How to Decide Which Dishes to Cut From Your Menu

Cutting a dish feels more personal than it should. Someone on staff probably likes it, a regular might ask for it by name once a month, and there is a nagging worry that removing it will upset exactly the wrong guest at exactly the wrong moment. That emotional weight is why so many menus carry three or four dishes everyone privately agrees should have been retired a year ago.

The clearest candidates for the chopping block

A dish is a strong candidate to cut when it combines low order volume with a middling or poor reaction from the guests who do order it. That combination, in menu engineering terms, is a dog: it isn't selling and it isn't delighting the few people who try it. There is no real argument for keeping a dog on the menu except sentiment, and sentiment does not pay for the ingredients sitting unused in your walk-in.

A dish that sells rarely but earns strong satisfaction from the guests who do order it is a different situation entirely. That dish is not failing on quality. It is failing on visibility, and the fix is better placement or a sharper description, not removal.

• Low order volume and low satisfaction together: a strong candidate to cut
• Low order volume but high satisfaction: a placement or description problem, not a quality one
• High labor or prep time relative to how often it sells, straining kitchen bandwidth for little return
• Ingredients that only exist on your shelf for this one dish, adding waste risk without enough volume to justify it

The mistake most owners make instead

The common failure mode is cutting based on gut feeling rather than data, usually driven by whichever dish an owner personally likes least or whichever one a vocal staff member complains about prepping. That approach cuts dishes for the wrong reasons roughly as often as it cuts them for the right ones, because personal taste and staff annoyance are not reliable proxies for what guests actually think.

The other common mistake is never cutting anything at all, letting a menu grow indefinitely because removing an item feels riskier than leaving it. A menu that only ever adds and never subtracts eventually becomes the bloated, choice-fatiguing spread that hurts guest satisfaction and kitchen efficiency at the same time.

Give a dish a fair trial before you retire it

Before cutting a dish that has genuinely never had a fair shot, consider giving it one deliberate push first: better placement, a rewritten description, maybe a server mention for a couple of weeks. If it still doesn't move after a real attempt to help it succeed, you can cut it with a clear conscience, knowing the decision wasn't premature. If you skip this step and cut based on a first glance at low sales, you risk losing a dish that just needed a better introduction, and you won't find out until it's already gone and a guest asks why.

Keep a simple record of what you tried and when, even something as basic as a note in a shared document. Six months from now, when you're staring at another underperforming dish, that record will save you from repeating a fix that already failed once.

It's also worth checking how a candidate for cutting connects to other dishes on your menu. A slow-selling side that's actually the only vegetarian option in its category serves a purpose beyond its own order count, and removing it might cost you guests who need that option even if the side itself rarely gets ordered on its own.

A calmer way to make the call

Before cutting anything, check three things: how often it actually sells, how the guests who order it rate it, and how much kitchen effort it demands relative to that volume. A dish that fails on all three is an easy cut. A dish that fails on only one deserves a targeted fix first.

TasteGraph's dashboard flags dishes that are genuinely underperforming under its Flag as at risk and low-engagement categories, drawing on real order counts and guest reactions rather than staff opinion, so the decision to retire a dish is backed by evidence instead of whoever complained about prepping it last.