In this article
Somewhere in every kitchen's recent past is a dish that ran as tonight's special for a couple of weeks, sold reasonably well, and then just — stopped. Nobody decided out loud that it had failed. It simply wasn't on the board the next Monday, and by Friday nobody could quite say why.
That disappearance is normal, and the research says it should be: guests genuinely want variety, and a kitchen that never changes its board is leaving real money on the table. But “guests want new things” is only half the argument that gets a dish onto tonight's specials list — and it's the half almost every kitchen already knows. The half nobody tracks is what the dish cost to get there, and whether it ever earned that back before someone quietly decided to stop making it.
That's not a rhetorical question. Every dish that gets tested costs chef hours, wasted trial batches, and a slot on the specials board that could have gone to something else — and none of that is free, even though almost no kitchen writes it down anywhere. A dish is usually judged on a feeling (“it sold okay”) rather than on the one question that actually matters: did it earn back what it cost to develop, in the time it was actually given to prove itself?
This article puts seven numbers side by side — six from research and industry data, one from a worked example — to put the real pull of a new dish (why guests want it, why it's worth testing) next to the real cost of testing one (in hours, ingredients, and the trial window most kitchens never measure against). The calculator further down takes your own numbers and tells you, for your own dish, whether it's already paid for itself — or how close it actually was when somebody decided to pull it.
HappyChef’s free new-dish pipeline board helps a kitchen move an idea from a napkin sketch through testing to a costed, permanent menu item; this calculator helps you check whether the dish that just came off your board earned its keep on the way.
Why most new dishes get judged on a feeling instead of a number
Variety-seeking is not a marketing slogan — it's a measured, well-documented part of how people eat. Even guests who genuinely like your food get bored of ordering the same thing, and a menu that never changes is quietly asking them to fight that instinct every time they walk in. That's the strongest argument for testing new dishes, and it's real.
But “guests want something new” says nothing about whether any particular dish is worth what it cost to get onto the board. Every tested dish has a real price tag — chef hours at the kitchen's own labour rate, ingredients used across however many trial batches it took to get the recipe right — and that number rarely gets written down anywhere, let alone compared against what the dish actually sold during however long it ran as a special.
So the decision to keep a dish or drop it usually comes down to a feeling: “it did fine”, “it didn't really take off”, “people seemed to like it”. None of those are wrong exactly — they're just answering a different question from the one that decides whether testing that dish made or lost the kitchen money. This article runs both numbers, on your own dish, so the two questions don't get mixed up.
The 7 numbers behind the dish on tonight's board
Six of these numbers come from research and industry data, one comes from a worked example — together they're exactly the model behind the calculator further down.
1. 44% — how much more people eat when a meal offers variety instead of repetition
In a widely cited 1984 study, Barbara Rolls and colleagues fed participants either four courses of the same food or four courses of different foods (sausages, bread and butter, chocolate dessert, bananas). The group offered variety ate 44% more overall — not because they were hungrier, but because the pleasantness of a food declines with every bite of it while unfamiliar foods on the same table stay appealing. Researchers call this sensory-specific satiety.
That's the actual mechanism behind “guests order more when something's new” — it isn't guesswork or a sales trick, it's a documented shift in appetite. A menu that never changes isn't neutral; it's quietly capping what a table is biologically inclined to order.
2. 46% to 59% — the share of guests, by generation, more willing to try new flavours than three years ago
Industry research firm Technomic found that willingness to try new flavours has risen across every generation it tracks: 46% of Gen X, 59% of Millennials and 48% of Gen Z say they're more open to trying something new now than they were three years ago.
The order matters as much as the numbers: it isn't a straight line from oldest to youngest. Millennials — not Gen Z — are the most open group, which is a useful correction if “novelty” has become shorthand in your head for “only the youngest tables care”.
Share who say they're more open to trying something new than they were three years ago.
Millennials, not Gen Z, are the most open group — a useful correction if “novelty” has quietly become shorthand for “only the youngest tables”.
3. 52% — the share of guests for whom something new actually decides which restaurant they walk into
The same research tracks a second, separate question: not whether guests will try something new once they're seated, but whether the availability of something new or limited-time influences which restaurant they choose in the first place. That share was 52% in the most recent measurement — up from 48% in 2021.
That's the part a kitchen chasing “reduce food waste with a special” usually misses: a genuinely new dish isn't only a retention tool for guests already walking through the door. It's also, measurably, an acquisition one.
4. 26% — how much higher the average check runs during a limited-time special
Checks that include a limited-time item run meaningfully higher than checks that don't — a commonly cited industry figure puts the gap at 26%. Some of that is the item itself; some of it is that a limited-time dish gives a table a reason to add something rather than order exactly what they always order.
That's the reward side of testing a new dish, and it's real. The rest of this article is about the side that reward gets weighed against — and how rarely that weighing actually happens.
5. Just over half — the share of tested dishes that actually earn a permanent spot on the menu
Industry surveys of operators running limited-time offers consistently find that only just over half of them ever earn permanent status on the regular menu. The rest — a large minority, often close to half — simply stop appearing, without the operator necessarily framing it as a failure.
That's the number this whole article is built around. Half is a coin flip, and a coin flip is exactly what you get when a decision is made on “did it feel like it was doing okay” instead of on whether the dish had actually paid back what it cost to test by the time somebody made the call.
6. €746 — what a tested dish actually cost before a single portion of it went out the door as “the special”
Take a dish that needed 18 hours of a chef's time to develop and test, at a blended kitchen labour rate of €32.00 an hour, plus €170 in ingredients burned across the trial batches that got the recipe right. That's €746 spent before the dish ever appeared on a chalkboard — a number that exists in every kitchen's own bookkeeping somewhere, just never added up as one line.
Run it as tonight's special for 14 days, selling 6 portions a day at a margin of €8.20 each, and the trial earns back €689 in that window — €57 short of the €746 it cost to get the dish onto the board in the first place. And this is a dish that sold every night, at a completely ordinary rate. It just didn't have enough days to catch up.
How many days it takes to break even on developing the dish, against how many days the trial actually ran.
Once the bar passes the dashed line, the dish would have paid for itself — if the trial had run that long. Cut it right at the deadline, and it never gets the chance.
7. 10 to 70 — the measured range of how strongly a guest resists trying something new
Not every guest wants the new thing, and that resistance is itself measured. The Food Neophobia Scale, developed by researchers Patricia Pliner and Karen Hobden in 1992, scores how reluctant someone is to try unfamiliar food on a ten-item questionnaire, from 10 (highly open) to 70 (highly resistant) — and real populations spread across that entire range, not clustered at one end.
That's the number that decides how a new dish should be introduced, not whether. A completely unfamiliar dish with no anchor to anything a guest already recognises loses the neophobic half of the room before they've even ordered; a familiar dish with one new twist — described well, per the research this site's own daily-specials article already covers — keeps both halves of the table happy at once.
Has your dish paid for itself yet?
Enter what this dish actually cost you to develop, how it ran as a special, and its price and cost per portion. The calculator works out the real development cost, how many portions it needs to break even, how many days that takes at your own sales rate, and where you actually stand at the end of the trial you gave it.
This is a calculation model based on your own numbers, not accounting advice. For a dish's exact ingredient cost — yield percentages and sub-recipes included — this site’s free recipe-costing tool is the better instrument; this page specifically runs the development-cost-versus-trial-earnings comparison that decides whether a tested dish actually earned its spot.
Has this dish paid for itself yet?
Enter your own numbers — the calculator works out the real development cost, break-even portions, days needed, and where you stand at the end of the trial.
—
A calculation model based on your own numbers, not accounting advice. For a dish's exact ingredient cost including yield percentages: this site's recipe-costing tool.
What this produces is deliberately an indication based on your own numbers, not exact accounting — every kitchen prices trial batches, staffs testing and runs its specials board differently. What the model gets right is the comparison almost nobody makes: what a dish actually cost to develop, against what it actually earned back in the window it was actually given.
A dish that's already in the black by the end of its trial has more than proven itself — every portion from here is straight margin. One that isn't yet, but is close, is exactly the case this article is about: the dish that gets pulled two days before it would have turned the corner, with nobody ever framing that as the reason.
Testing a dish without guessing: the plan
No new paperwork — three moments where thirty seconds of arithmetic replaces a feeling.
Before you test a dish
- Log the hours and the trial-batch ingredients as you go, not from memory afterwards — that's the number the calculator needs, and it's the one almost no kitchen has ever written down.
- Decide the trial length and a realistic daily-sales target before service starts, so the go/no-go moment isn't decided by mood on a random Tuesday.
While it's running as tonight's special
- Track daily portions sold like any other special — this site’s free daily-specials pricing calculator works just as well for a dish you're testing as for one clearing surplus stock.
- Brief the floor on it every service. How confidently it's pitched decides a real share of whether it sells at all — not just its price.
At the go/no-go moment
- Run the calculator above before deciding anything: has the dish actually paid back what it cost to test, or does it just need a few more days that a fixed trial length never gave it?
- If it's cut, don't lose the work — move it, already costed, into this site’s free new-dish pipeline board or straight onto a future specials run instead of starting from a blank page next time.
The number that actually decides it
A dish that “sold fine” for two weeks says nothing on its own about whether testing it made the kitchen money. What decides that is a number almost nobody runs: what the dish cost to develop, against what it earned back in the actual window it was given to prove itself.
The gap between a dish that genuinely wasn't worth keeping and one that was cut two days too early rarely lives in the recipe. It lives in a sum most kitchens never do — and the difference between guessing and running it is thirty seconds with a calculator.
Run your own numbers through the tool above before the next dish quietly disappears from the board. If it's already paid for itself, you know to keep going. If it hasn't, you'll know exactly how close it was — and whether it deserved a few more days.