# A restaurant forecast is not a spreadsheet
A restaurant forecast is not a spreadsheet. It is a weekly decision system that tells you what to buy, who to schedule, and what to push, before the week happens.
TL;DR: Most restaurant forecasts fail because they are treated like a finance artifact instead of an operating cadence. Your goal is not a perfect prediction. Your goal is fewer surprises, tighter labor, less waste, and faster decisions. Build a forecast your team can run every Monday, and you will feel the difference inside two cycles.
What a forecast is supposed to do
A forecast exists to answer three operator questions:
- How busy are we going to be, by day and by shift?
- What does that mean for labor, prep, and purchasing?
- What do we need to change this week to protect the margin?
If your forecast does not create a decision, it is not a forecast. It is a report.
I have watched this play out in every kind of room. Independent cocktail bars, high-volume casual, hotel outlets, you name it. Someone builds a spreadsheet, maybe it is even a good one, and then it just sits there while the week happens to them.
The reason is simple.
Spreadsheets do not run restaurants. People do.
The real enemy is surprise
Restaurants can survive a slow week. They can survive a busy week. They can even survive a weird week.
What hurts is a surprise week.
Surprise shows up as:
- Labor gets scheduled wrong, and now you are either drowning or overstaffed.
- Ordering is off, and now you are 86ing things, or throwing them away.
- The wrong menu items get pushed, and average check slides.
- Your best people get burned out because you keep asking them to bail the boat out.
Forecasting is not about being right. It is about reducing surprise.
That is why the highest value forecasting systems I see are boring.
They are not a dashboard someone checks when they feel anxious. They are a weekly rhythm.
Why the spreadsheet version fails
The spreadsheet forecast usually fails for one of these reasons:
1) It is built for accuracy, not action
A forecast that is 92% accurate but never changes a decision is worse than a forecast that is 75% accurate and changes three decisions every week.
Accuracy is a vanity metric if it does not drive behavior.
2) It lives in one person’s head
The spreadsheet is owned by one manager who is good at Excel.
When they are out, the forecast is out.
Or worse, they are the only one who understands why the numbers moved, so the team cannot learn from it.
3) It is too slow
If it takes four hours to update, it will not get updated.
A forecast has to be cheap to run. The system has to do the heavy lifting so the operator can make calls.
4) It ignores the actual drivers
Most forecast sheets are built around last year vs this year and a generic growth assumption.
That misses what actually moves demand:
- Weather
- Local events
- Reservation pacing
- Private dining inquiries
- Social spikes
- Tourism cycles
- Construction and street closures
- Menu changes, price changes, and hours changes
If the forecast does not ingest signals from the real world, it will always feel wrong.
A forecast is a cadence: Monday build, midweek check, Sunday closeout
If you want a forecast that works, treat it like a weekly operating meeting.
Here is a structure I like because it matches how restaurants actually move.
Monday: Build the week
On Monday morning, you assemble the inputs and publish a simple view of the week.
Your output should fit on one page.
- Covers by day
- Sales by day
- Labor target by day (hours and dollars)
- One or two notes about what is driving the swing
Then you make decisions.
- Adjust schedule.
- Adjust prep and pars.
- Decide what you are pushing.
- Decide what you are protecting.
If the forecast does not end in a schedule change or an ordering change, you did not finish.
Midweek: Check the pace
Wednesday or Thursday, you do a short check.
- Are reservations pacing above or below?
- Are walk-ins tracking?
- Did weather shift?
- Did something happen in the neighborhood?
This is where you correct course.
Maybe you cut a bartender on Thursday. Maybe you call in a prep cook on Friday. Maybe you push a high-margin item because you can see the weekend building.
Sunday: Close the loop
On Sunday, you do the postmortem while it is still fresh.
- What did we predict?
- What happened?
- What signal did we miss?
- What should we watch next time?
The point is not blame. It is calibration.
If you close the loop every week, your forecast gets smarter without anyone “working on the forecast.”
What signals to track (and what to ignore)
This is the part most operators want to skip because it sounds like work.
It is work, but it is not that much work.
You are not building a research project. You are building a decision engine.
Here are signals that tend to matter.
Reservation pacing
Not just total covers, pacing.
- How far out are you filling?
- Are you filling the early seat or the late seat?
- Are you getting same-day bookings or not?
Pacing tells you if you are heading into a slam, a slow roll, or a dead night that needs help.
Private dining and large party inquiries
These are early warning signals for staffing and ordering.
Even if the event does not land, the inquiry volume tells you what people are trying to do.
Weather, but in a way a restaurant can use
The goal is not the meteorology forecast.
The goal is a simple rule set.
- If rain hits between 4pm and 9pm, walk-in bars get weird.
- If it is 85 and sunny, patios and frozen drinks spike.
- If it is 35 and windy, delivery and comfort food win.
Write down what is true in your room.
Menu changes and price moves
If you changed the menu, you changed the demand curve.
If you raised prices, you changed the mix.
Forecasting off last year with no adjustment is how you get blindsided.
Neighborhood friction
Street closures, scaffolding, subway work, sports games, concerts, graduation weekends.
Your restaurant does not live in a spreadsheet. It lives in a city.
The demand reality operators are dealing with right now
Two things can be true at the same time.
- People still want to eat out.
- They are getting pickier about how they spend.
The National Restaurant Association reported that eating and drinking places registered total sales of $101.0B in April 2026, up from $100.3B in March, marking the third straight monthly increase. (https://restaurant.org/research-and-media/research/restaurant-economic-insights/economic-indicators/total-restaurant-industry-sales/)
At the same time, YouGov reports that 37% of Americans say they are eating out less often than a year ago, and 54% say they have adjusted how they dine out to spend less. (https://yougov.com/en-us/articles/53259-rising-costs-are-changing-how-americans-dine-out)
If you are an operator, that combination shows up as volatility.
- The room can still be full, but average check is softer.
- Guests still come, but they trade down.
- Drinks get skipped.
- Add-ons get skipped.
- Midweek gets weird.
That is not a reason to panic. It is a reason to get tighter.
Forecasting is the “tighter” system.
What I would do this week if I were running your place
If you are reading this and thinking, “cool, but I do not have time,” I get it.
Here is the smallest version that still works.
- Pick one weekly moment. Monday morning is clean.
- Make a single page forecast, covers and sales by day.
- Add one line: “what is driving this week.” Weather, events, a promotion, whatever.
- Make one schedule change based on it.
- On Sunday, write down one thing you learned.
That is it.
You are building the muscle, not the model.
Once that rhythm exists, you can add better inputs, better tooling, and eventually automation.
But the cadence comes first.
Why I care about this
I spent 10 years behind the bar and another 10 years building systems for restaurants and brands.
The operators who win are not the ones who can predict the future.
They are the ones who can adjust faster than the week can hit them.
A forecast is how you adjust before you have to.
Jason
