
Nobody can tell you how much time your manager loses answering the same questions: you have to measure it in your own place. The method takes three steps. For two weeks, log every interruption (time, topic, duration, who else could have answered). Then sort each line into three cases: real decision, information already written down, information never written down. Finally, multiply the last two cases by a loaded hourly cost that you replace with your own.
Saturday, 8:15 pm, at the end of the pass
Lamia runs the floor of a sixty-cover restaurant. First seating is under way and tickets are coming in. At the end of the pass, three people are waiting for her at the same time. An extra wants to know where the cloths for the round table are kept. A server wants to know whether today's dessert contains tree nuts. A commis asks whether he can swap a side because they are short on leeks.
Michel, the chef, sends two plates. They sit under the heat lamp while Lamia answers all three. The first answer has been in the prep binder for a year. The second is on the dessert's recipe card, which the server could not find. The third is a real decision: only Michel or Lamia could make it.
The next day, Lamia takes a notebook out of her pocket. She has decided to count.
(Fictional scene and characters.)
Why this time never shows up
Each interruption is short
A question takes a minute, sometimes less. But one minute repeated thirty times a shift, five days a week, adds up to hours that nobody scheduled.
It blends into normal work
Answering questions is part of a manager's job. So the repeated question looks like useful work, when it is often making up for information filed in the wrong place.
The cost lands at the worst moment
Questions come when the team is under pressure, so during service. That is also when the manager is most needed elsewhere: at the pass, on the floor, with a guest.
You count questions, not causes
Saying "I always get the same questions" gives you nothing to act on. You need to know which ones, and why they reach you rather than a procedure card or a colleague. We covered the other side, the instruction you keep repeating, in how many times should you repeat an instruction before changing the system. This article is about measuring the time and putting a number on it.
Keeping a two-week interruption log
Two weeks is enough to even out a quiet shift and a slammed one. Tell the team before you start: you are logging topics to improve the procedures, not people.
- Pick the format. A pocket notebook or a note on your phone. You need to be able to write a line in ten seconds, standing up.
- Write one line per interruption. Four columns: the time, the topic ("dessert allergens", "where the cloths go", "side swap"), the estimated duration in minutes, and who else could have answered (nobody else, a colleague, a written card).
- Never write the name. Not the person's name, not their initials. In a small team, leave out even the role if that is enough to identify someone.
- Count messages too. A call or a text on your day off about an operational question is an interruption. Log it separately.
- Fix nothing during the two weeks. If you rewrite a card on day three, your starting figure will be wrong. Note the idea and fix it afterwards.
Sorting every line into three cases
At the end of the two weeks, go back through each line and classify it: each case calls for a different fix.
| Case | How to recognise it | Example | What to do with it |
|---|---|---|---|
| Real decision | Nobody else could have decided | Swapping a side, taking a table of ten at 10 pm | Nothing to remove: this is the manager's job. You can only group these better |
| Information already written down | The answer exists in a card, a binder, a message | The dessert's allergens, the closing order | Make the card findable at the station, at the moment of the task |
| Information never written down | The answer only exists in someone's head | Where the round-table cloths go, the code for the bin store | Write it once, in a few lines, and assign it to a role |
The "who else could have answered" column is your cross-check. A line sorted as a real decision where "a colleague" could have answered deserves a second look: either the colleague does not dare, or the rule is unclear.
The calculation, with assumptions to replace
The figures below come from no study. They are assumptions to show how the calculation works: replace each one with your own.
Take a manager with a loaded hourly cost of €25: replace it with yours, which your accountant or payroll can give you. Suppose Lamia's log shows 60 interruptions in a week, averaging two minutes each, and that 40 of them fall into the last two cases (information already written down or never written down).
Direct time per week = avoidable interruptions × average duration 40 × 2 min = 80 min
Recovery time = avoidable interruptions × minutes to pick the thread back up (assumption: 1 min) 40 × 1 min = 40 min
Total avoidable time per week = 80 + 40 = 120 min, or 2 hours
Yearly value = 2 h × weeks open (assumption: 45) × loaded hourly cost (assumption: €25) 2 × 45 × 25 = €2,250
Three precautions before you show that figure to anyone. The one minute of recovery time is a cautious assumption: set it to zero if you would rather count only direct time. The amount is not a saving: it only becomes one if those two hours go to something else (training, planning the week, being on the floor). And do not count the waiting time of the person who asked as certainly lost: note it separately, as a possible cost.
To go further in costing a tool, the full method is in how to measure restaurant software ROI in 90 days.
What interruptions cost besides time
Focus
In 2008, Gloria Mark, Daniela Gudith and Ulrich Klocke (University of California, Irvine) presented a study at the CHI conference called "The Cost of Interrupted Work: More Speed and Stress". In a lab setting, on an email task, people who were interrupted finished just as fast or faster, but reported more stress, frustration, time pressure and effort. The experiment was not run in a kitchen, but the mechanism is familiar: you catch up by speeding up, and you pay for it in fatigue.
Service
While the manager answers, something is waiting: a plate under the lamp, a guest who wanted the check, a delivery to check in. Mark the interruptions that land in the middle of the rush with a separate sign in your log. That is often where you will find the first three questions to deal with.
The team
A team that has to go through the manager for every piece of information learns to wait rather than look. On the day the manager is off, it looks less well. That is the subject of independent, not dependent: measuring that the team is learning.
Common mistakes
- Counting per person. The log turns into a file on people, the team stops asking, and the number drops without anything being fixed. On this point, read can your manager read your questions to an assistant.
- Filing everything under "already written down". If the card exists but nobody can find it at the station, it is not really available. The problem is where it lives.
- Trying to remove real decisions. A garnish swap or a late table of ten stays with the manager: you can batch these decisions, not make them disappear.
- Presenting the amount as a saving before the time has been put to other use (see the caveats under the calculation).
What to measure
- The number of avoidable interruptions per week (the last two cases in the sort), before and after your fixes.
- The share of those interruptions that land in the middle of service.
- The three topics that come up most. Those are your first three cards to write or move.
Repeat the same week of logging a month after your fixes. If avoidable interruptions do not go down, the card is in the wrong place, or it answers the question badly.
Where Tsuno San fits
Tsuno San, the smart voice assistant connected to the Tsuno environment, can take a question like "where do the round-table cloths go?" instead of the manager, provided the answer has been written down. It only sees what the person can already see. Its current status (voice available as a demo, hands-free still being explored) is set out in Tsuno San, the work companion for frontline teams.
San does not make the real decisions. It does not replace the manager and does not evaluate anyone. Nor does it create the information that is missing: the third case in the sort is still a writing job the team has to do once. What a good tool should do, and what we are working on: show which topics come up most, without ever saying who asked. We do not present that as done. In the meantime, Lamia's notebook works with any tool.
The features that keep information in order (checklists, site documents, skills) are listed on the features page. The Tsuno San page presents the assistant, and the online demo lets you ask questions of a fictional group of three restaurants.
Further reading
Once you have your number, test a tool on three simple measures with measuring a pilot in 30 days: three numbers, not thirty. To understand why new hires keep coming back to you, read why new restaurant employees often say "nobody told me". And for a manager alone with the questions late at night, see AI that reassures a manager alone at closing time. The pillar page AI for frontline managers: analyse, advise, reassure widens the subject, and the Tsuno San FAQ answers the questions managers ask most.
Key takeaways
You count the time lost answering the same questions in your own restaurant, with two weeks of logging and a sort into three cases, then a calculation built on your own assumptions. Real decisions stay with the manager; the rest is fixed by writing down what is missing and putting the cards where the work happens.
Frequently asked questions
How much time does a restaurant manager spend answering the same questions?
There is no reliable figure that holds for every restaurant. The only useful number is yours: for two weeks, note every interruption, how long it took and what it was about, then add it up. Most managers are surprised by the total, one way or the other.
How do you keep an interruption log without it eating more time?
One line per interruption, in ten seconds: the time, the topic of the question, an estimated duration and who else could have answered. A pocket notebook or a note on your phone is enough. You group and calculate only at the end of the two weeks.
Should you write down who asked the question?
No. Write the topic, never the name. The goal is to find the information that is missing or badly placed, not to know who asks. A team that knows it is being counted stops asking and starts guessing.
How do you turn that time into money?
Multiply the hours lost per week by the manager's loaded hourly cost, then by the number of weeks you are open. Use your own hourly cost and your own weeks, not an industry average. Remember that the time is only saved if it is spent on something else.
Is every repeated question wasted time?
No. A real decision (a gesture for a guest, a substitution, a difficult table) belongs to the manager and should stay there. What costs you is information that is already written down but that nobody can find, and information that was never written down at all.
What do interruptions cost besides time?
Focus, and sometimes the quality of service. A University of California, Irvine study (Mark, Gudith and Klocke, 2008) found that people who are interrupted make up for it by working faster, but report more stress, frustration and time pressure.