When to raise menu prices and how to do it without losing covers

Most operators wait too long, then move too much at once. Costs creep for eighteen months, nobody adjusts, and one day the whole menu jumps. The regular does not object to paying more. They object to being surprised.

Guests track a handful of prices, not a menu. Which ones they track is the part worth knowing before you touch anything.

Two numbers set the frame. The National Restaurant Association estimates that average wholesale food prices are up 35% since 2019, average hourly earnings of restaurant employees are up 41% since February 2020, and total expenses for an average restaurant rose 36% between 2019 and 2026. Against that, the Association's July 2026 outlook projects total restaurant and foodservice sales growing 4.3% in nominal terms and only 0.8% after inflation, and states plainly that much of the growth is being driven by higher menu prices rather than by traffic.

Read those together and the position is uncomfortable but clear. The industry's revenue growth is mostly price, and your costs moved first. The question is not whether to price. It is which items and when.

When is it actually time to raise prices?

When one of four things is true, and none of them is a feeling.

Your cost on a specific item has moved and stayed moved for two or three periods. A one week spike on produce is weather. A sustained shift in a protein price is a new baseline.

Your labor has structurally changed. A wage increase, a scheduling change you cannot undo, a station you now have to staff differently. Work the schedule first, though, because four points of labor variance is a cheaper fix than asking guests for the same money.

The item is underpriced for what it delivers. Some dishes sit at the same number for years while the portion grows and the plate gets more expensive to send.

You are turning tables away. Full at the times you want to be full is the clearest signal there is, and the one people are most reluctant to act on.

Not reasons: a bad month, a competitor's menu, or a general sense that things cost more now. Those produce increases you cannot defend when a guest asks, and someone will ask.

Which items should move and which should not?

Move the ones guests do not price check. Leave the ones they do.

Every restaurant has three or four anchor items guests carry in their heads. The burger, the house roll, the lunch special, the wine everybody orders. A regular can quote those, and a move on them reads as a price increase across the whole restaurant.

Everything else has more room than operators assume. Sides, add ons, specialty items, anything with a modifier. A guest who orders your short rib twice a year does not know what it cost last time.

Practical ordering:

• Start with high margin items that sell moderately. Small moves there produce real dollars with almost no visibility.
• Fix the items losing money, even if the move is larger, because carrying a loser is worse than losing a few orders of it.
• Leave the anchors alone this round, and revisit them later, on their own.

There is a fourth option people forget. Change the item instead of the price. A different cut, a resized portion, a new garnish. A dish that changed has no old price to be compared against, which is why moves at a seasonal menu change draw so little reaction.

How do you make the increase land quietly?

Attach it to something the guest already expects to be different.

A new menu, a seasonal rollout, a printed redesign. All of those reset expectations by themselves. Changing prices on an unchanged menu is the version guests notice, because the only thing different is the number.

The rest is craft. Keep the moves small and uneven so the menu does not read as a systematic adjustment. Avoid moving every item in one section. Do not round everything to the same ending, which looks mechanical.

Tell your staff before the guests see it. Servers get asked directly and they need an answer that is not an apology. Give them one true sentence: the cost of that dish went up and we adjusted it rather than cutting the portion. A server who apologizes for a price teaches the guest to be annoyed by it.

Do not print an explanation on the menu. Guests who would not have noticed now read a paragraph about your costs, and a quiet change becomes an announcement.

How much volume can you afford to lose?

More than you think, and this single calculation settles most pricing arguments before they start.

Take one item. It sells 340 units a month at $19, and it costs you $6.40 to make. Contribution per unit is $19 minus $6.40, or $12.60. Monthly contribution from that item is 340 times $12.60, which is $4,284.

Now move it to $20. Contribution per unit becomes $13.60, because the dollar you added carries no additional food cost.

Here is the question that matters. How many units can you lose and still be even?

$4,284 divided by $13.60 is 315 units. You currently sell 340. You can lose 25 orders a month, roughly one order in fourteen, before the increase has cost you anything.

That is the whole argument for moving a price. A one dollar move on a $19 item does not need the guest to be indifferent. It needs the guest to be less than one in fourteen likely to walk away from the dish because of it.

If units hold, the same move is 340 times $1, or $340 a month and $4,080 a year, almost all of it contribution.

Now run it backwards, which is the version nobody does. Say you leave the price alone and your cost on that item drifts from $6.40 to $6.90 over a year, which is modest against a 35% rise in wholesale food prices since 2019. Contribution falls to $12.10, monthly contribution to $4,114, and you have quietly lost $170 a month, or $2,040 a year, on one dish. Nobody decided that. It just happened while the menu sat there.

Build that table for your top twenty sellers. Price, cost, contribution, units. It fits on one page and it is the entire decision.

Can you just test the price?

Not in any way that deserves the word test, and it is worth knowing why before someone sells you the idea.

A valid controlled test needs volume, and the amount is not a number you can look up. It falls out of how much your unit counts bounce week to week and how small a drop you want to catch. At 95% confidence and 80% power the arithmetic carries a constant of 2 × (1.96 + 0.84)², which is 15.68, call it 16. For counts, the smallest relative change you can detect is roughly the square root of 16 divided by the units on each side.

Put your own dish through it. An item selling 340 a month splits to 170 per side, and the square root of 16 over 170 is 0.31. The smallest change you could detect is a 31%derived collapse in orders. A 5% price rise that cost you 8% of orders, the outcome you actually care about, would be invisible. And that is before the practical problem: two prices at once means printing two menus and deciding which guest gets which.

So price testing in a restaurant is not an experiment. It is a before and after observation with season, weather, staffing, and every other thing moving at the same time.

Say that out loud rather than pretending otherwise. The defensible method is the one described here: change a small number of items, hold everything else still, watch unit counts on those specific items for four weeks against the same items the prior month, and accept that you are reading a signal rather than proving a cause.

That is not a weaker method than a split test. It is the honest description of the only method available, and it is what makes the four week watch window in the next section a discipline rather than a formality.

What should you watch after the change?

Units on the items you moved, for four weeks, compared against the same items the month before.

You are looking for a real drop in a specific item, not a change in total sales. Total sales moves for a dozen reasons and teaches you nothing about pricing.

Two reading mistakes. Do not judge in week one, because regulars cycle through on their own schedule. And do not compare a month with a holiday in it to one without.

Before you conclude the price caused a cover drop, check whether anything else moved. In the restaurants I ran, a sudden slow stretch was more often a schedule or an equipment failure than a pricing decision. Confirm you are still visible where people look, because a restaurant that disappears from the map at dinner time produces the same symptom for unrelated reasons.

If you run more than one restaurant, roll the change out at one and watch it for a month. The reasons two locations perform differently usually include a different guest with a different tolerance.

What to do this week

Pull ninety days of item mix and list your top twenty sellers with current price and current food cost. That table is the entire decision.

Mark your three or four anchor items. Be honest about which ones your regulars can quote. Ask two servers if you are not sure, they will know immediately.

Pick five items that are not anchors, sell moderately, and have moved in cost. Move those and only those.

Schedule the change for a menu update rather than a random Tuesday. If none is planned, plan one.

Brief the staff with a single sentence they can say out loud. Then watch units on those five items for four weeks before touching anything else.

Questions about a price move?

Email me at eric@seod.com with five items you are thinking about, their current price, and what each costs you to make. I will tell you which ones I would move, which one I would leave alone no matter what the math says, and whether I would change the dish instead of the number.

Sixteen years running multi unit restaurants and more than $54M in annual P&L, and I have watched price moves land quietly and land badly. The difference was timing and which items moved. Making the call yourself means holding your own numbers, which is the point of owning the data you should be able to take with you.

One prerequisite. If your labor is set by feeling rather than volume, price is the wrong lever, and staffing to a forecast comes first.

Otherwise, there is more on running the operation.