Improve restaurant operations in ninety days

To improve restaurant operations, most venues try to fix everything at once. A new manager arrives, or the owner reads a book, and the list of changes runs to twenty items: a new rota, new checklists, a new menu, new suppliers, new training, new systems. Six months later, a few have stuck, most have faded, and nobody can say which made a difference. The team is tired of change, and the venue runs much as it did before.

There is a better way. Choose three changes a quarter, no more. Choose them by the size of the problem they solve, not by how easy they look. Define what done looks like and how you will measure it before you start. Then spend ninety days making each change stick, measure the result, and choose the next three. Over a year, twelve changes that worked will transform a venue more than fifty that were started and abandoned.

This guide sets out how to find and size the problems, how to choose three, how to plan the ninety days, how to make the changes stick with standard operating procedures and training, and how to tell whether they worked. It ends with examples of changes worth making, a worked example and the common mistakes.

Ninety days is long enough to make a real change and see its effect, and short enough to keep attention. A change that takes a week to introduce needs several more weeks to become habit, and several more for its effect to show in the numbers. Thirteen weeks gives time for all three.

It also fits the business calendar. A quarter is a natural period for reviewing the numbers, and it allows four rounds of change a year. Shorter cycles tend to produce changes that are announced but never embedded; longer ones lose momentum.

Every change needs attention: someone to lead it, time to train the team, and a manager to check that it is happening. A venue has limited attention, most of it spent running service. Three changes is roughly what a management team can lead well alongside the daily work.

Three also makes the result readable. If ten things change at once and food cost falls, nobody knows which change caused it. With three, each chosen for a different problem and measured separately, the effect of each can be seen.

The natural instinct is to start with the easy changes: a new checklist, a tidier store room, a fresh coat of paint. They feel productive and can be finished quickly. But an easy change to a small problem produces a small result. Meanwhile, the large problem, such as a kitchen that cannot keep up on Saturdays or a labour cost far above target, goes untouched.

Choose by size. Each change has an opportunity cost: the attention spent on it could have gone to something else. Spend that attention where the problem is largest, even if the fix is harder.

Start by listing every problem the venue has, without judging them yet. Ask the managers, the chef, the team and the guests. Look at the numbers. Walk through a service and watch. Write everything down: slow tickets, high waste, staff leaving, complaints about noise, a booking page that loses guests, late deliveries, a cash-up that never balances.

A long list is normal. The point of this step is to see everything before choosing, so the choice is deliberate rather than driven by whatever happened last weekend.

The numbers show many problems directly. Look at prime cost, the cost of food, drink and labour as a share of sales, against its target. Look at covers and spend per head by service against last year, the no-show rate, ticket times, waste and the share of guests who return.

Look for numbers that are far from their targets, that are moving the wrong way, or that vary widely from week to week. Each one points to a problem worth listing. Our guide to restaurant KPIs worth tracking every week sets out the numbers to start with.

Guests show problems the numbers miss: slow service at the bar, a dish that is often sent back, a room that is too loud, a booking page that is hard to use, staff who seem rushed. Read feedback and reviews for themes rather than single comments. A complaint that appears every week is a problem; one that appears once may not be.

Ask the floor team too. Servers hear what guests say at the table, including the things guests never write down.

The team shows problems that neither the numbers nor the guests reveal: a rota that changes too late, a section that is always understaffed, a piece of equipment that breaks every week, a process that wastes time, a manager who is never available. High staff turnover is often the clearest sign that something is wrong.

Ask the team directly, in a short conversation or a simple anonymous form: what makes your shift harder than it needs to be? The answers are often surprising and specific.

Look at the venue's week as a system. Which recurring jobs are done late, done badly or missed? Is the rota published on time? Are orders placed on the right days? Do the daily checklists get signed? Is the cash-up done every night? Each failure in the weekly routine causes problems later in the week.

Our guide to restaurant operations: the week as a system sets out the eight recurring jobs to check.

For each problem on the list, estimate its size: how much it costs the venue each month, in money, guests or team. Some are easy to size: waste can be weighed and costed, labour above target can be calculated. Others take a little judgement: slow tickets cost turns on busy nights and lose return visits; staff turnover costs recruitment, training and service quality.

Rough estimates are fine. The aim is to rank the problems, not to calculate them exactly. A problem worth a few thousand a month is larger than one worth a few hundred, whatever the precise figures.

Some examples of putting a value on a problem, with illustrative numbers. A labour cost two points above target on sales of 100,000 a month is 2,000 a month. Waste of 40 kilograms of food a week at an average cost of 6 per kilogram is about 1,000 a month. Losing one table turn on each of eight busy nights a month, at 300 a turn, is 2,400 a month.

For problems that are harder to value, such as a stressful rota or a noisy room, estimate their effect on something measurable: staff leaving, guests not returning, lower spend. Even a rough value makes them comparable with the others.

For most venues, prime cost is the first lens to apply. Food, drink and labour are the largest controllable costs, and problems that affect them usually affect profit most. A problem that moves prime cost by a point is often worth more than one that improves a guest satisfaction score by the same proportion.

That does not mean every change should be about cost. Guest and team problems matter too, and some of the largest problems are about guests who do not return. But looking at prime cost first stops the biggest cost problems from being overlooked. Our guide to prime cost, the one number to watch weekly covers how to calculate it.

With the problems sized, choose three. Usually, these are the three largest. Sometimes, one large problem needs another to be solved first: a kitchen rota cannot be fixed until the forecast is reliable, for example. In that case, choose the problem that unlocks the others.

Check that the three can be led at the same time. Three large changes in the kitchen, all led by the head chef, may be too much. Spreading them across the kitchen, the floor and the management team makes them easier to deliver.

A useful pattern is one change about cost, one about guests and one about the team. That keeps the quarter balanced: profit improves, guests notice something better, and the team feels its concerns have been heard. It also spreads the work across different people.

The pattern is a guide, not a rule. If the three largest problems are all about cost, start there. But a quarter that only cuts costs, while guests and the team see no improvement, can damage the venue in ways the numbers will show later.

Before starting, define what done looks like for each change, and how you will measure it. "Reduce waste" is not a change; "weigh and record kitchen waste every day, and cut it by a quarter by the end of the quarter" is. "Improve the rota" is not a change; "publish the rota by Wednesday for the following week, every week" is.

Choose one measure for each change, record its starting value now, and decide when you will check it. Without a baseline, you cannot tell whether the change worked.

Plan each change across the ninety days, with an owner, a start date and the steps in between. Most changes follow the same three phases: understand and start, change and steady, hold and measure. Each phase is roughly a month.

Write the plan on one page per change, and review progress at the weekly operations meeting.

In the first month, understand the problem properly and start the change. Watch how the work is done now, talk to the people who do it, and find the real cause. A slow kitchen may be caused by the layout, the menu, the ticket flow or the rota; the fix depends on which.

Design the change with the people who will carry it out, explain why it matters, and start it, ideally on a small scale first: one service, one section, one week.

In the second month, roll the change out fully and make it steady. Train everyone involved, write the new way of working down, and check every day that it is happening. Expect resistance and slips: old habits are strong, and a busy Saturday will test any new process.

Fix problems as they appear. If the change does not work as designed, adjust it rather than abandoning it. Most changes need two or three adjustments before they settle.

In the third month, hold the change and measure the result. The change should now be part of normal work, with no special attention needed. Check the measure against the baseline and the target. Look for side effects: did the change cause problems elsewhere?

At the end of the ninety days, decide: keep the change, adjust it, or reverse it. Then choose the next three.

The changes need a regular place to be discussed. An operations meeting once a week, short and focused, reviews the progress of each change: what was done, what is blocking it, and what happens next. It can be part of the weekly review or a separate meeting with the people leading the changes.

Keep it short: five minutes per change is usually enough. The purpose is to keep attention on the changes, not to redesign them every week. Our guide to a weekly review that takes thirty minutes covers the wider meeting.

A change that lives only in people's heads fades when they leave or forget. Write each change into the venue's standard operating procedures: the step-by-step description of how the job is now done. A new waste routine, a new cash-up process, a new way of pacing tables: each needs a short, clear procedure.

Keep procedures short, with steps in order and photos where they help. Put them where the work happens: the kitchen wall, the bar, the office. Use them in training. A procedure that is written but never used will not hold the change.

Every change needs training: explaining why, showing how, and checking that people can do it. Train in short sessions, close to the work, ideally during quiet periods. Show, then let people try, then check.

Train everyone who is affected, including people who work only occasional shifts. A change that the weekend team never heard about will fail on the busiest nights.

At the end of the quarter, compare the measure with its baseline and target. Did waste fall by a quarter? Is the rota published by Wednesday every week? Did ticket times on Saturdays drop below the target? If yes, the change worked. If the measure moved partly, the change may need more time or an adjustment.

Be honest. A change that did not work is useful information, as long as it is recognised and learned from. Pretending it worked wastes the next quarter.

Compare carefully. A measure that changes over a quarter may change because of the season, not the change. Waste falls in quiet months; labour cost rises in December. Compare with the same period last year where possible, or with the weeks just before the change, adjusted for any obvious seasonal effect.

Where you can, compare like with like: Saturday ticket times before and after, not the average for the whole week.

If a change does not work, find out why before deciding what to do. The cause may have been misunderstood; the change may not have been carried out as designed; the training may not have reached everyone; or the change may simply have been the wrong answer.

Decide whether to adjust it for another quarter or replace it with something else. Do not keep a change that is not working just because effort has gone into it.

Every change can have side effects. Cutting labour can slow service. Cutting waste by preparing less can cause dishes to run out. Raising table turns can make guests feel rushed. Watch for them, by looking at the other numbers and listening to guests and the team.

A change that improves one number by damaging another has not improved operations; it has moved the problem.

At the end of each quarter, update the list of problems, size them again, and choose the next three. Some problems will have shrunk because of the last quarter's changes; new ones will have appeared. The list never ends, and that is normal.

Over a year, the venue makes twelve deliberate changes, each measured. That is how good venues improve steadily without exhausting their teams.

Changes that often solve large problems include:

Each is specific, measurable and achievable in a quarter.

Some changes are tempting but rarely solve a large problem: a full menu redesign when the menu is not the issue, new uniforms, a new logo, rearranging the office, a long policy document. They may be worth doing at some point, but they should not take one of the three places unless they address a large problem.

Software changes can fall into this category too. A new system can help, but only if it solves a problem on the list. Our guide to choosing restaurant software without a demo haze covers how to tell.

In a small venue, the owner may lead all three changes, and the operations meeting may be a coffee with the chef. That is fine. Three changes may even be too many; one or two may be the right number. The principles still apply: choose by size, define done, measure, and give each change time.

A small venue often gains more from each change, because a single problem can be a large share of its costs or guests.

In a group, each venue can choose its own three changes, based on its own problems. The group adds value by sharing what worked: a waste routine that cut food cost at one venue can be tried at another. Some changes, such as a new system or a shared supplier, may be chosen at group level and rolled out across venues.

Keep the venue changes local. A group that imposes the same three changes on every venue, regardless of their problems, will waste effort where those problems do not exist.

Here is an illustrative venue's quarter. The figures are examples.

Each change was written into the procedures and trained. The next quarter's list started with waste and staff turnover.

In servearound, analytics helps with both ends of the ninety days: finding the problems and measuring the changes. Every page reads for a quick range or one you pick, compared with the period before, and flags small samples. Kitchen analytics shows median and P90 prep time, on-time rate, remake rate, where the minutes go, the items that cost most time and the slowest tickets. Reservation analytics shows no-show and cancellation rates by meal period, source and party size, and how late cancellations arrive.

Table analytics shows utilisation and the gap between seatings, payment analytics shows voids and comps by member, and the menu report shows margin once dish costs are set. Labour and stock costs live outside servearound. See Analytics, or get in touch to talk through your quarter.

Three a quarter is a good number for most venues; one or two for a small one.

By size: how much each costs in money, guests or the team. Start with the largest.

Because easy changes to small problems produce small results, while the large problems go untouched.

Record a baseline before starting, measure the same thing at the end, and compare fairly, allowing for the season.

A short weekly meeting that reviews the progress of each change and what happens next.

So they survive when people leave, forget or are away, and so new starters learn the new way.

Find out why, then adjust it or replace it. Do not keep it just because effort has gone into it.

Yes. Each venue chooses its own changes, and the group shares what worked.

To improve restaurant operations, make three changes a quarter, chosen by the size of the problem rather than ease. List every problem from the numbers, the guests, the team and the weekly routine; size each one, with prime cost as the first lens; and choose three, ideally balancing cost, guests and the team.

Define done and a measure for each, plan the ninety days in three phases, and review progress at a short weekly operations meeting. Lock each change into standard operating procedures and training, measure against a fair baseline, watch for side effects, and then choose the next three. Twelve changes that worked will do more than fifty that were started.