The Rescue Job

Stopping the Bleeding in Week One Without Burning Trust

There is one week where you can act fast without breaking the diagnostic phase. Here is what belongs in it, and what does not.

Kiyong (Eric) Lee·Operations Executive·August 2026·9 min read

The short version

The no-change rule in a turnaround applies to the operating system, not to items that are literally bleeding cash and can be stopped without disturbing anything.
Week one fixes are narrow: standing autopays, uncontrolled overtime, petty cash, and unbilled comps. Nothing else.
Before you touch anything, publish a list of what is not changing in the first 30 days. That message buys you the room to move on the short list.
Every week one fix goes through the general manager, not around them. The manager owns the message. You own the substance.
What must wait: the schedule, the menu, the vendor renegotiation, the firing. Those are execution phase, not week one.

Every turnaround has a version of the same conflict in week one. The board wants to see fast action. The operator's textbook says not to change anything for 30 days. Both instincts are right, and they are not actually in conflict once you look at what belongs in each bucket.

There is a small, well-defined set of things that must be fixed in week one because they are literally hemorrhaging cash into the wall every day you wait. And there is a much larger set of things that feel urgent but must not be touched until the diagnostic is done. Confusing the two is the single most common mistake I see in first-time turnaround operators, and it is the mistake I made the first time too.

This is the short list of what belongs in week one, why it is safe to touch, and what does not belong even though it will feel like it does.

The rule underneath the rule

The reason the standard turnaround advice says "do not change anything for 30 days" is that changes contaminate the diagnostic signal. If you rebuild the schedule in week two and labor cost drops in week four, you cannot tell whether the drop is your new schedule or the underlying demand curve moving. You have lost the ability to read the picture.

But this rule has a hidden condition. It applies to changes that affect how the restaurant runs. A canceled autopay on a discontinued vendor does not change how the restaurant runs. A capped petty cash card does not change how the restaurant runs. Requiring a comp code at the point of sale does not change how the restaurant runs. These fixes stop cash bleed without disturbing the operating system, so they do not contaminate anything.

The 30-day no-change rule protects the diagnostic signal, not your calendar. Any fix that does not touch the operating system does not touch the diagnostic. Move on it now.

The trick is knowing which is which. The list below is the one that survives the test in almost every restaurant turnaround I have run.

Fig. 1 · The week one action set lives inside the 30-day diagnostic, not against it.

The week one action list

1. Cancel the standing autopay leaks

In the first 48 hours, pull a complete list of every recurring credit card charge and every automated ACH pull from every operating account across every location. In an underperforming restaurant group of any real size, this list has surprises on it. Software subscriptions from a manager who left 18 months ago. A pest control contract on a location you closed. An SEO retainer nobody remembers signing. A commercial insurance rider on a piece of equipment you sold in 2023.

In one $30M group I ran the diagnostic on, the autopay audit surfaced $2,400 a month in charges servicing operations that no longer existed. That is $28,800 a year, recovered by making phone calls, and it did not touch a single thing anyone actually does inside the restaurants. This is the safest and easiest fix in the entire turnaround. It should be done in week one, every time.

The only caveat: do not cancel anything you are not sure about. Ambiguous charges get paused, not canceled, and then routed into the diagnostic. The rule is only cancel the ones that are obviously dead.

2. Freeze uncontrolled overtime

Overtime is the loudest bleed in most underperforming restaurants and the temptation is to attack it with a schedule rebuild. That is the wrong move in week one. The right move is a brake, not a redesign.

Put a rule in place, in writing, by end of day three: any shift that would push a line employee over 40 hours in the workweek requires manager approval in the scheduling system before the employee clocks in. Not the general manager. The shift manager. It is a small procedural change that flags the overtime hours before they happen instead of after.

Notice what this does not do. It does not cap overtime. It does not fire anyone. It does not change anyone's regular schedule. It just makes the overtime visible in the moment it would occur, and it forces a person to make an active decision instead of passively accumulating it. In practice, this single control usually stops 30 to 60 percent of the overtime bleed inside a week, because most uncontrolled overtime is not decisions. It is drift.

3. Cap the petty cash and force receipts

Every underperforming restaurant has a petty cash habit somewhere. Someone runs to the hardware store for a $60 part. A vendor delivery is short and the manager peels off $200 to make up the difference on the spot. Staff meals are being covered from the drawer. None of this is theft. It is convenience that has gotten expensive.

Week one fix: petty cash gets moved to a card with a hard $500 daily cap per location and a next-business-day photo receipt requirement in the accounting system. No receipt in 24 hours, no card use tomorrow. Boring, effective, and it does not touch how the restaurant runs. The staff meal habit and the vendor short habit both surface into the diagnostic instead of hiding in a drawer.

4. Kill the unbilled comp channel at the point of sale

Unbilled comps are the silent margin killer in most underperforming groups. Managers comp meals for regulars, for service recoveries, for staff, and often for reasons that never get written down. The comps go through the POS as adjustments without codes, or as voids, or as manager overrides that never surface on the daily P&L.

Week one fix: starting the next shift, require a written comp code with a manager PIN for every dollar comped. Do not change the comp policy yet. The policy stays exactly as it is. Only the tracking changes. This is critical, because if you change the policy in week one the general managers will read it as an accusation and the whole team goes defensive. If you only change the tracking, they experience it as a system requirement, not a judgment.

The signal alone reduces comp bleed by 20 to 40 percent inside a week, without any policy change, because people comp less when they know it will show up on a report someone will read. The remaining comps, now visible with codes and reasons, become material for the diagnostic.

What must not happen in week one

The reason the week one list is short is that the list of things that feel urgent is much longer, and most of what is on the "feels urgent" list must not happen yet. The temptation to move on these things is real. I have moved on them myself in past engagements and paid for it.

Do not rebuild the schedule

This is the biggest one. Labor variance is the loudest number, the schedule is the obvious lever, and the operator's instinct is to redesign it in week one. Do not. You do not yet know the demand curve well enough to build a better schedule. Rebuilding early will produce a schedule that is different, not better, and it will corrupt the diagnostic for two weeks while you cannot tell whether labor cost is moving because of your change or in spite of it. Schedules get rebuilt in the execution phase, after the demand curve is understood.

Do not change the menu

Cutting the two worst-performing items in week one looks obvious and feels responsive. It is neither. Menu items are load bearing in ways you cannot see yet. That $18 salad you were going to cut might be the item half your regulars order for lunch, and the guests who follow those regulars in do not know they exist until you check the ticket data properly. Menu decisions are for month two at the earliest.

Do not renegotiate vendors

You do not have the volume data to negotiate from yet. Any vendor concession you extract in week one comes at the price of the relationship, and it is a fraction of what you will get in month three when you can walk into the negotiation with 90 days of clean data and a clear ask.

Do not fire the general manager

The temptation to move fast on the general manager is the strongest and the most damaging. Firing in week one teaches every remaining general manager to hide problems from you, which is the exact opposite of what you need during a diagnostic phase. The manager stays, whatever the manager's future ends up being. That decision is for month three or four, not week one.

The list of things that must wait is longer than the list of things that must happen. Discipline in what you do not do is the whole game in week one.

The message that buys the room

The last thing that belongs in week one, and it is arguably the most important, is the message to the operating team. Before you cancel a single autopay or install a single comp code, send one written message by end of day two to every general manager and every kitchen lead. The message has two parts.

Part one: here is what will not change during the first 30 days. Menu, prices, staffing structure, vendor relationships, standing schedules, reporting lines. Name the specific things people are afraid of, and name them as safe.

Part two: here is the small set of items I am going to touch this week, and why. Standing autopay audit. Overtime approval requirement. Petty cash cap. Comp code requirement at POS. These are procedural. They do not change your job.

That message is what buys you the room to move fast on the short list. Without it, the team reads any change in week one as the beginning of everything changing, and morale collapses before you get to the actual work. With it, the team reads the changes as bounded and procedural, and they wait for the diagnostic like adults.

What the whole thing is really doing

Week one is not the turnaround. The turnaround is the 9 to 11 months that follow it. What week one is doing is three things, in this order.

First, it stops the fastest bleed while the real work is being designed. That is real money, usually $8K to $25K a week in ongoing recovery for a mid-sized multi-unit group, but it is not the point. It is the byproduct.

Second, it demonstrates to the operating team, the board, and the founder that the operator can execute. The team needs to see the operator do something concrete and land it cleanly in week one, or the diagnostic month will feel like paralysis.

Third, and most importantly, it establishes the pattern of the whole engagement: narrow, specific, high-discipline changes with named owners and clear reasons. If week one is careful, the executed team assumes months two through eleven will be careful too. If week one is chaotic, they brace for chaos, and they behave defensively for the rest of the engagement.

Move fast on four things. Leave the rest alone. Tell people what is safe before you touch what is not. That is week one.

Staring at a broken P&L right now?

If you are one week into a turnaround and trying to figure out what belongs in the next seven days, I am happy to trade notes. I take on a small number of restaurant and hospitality turnaround engagements each year, mostly in the $10M to $60M revenue range.

Get in touch

Questions people ask

What should you fix in the first week of a restaurant turnaround?

Fix only the items that are actively hemorrhaging cash and can be stopped without changing how the restaurant runs. That short list usually includes canceling standing vendor autopays on discontinued programs, freezing uncontrolled overtime with a manager-approval requirement, moving petty cash to a capped card, and forcing a written comp code on every discount at the point of sale. Everything else waits for the day 30 diagnostic.

Does the no-change rule in a turnaround apply to week one?

The no-change rule applies to the operating system: SOPs, menus, schedules, staffing structure. It does not apply to items that are actively bleeding cash and can be stopped without disturbing the operating system. A canceled autopay on a closed location does not change how the restaurant runs. A rebuilt schedule does. Know the difference and act on it fast.

How do you move fast in week one without burning trust?

Announce what is not changing before you announce what is. Sending general managers a list of the things that will remain stable through the diagnostic buys you the emotional room to act quickly on the small number of items that must stop bleeding immediately. People fear change they cannot bound. Bound it explicitly.

What is the biggest week one mistake in a turnaround?

The biggest mistake is rebuilding the schedule in week one. It always looks like the obvious fix because labor variance is the loudest number on the P&L. But changing the schedule before you understand the demand curve corrupts the diagnostic signal for at least two weeks, and the new schedule is almost never better than the old one when built on incomplete data. Wait for day 30.

How much cash can week one actually recover?

For a $30M multi-unit group with three underperforming locations, the week one fixes typically recover $8K to $25K per week in ongoing bleed once fully implemented. That is not the turnaround. That is buying breathing room while the real turnaround work gets designed. But it is meaningful, and it demonstrates to the team and the board that the operator can execute.

Should you communicate the week one changes to line staff?

Not directly, and not as changes. Every week one fix should be routed through the general manager or the kitchen lead. Line staff should experience the changes as small process adjustments coming from their direct manager, not as edicts from a new outside operator. The manager owns the message. The operator owns the substance.

What week one action feels urgent but should wait?

Firing a general manager. Changing a menu. Rebuilding the schedule. Renegotiating a vendor contract. Repricing anything. All of these feel urgent in week one because the bleed feels urgent. None of them belong in week one. They belong in the execution phase, after the diagnostic tells you the sequence.