is the queue rather than the kitchen, because guests order from a seat instead of standing in a line. And pre-ordering moves revenue outside the trading window entirely, which is the only way to genuinely beat a fixed-window constraint. Guests browse, order and pay before they arrive, and the product is produced against a known order rather than a forecast.

Two cautions from experience. Mobile ordering with insufficient fulfilment capacity produces a worse guest outcome than no mobile ordering, because the guest has paid and is now waiting. And anything requiring an app download will see a fraction of the adoption of a QR code that opens a menu directly. Our Mobile Order and Pay product is built around no-download QR ordering with pre-ordering, a stand-alone kitchen display or an integration into your existing POS, and payment through Stripe or FreedomPay including Apple Pay and Google Wallet.

Labour: Deploy Against the Curve

Most venues staff by outlet and by shift. The demand curve does not respect either. Intermission needs three times the capacity of the second period at the same outlet, and no amount of scheduling fixes that if staff are fixed to a station.

The practical approach is to hold a share of your crew as floating capacity that can be moved to where demand actually appears, and to make moving them a single action rather than a series of radio calls. Listo supports dynamic staff assignment and mass staff assignment for reshuffling coverage across locations mid-event, which matters most on the nights when two people call out an hour before doors.

Across its venues Listo reports an average 15 percent reduction in monthly labour hours. The mechanism is not fewer people on the floor. It is less time spent walking to find someone, waiting for an answer and repeating instructions.

Menu Engineering for a Two-Hour Window

Menu design in a venue is an operations decision that gets made by marketing. The constraint is not what guests would enjoy, it is what your line can produce at rate during a fifteen-minute intermission.

The useful discipline is to score every item on two axes: contribution per unit and seconds of line time consumed. Items that are high contribution and fast belong at the front of the board and in every upsell prompt. Items that are high contribution and slow belong in premium areas and in pre-order, where production happens against a known order rather than in a queue. Items that are low contribution and slow should not exist in a general-admission stand, however popular they are, because they consume the capacity your fast items need.

The mistake worth avoiding is assuming popularity justifies keeping something. A slow item with a long queue behind it is not producing its own revenue, it is suppressing everything behind it in the line. Measuring transactions per hour by stand before and after a menu change is how you settle those arguments with numbers rather than opinions.

One practical move that works in most buildings: designate at least one point of sale per concourse as a limited fast menu, three or four items, single price point, no modifications. Throughput on those lines routinely runs at multiples of a full-menu stand, and they absorb the guests who want speed over choice.

Cashless, Card Readers and the Failure You Cannot Work Around

Most large venues are now effectively cashless, which is an operational gain almost everywhere and a single point of failure in one specific way. When a card reader fails in a cash-accepting operation, the till keeps trading. In a cashless building, that till is finished for the night unless somebody swaps the hardware.

Three things are worth having in place before the season starts. Spare readers held on each concourse rather than centrally, because a twenty-minute round trip to a back-of-house store is the whole peak. A defined swap procedure that a stand lead can execute without a technician. And a routed request path that puts a hardware fault in front of whoever holds the spares in seconds rather than through a supervisor.

This is the clearest example of why ranking assets by replacement cost misleads. A card reader is among the cheapest items in the building and among the most expensive to lose during a peak, because in a cashless operation its failure closes a point of sale completely.

The Four Numbers to Manage

• Lost trading minutes per point of sale. Minutes a till could not transact, by cause. Almost nobody has this and it is the most actionable number in venue F&B.
• Per-cap by area. Split general admission, premium and in-seat. A building-wide per-cap hides which of the three is underperforming.
• Stockouts per event by stand and item. Ranked. This becomes next season's par levels.
• Median service response time in premium areas. Read by hour. This is the closest thing venue F&B has to a leading revenue indicator.

Where to Start

Pick your three highest-volume points of sale and instrument lost trading minutes for one event cycle. Log every outage, stockout and fault as a routed request with a timestamp, and total the minutes at the end.

Most operators are surprised twice. The total is larger than they expected, and the dominant cause is not the one they would have named. That single number, in trading minutes rather than work orders, is usually enough to fund whatever you decide to do next. The wider operating model is set out in our intelligent venue management guide.