Planning a coffee-shop menu around morning service capacity

Start with the morning intervals, not the daily average

The published cafe has a 170-transaction mature day, 10 opening hours and a 30-transaction hourly planning ceiling. The day averages only 17.0 orders an hour. That arithmetic can conceal a morning queue. The model has no measured drink mix or arrival profile, so this article adds a separate worksheet rather than treating its daily capacity check as a peak-service trial.

Assume 60.0% of those transactions arrive during a 3-hour morning. That is 102 orders, averaging 34.0 an hour, already above the retained planning ceiling. Within that morning, the fictional busiest hour below contains 48 orders, with the remaining morning orders arriving in the other two hours. Count in 15-minute blocks and combine counter and online requests. Neither the morning share nor the burst is an industry average.

Define the promise as a complete drink-and-food order ready for collection. A payment timestamp, espresso shot finish and final pickup are different events. Keep those times separate when diagnosing a queue.

Build the menu from paid station work

This worksheet assigns one person to the drink station and another to order/payment/food/handoff. It is a hypothetical trial roster, not a roster encoded in the existing model. Replenishment, cleaning, interruptions and remakes receive a chosen 80.0% usable-time factor. That leaves 720 active seconds per position in each 15-minute block. Replace both the times and the allowance after testing the actual team.

Multiply each family share by its active seconds and add the results. The original mix needs 75.5 seconds of barista work per mixed drink order and implies 38.1 order-equivalents per hour at that position. The candidate mix needs 49.0 seconds and implies 58.8. The cashier/handoff assumption is 42 active seconds per order, implying 68.6 per hour.

These are station workload estimates, not a validated whole-shop capacity. Time grinder/steam occupancy separately from active hands, and test shared milk/refrigerator access. Two group heads cannot establish the rate of a staffed bar. The selected machine installation sheet confirms connections, not orders per hour. Manufacturer specification boundary

Check whether a short burst leaves unfinished work

The busiest block requests 18 orders, equivalent to an arrival rate of 72.0 per hour. Even the faster candidate cannot process that block without some backlog under these assumptions. Use unfinished work as a first screening measure rather than promising a wait time from the daily average.

The recurrence is: next backlog = the greater of zero and previous backlog plus arrivals minus the block service allowance. The published ceiling allows 7.5 order-equivalents per block; the original and candidate station worksheets allow 9.5 and 14.7. Calculations keep the unrounded rates. Decimal backlogs describe work equivalents in a continuous-flow screen, not fractional drinks handed to customers. It assumes work becomes available through each block; clustered arrivals and actual item sequences need a timed trial.

The published-rate column ends with 18.0 unfinished order-equivalents. The original mix worksheet still ends with 11.4, while the candidate clears its burst by the final block. This does not establish a customer wait-time guarantee or change the published 30-per-hour model input. It identifies a candidate worth testing, provided customers actually choose that mix and upstream prep, equipment and handoff can support it.

Possible changes include making a batch-brewed option easy to order, restricting complex modifiers during the rush, simplifying milk-drink sizes or scheduling online release to the measured capacity. Moving orders to a separate screen does not create a second bar. A faster item only helps when it replaces constrained work that customers were otherwise going to request.

Compare contribution with scarce bar time and batch yield

A higher single-drink margin is not the only peak decision. The following isolated prices and recipe costs are hypothetical, and the payment-fee allowance remains the case assumption of 3.0% of sales. Brewed ingredients cost $0.350 per produced portion, milk-drink ingredients $1.050 per sold drink, and packaging $0.350 per sold drink. The direct contribution calculation subtracts these costs and the fee once.

The brewed item has lower contribution per drink but more contribution per assumed barista active minute. That ratio diagnoses a scarce-work tradeoff; it is not shop profit, utilization or a reason to ignore customer preference. The existing $8.25 ticket is a whole customer transaction that can include food. Do not substitute these single-drink prices into it or invent a food attachment to force the worksheet to match.

Batching moves preparation earlier and can waste ingredients. If 12 brewed portions are made and 10 are sold, ingredient cost is spread across the sold portions, while cups are purchased only for the sold drinks in this example. Direct ingredients plus packaging become $0.770 per sale, reducing contribution to $2.625. Paid brewing, refill and cleanup time still sits in the roster. Measure yields rather than assuming every batch sells out.

Keep preparation and cold holding inside the service plan

A short menu must preserve its food controls when the line is busy. Locate milk and food storage, handwashing, washing and waste access so replenishment does not block service. Record the recipe steps, batch method and storage/holding arrangement with the equipment layout. In the King County example, the menu and preparation method inform the equipment review. Its plan guide is used here for scope, not for older printed fee figures. Menu and equipment review

As a Washington-specific example, the general cold-holding limit for time/temperature control for safety food is 41°F or less, with the rule's stated exceptions. Do not improvise time without temperature control because orders are waiting; use the actual approved procedure and responsible authority. Other jurisdictions may apply different adopted requirements. Current Washington section and exceptions

Before expanding beyond drinks and bought-in pastries into cooked breakfast items, check preparation, ventilation, wash-up, storage and approvals. A new hot-food process is a changed operating scope, not a free way to increase the average ticket.

Price additional cover without charging existing labor twice

First map receiving, brewing/prep, service, breaks and cleanup to paid shifts. The published case already includes employee payroll and compensated owner work; no unassigned free owner is created by this worksheet. Moving an existing funded hour into the peak may be a roster change, while adding an hour is a cost change. A second barista also needs usable work space and shared equipment capacity.

If the test truly requires an additional 3 paid hours on 5 days each week, use 52/12 weeks per month. At the existing $18.00 hourly wage and 12.0% modeled payroll loading, incremental cover costs $1,310.40 a month. At the retained exact transaction contribution of $5.425, that is 8.1 additional transaction-equivalents per trading day to cover the added pay alone. Wage/loading are assumptions, not a current hiring quotation or complete statutory rate.

This staffing comparison is not added to the published model. Avoid double counting if hours can be redistributed inside the existing roster. If an actual trial supports different served volume or more paid labor, revise and rerun the monthly case together rather than changing only an appealing revenue line.

Run the mixed-order trial before adopting the menu

Use the actual counter, equipment and paid team. Release a realistic mixed sequence including food, modifiers, online orders, refills and remakes, rather than timing one clean latte repeatedly. Run the original and candidate menu under comparable conditions, then repeat the peak burst. Save unsuccessful periods as well as successful ones.

Choose the menu and ordering promise only after the observations show that the paid team can recover from the relevant burst while preserving yield, food controls and contribution. If demand chooses more milk drinks than the candidate assumes, rerun that mix; do not label a theoretical rate as the new business capacity. The transaction economics guide connects confirmed service volume to contribution, while the monthly cash scenarios show why extra cover needs both earnings and cash support. Keep the trial evidence with the operating plan.