Operations
Part
3
10
min read
A faster kitchen can mean more revenue
How waiting times in your kitchen arise, why stations block each other, when speeding up can mean extra revenue and how to measure ticket times yourself.

Do this today
Tonight, measure twenty tickets with paper and a clock.
According to CBS (Statistics Netherlands), revenue at eating and drinking establishments in the second quarter of 2026 was 3.5% higher than a year earlier (restaurants: 3.2%). According to KHN Kompas, KHN's own research (Royal Dutch Hospitality Industry Association; sample and comparison period not stated), average revenue per venue fell by 1.3% in that quarter and the number of tickets decreased by 3%. The two sources do not measure the same thing and do not point in the same direction; both cover a whole quarter and not a single venue. They do not say how busy your Friday evening is, and that is what this is about: what does it mean if tickets leave the kitchen faster? This is part 3 of the series Where does the profit in your restaurant go? You will read how waiting times arise in a kitchen, why stations can block each other, when a faster kitchen can mean extra revenue and when it cannot, and how to measure ticket times yourself. Numbers in examples are made up, unless a source is given. Data as of 18 November 2026.
How waiting times arise
Waiting time usually arises where work comes in faster than a station can finish it.
Peaks in tickets. Tickets can arrive in waves, for example a party of twelve. The hourly average hides the wave.
A station that is the bottleneck. The station with the most work relative to its capacity limits how many dishes per hour leave the kitchen. When the inflow gets close to that limit, the waiting time rises first slowly and then quickly, according to queueing theory (Little, 2011).
Work that piles up. Worked example: a grill finishes 24 steaks per hour (one every 2.5 minutes); at the peak, 30 come in. The pile grows by 6 per hour: after an hour, the next steak waits 6 × 2.5 = 15 minutes before a cook starts on it.
Little's law. In 1961, John Little showed that in a queueing system the average number of units in the system equals the average number arriving per unit of time, multiplied by the average time a unit spends in the system (under conditions, such as finite averages and stationary processes). For your kitchen: ticket time = average number of open tickets divided by the number of tickets per minute. If you know two of the three numbers, you can calculate the third. Over an evening on which your kitchen starts empty and ends empty, this holds exactly for the measured averages (Little, 2011). It is about averages, not about a single ticket.
Worked example: if a ticket comes in every 3 minutes (20 per hour) and on average 5 tickets are open, the ticket time is 5 × 3 = 15 minutes; with on average 10 open tickets it is 30 minutes.
Conversely: if 5 tickets stay open and the ticket time falls to 12 minutes, then 5 ÷ 12 × 60 = 25 tickets per hour can pass through.
Why stations block each other
A table usually gets its dishes only when the slowest dish is ready: the pass waits for the last plate. Worked example with made-up numbers: table 12, four guests, four main courses. Each dish below shows how long it waited before a cook started on it and how long the preparation took.
Cold kitchen, goat cheese salad: 2 minutes waiting + 4 preparing = ready after 6 minutes
Fryer, fish and chips: 3 + 8 = ready after 11 minutes
Pan, pasta with prawns: 3 + 10 = ready after 13 minutes
Grill, steak: 6 + 12 = ready after 18 minutes
The table gets its food after 18 minutes. An average dish is ready after 12 minutes ((6 + 11 + 13 + 18) ÷ 4), but the guest experiences 18. The salad waits 12 minutes at the pass, the fish and chips 7 and the pasta 5.
What do you gain by speeding up? If a cook starts on the steak after 1 minute instead of after 6, it is ready after 1 + 12 = 13 minutes and the table gets its food after 13 minutes: 5 minutes faster. If the steak is ready even earlier, nothing changes: the pasta (13 minutes) then decides when the table gets its food, and only the slowest dish counts. Blockages also arise like this:
A dish with two stations, such as a steak from the grill with a salad from the cold kitchen, is only ready when the slowest part is ready.
Batches: a station that works in batches can make the first ticket wait until there are enough tickets.
Starting just too early or too late: a dish that starts too early sits waiting at the pass; a dish that starts too late holds up the whole table.
When a faster kitchen can mean more revenue
A faster kitchen can free up capacity. Revenue only arises when someone uses that free capacity: a guest who would otherwise have walked away or a waiting party that now gets a table sooner. Kimes (2004) calls the hours in which guests wait for a table the busy hours; in her case study of one American restaurant (Chevys Arrowhead, data from 2001) those were ten hours per week, accounting for about a third of annual revenue. Only in such hours does the time you save count for your revenue.
How to work it out. Worked example with made-up numbers, amounts excluding VAT:
Seat minutes: 38 seats × 300 minutes (the dining room is in use from 17:00 to 22:00) = 11,400
Now: dining duration 100 minutes (from being seated until the table is ready again): 11,400 ÷ 100 = 114 guests, on average three times per seat
Five minutes shorter: 11,400 ÷ 95 = 120 guests, on average 3.2 times per seat. That is only possible if tables become available spread across the evening (see Thompson, 2009). With fixed sittings, only three fit into 300 minutes even at 95 minutes: 114 guests, extra revenue €0.
Extra capacity without fixed sittings: 120 − 114 = 6 guests (5.3% more)
Extra revenue: the smaller of the extra capacity and the number of waiting guests, times the revenue per guest. At €45 per guest and at least 6 waiting guests: 6 × €45 = €270 per Friday evening. If 2 guests are waiting: 2 × €45 = €90. If nobody is waiting: €0.
You do not necessarily gain those five minutes in the kitchen. At Chevys, the time from order to main course averaged 11 minutes and 31 seconds of a meal of 53 minutes and 15 seconds, more than a fifth (times measured for a hundred parties on busy weekend evenings).
Also treat the outcome as a ceiling. Thompson (2009) tested this calculation in a simulation of 1,296 scenarios with only guests without a reservation. With a 5% shorter dining duration, the calculation predicts 5.3% more revenue; in the simulation it was 1.1% on average, and 2.1% in the twelve most favourable scenarios. The revenue increase was larger with more demand and a longer peak. A simulation is not a measurement in real venues.
Faster is not always better for the guest either. In an American study of 151 guests of table-service restaurants (convenience sample, recalled experiences; Noone et al., 2009), satisfaction with the meal was highest at a perceived pace between too slow and too fast; it declined on both sides. A faster kitchen can improve the guest experience, as long as guests do not feel rushed.
More revenue does not by itself mean more profit: extra guests also bring purchasing costs and labour hours. You can read how to calculate that in part 2; the table turnover rate as a regular figure is covered in part 4.
How to measure ticket times
You only know how long a ticket stays in your kitchen once you measure it. At Chevys, a student recorded the times of a hundred parties; you start with twenty tickets.
Choose your peak. Measure twenty tickets in a row in the peak hour of a busy evening, and optionally a quiet hour for comparison.
Note for each ticket the time it came in and the time it went out (the last dish of the table leaves the pass; with courses you measure per course, from the moment it is called). If you want to know where the time goes, also note per station when it started and when it was done.
Calculate per ticket: ticket time = time out minus time in. Per station: waiting time = start minus time in; preparation time = done minus start.
Average: all ticket times added together, divided by the number of tickets.
90th percentile: sort the ticket times from short to long. With twenty tickets, that is the 18th (0.9 × 20 = 18): the longest time that remains if you leave out the two slowest tickets.
Worked example with made-up numbers. Twenty ticket times in minutes, sorted: 9, 10, 10, 11, 11, 12, 12, 12, 13, 13, 14, 14, 14, 15, 16, 16, 18, 25, 31, 38. Average: 314 ÷ 20 = 15.7 minutes. 90th percentile: the 18th value, 25 minutes (a spreadsheet function that interpolates gives 25.6). The average sounds calm, but two of the twenty tables waited more than half an hour (31 and 38 minutes). Suppose the grill was the last station to send out its dish on four of the five longest tickets: then that is where you start.
Twenty tickets give a first impression, not certainty; repeat the measurement on another evening. We did not find a fixed standard for the time between ticket and plate: it differs per concept. So choose your own target, for example a maximum for the 90th percentile, and measure whether you reach it.
Common mistakes
Looking only at the average. It hides the tables that wait a long time.
Measuring a quiet hour. Queues arise mainly at the peak.
Counting dishes that are waiting on purpose. A main course that has not yet been called is not a delay.
Assuming that faster always means more revenue. Without waiting guests, it yields no extra revenue.
Do this today
Tonight, measure twenty tickets with paper and a clock. For each ticket, note the number, the time the ticket came in and the time the last dish of the table left the pass, starting in the peak hour. After service, calculate the average (all ticket times added together, divided by twenty) and write down the longest five in a list. By our estimate, this takes a few seconds per ticket during service and then about ten minutes of calculation; you need no purchase and no system.
Where technology helps
Part of what you do above with paper and a clock can also be done on a screen. The Guestit kitchen displays (Kitchen OS) are meant to make those times visible. Each dish goes to a station (grill, fryer, cold kitchen, bar or dessert); a dish for two stations shows both tasks. The pass shows per course what is ready, what it is still waiting for and how long it has been waiting. If you choose two steps (first “Start all”, then “Ready”), the preparation time is measured exactly. The screen shows, among other things, the average ticket time against your target, the late orders and the slowest station, and you set thresholds per station for Warning and Late. As above, look not only at the average but also at the late tickets. The screen shows the times; what you do with them determines whether anything changes.
How we put this together
What we read: Little (2011) via a copy from the University of Massachusetts; Kimes (2004), Thompson (2009) and the author manuscript of Noone et al. (2009; we could not read the tables and figures) via the Cornell repository eCommons; the CBS and KHN news items (from KHN only the public summary; the full analysis requires a KHN Kompas account). Of Little (1961) we read only the abstract via Crossref. In Crossref, queried per DOI on 18 November 2026, none of the journal articles has a correction or retraction listed. The restaurant studies are American or simulated and do not necessarily apply to the Netherlands. Could not be verified and therefore left out: research on waiting time and satisfaction in Dutch restaurants, a standard for the time between ticket and plate, and Goldratt's theory of constraints (we did not open a source).
The image for this article is an illustration (AI-generated), not a photo of a real venue and not a screenshot of Guestit.
Sources
Statistics Netherlands (CBS), Omzet horeca groeit in tweede kwartaal 2026 met ruim 2 procent (1 September 2026)
KHN, Horeca blijft overeind, maar vraag en rendement staan onder druk (5 August 2026, updated 27 August 2026; public summary of KHN's own research, KHN Kompas; sample and comparison period not stated)
Little, A Proof for the Queuing Formula: L = λW, Operations Research, INFORMS (1961; only the abstract read, via Crossref)
Little, OR Forum: Little's Law as Viewed on Its 50th Anniversary, Operations Research, INFORMS (2011; text read, via a copy from the University of Massachusetts)
Kimes, Restaurant Revenue Management: Implementation at Chevys Arrowhead, Cornell Hotel and Restaurant Administration Quarterly, SAGE (2004; one restaurant in the United States; text read, via Cornell eCommons)
Thompson, (Mythical) Revenue Benefits of Reducing Dining Duration in Restaurants, Cornell Hospitality Quarterly, SAGE (2009; simulation study; text read, via Cornell eCommons)
Noone, Kimes, Mattila and Wirtz, Perceived service encounter pace and customer satisfaction: An empirical study of restaurant experiences, Journal of Service Management, Emerald (2009; convenience sample in the United States; author manuscript read, without the tables and figures, via Cornell eCommons)
About Guestit
Guestit is one system for reservations, ordering, preparation and payment.
Want to see this work in your business?
Book an introduction and we’ll look together at which of these ideas fit your business. If Guestit isn’t a fit, we’ll say so.

The operating system for hospitality. Connects every step of the guest journey with the processes behind it.
© 2026 Guestit
Privacy · Terms


