The Ultimate Guide to Wine List Engineering

Written by Marco Azzari on 14 September 2026

Your best-selling wine is probably not your most profitable one, and your most profitable one is probably sitting in the cellar. Most lists I read have both problems at the same time, and nobody has noticed, because nobody has ever put the two numbers side by side.

Wine list engineering does exactly that. It's menu engineering, the method kitchens use to cost and rank every dish, applied to the wine list. By the end of this guide you'll have your own list plotted on one chart, every wine sorted into one of four boxes, and a clear next step for each one.

What is wine list engineering?

Wine list engineering is menu engineering for the wine list. You take every wine you sell, work out how much money each one leaves you per bottle or per glass, count how many you sold, and plot the two against each other. Every wine lands in one of four boxes, and each box comes with its own instruction: protect it, reprice it, push it, or cut it.

Food menus get this treatment all the time. A chef costs a plate down to the garnish, and a good F&B manager knows which secondi carry the kitchen. Wine rarely gets the same care. In a lot of venues I've worked in and around, the list was priced with one multiplier and a good feeling on the day it was printed, and nobody has read it line by line since.

The method asks two questions of every wine: how much does it earn you each time you sell it, and how often do you sell it? Neither number means much on its own. A bottle that earns €50 and sells forty times a year brings in less than one that earns €20 and sells four hundred times. You only see what the list is really doing when the two numbers sit on the same chart.

Why it's worth an afternoon of your time

Restaurants in Italy run on thin margins. Once food, staff, rent and taxes are paid, a typical restaurant keeps only a few cents of every euro it takes, and wine is often 10 to 20% of revenue in a sit-down restaurant.

At a few cents per euro, small amounts of money carry a lot of weight. Take a restaurant turning over €1 million a year. It keeps around €35,000, and its wine sales are roughly €150,000. Recover three points of margin on that wine and you've added about €4,500, which is an eighth of the year's profit.

Wine is also the cheapest place in the building to find that money. A plate passes through four or five pairs of hands before it reaches the table, while a bottle that's already in the cellar only needs opening. When you correct a price or a pour, nobody's shift gets longer, so whatever you recover goes straight to the bottom line.

The matrix is just as useful for what it tells you to leave alone. When a manager suspects the list is underperforming, the temptation is to change everything at once. The chart points you to the handful of wines that matter and lets you leave the rest alone.

And there's a benefit your guests will notice. A list that has been through this exercise is usually shorter and easier to read, because the wines nobody orders are gone and the good ones nobody understood finally get a proper description. Guests choose with more confidence, and your floor team has a list it can actually recommend from.

Number one: the wine margin, in euros

The first number is what each wine leaves you every time you sell it. The formula is simple:

Wine margin = menu price − VAT − bottle cost

Each of the three parts has a trap, so let's take them one at a time.

Take the VAT out first

Your menu prices include 10% VAT (IVA), the rate for food and drink served at the table, wine included. That slice belongs to the Agenzia delle Entrate from the moment the guest pays. To find the part that's yours, divide the menu price by 1.10.

A bottle listed at €34 becomes 34 ÷ 1.10 = €30.91. The other €3.09 is VAT.

Two mistakes are common here. The first is skipping this step entirely, which is how a list priced at three times cost ends up earning 2.73 times cost once the tax office has taken its share. The second is multiplying by 0.90 instead of dividing by 1.10. The 10% was added on top of your net price, so taking 10% off the total removes slightly too much and makes every wine look a little worse than it is.

Subtract what the bottle really cost

Use the price on your latest supplier invoice, before VAT, because a VAT-registered venue deducts that VAT. The latest invoice matters more than people expect, since suppliers raise prices quietly and the cost you remember from the day you listed the wine may be two increases old.

For our €34 bottle that cost €10: 30.91 - 10.00 = €20.91 of margin per bottle.

For glasses, divide by the glasses you really pour

By-the-glass needs one more step. Divide the bottle cost by the number of glasses you actually get out of a bottle, then subtract that from the net glass price.

The number on paper is easy: 75 cl divided by your pour. Costing sheets usually assume six calici of 125 ml. Real venues get fewer, often four or five, once you count the generous pours and the bottle opened for one table on Tuesday and poured away on Thursday.

Here's the same Lugana sold at €7.50 a glass. The net price is 7.50 ÷ 1.10 = €6.82. At six glasses a bottle, each glass costs €1.67 and leaves €5.15. At five glasses, each costs €2.00 and leaves €4.82. That's 33 cents per glass you thought you were earning and never saw.

To find your real yield, spend two weeks comparing the glasses rung on the till with the bottles opened for the glass programme. Use that figure in the sheet, not the one on the costing card.

Why euros, and not percentages

Most lists are judged by margin percentage, and percentages flatter cheap wines. In the example later in this guide, a Nero d'Avola earns 76% margin and a Barolo earns 58%, the lowest on the list. Yet every bottle of Barolo leaves €50 and every bottle of Nero d'Avola leaves €19.

Rent and wages get paid in euros. Percentages are useful for setting prices, but when you're deciding which wines deserve their place, the matrix has to run on the money each sale actually leaves behind.

Number two: how many you sold, and over which period

The second number is units sold per wine: bottles for the bottle list, glasses for the glass list. Your till's sales-by-item report has it. The harder question is which period to measure.

A full year is the best default. It captures seasonality on its own terms, from rosato and crisp whites on the terrace in July to spumante in December. Any shorter window rewards whatever happened to be in season and punishes the rest.

A season works if your venue or your list is seasonal. A restaurant in a ski resort, a beach restaurant open from May to September, or a venue that prints a summer list and a winter list should compare each season with itself: this winter against last winter, not winter against summer.

A month is for a first read. If the list is new and there's no year behind it yet, or you've just made a change, a month gives you an early picture. Read it knowing the season is shaping the numbers.

A week is too short for the whole list. One table ordering six bottles of the same red can move it from one box to another. Use a week only to check a single change, such as a new pour size or a new wine by the glass, and compare it with the same week before the change.

Whichever period you choose, a few rules keep the comparison fair:

• Use the same period for every wine. Mixing a year for some lines and three months for others makes the chart meaningless.
• Scale up wines that weren't there the whole time. A wine listed halfway through the year, or out of stock for six weeks, lost those sales because it was missing from the list. Divide its sales by the weeks it was available and multiply by the weeks in the period, or leave it out until it has a full period behind it.
• Take out one-off events. A wedding that ordered sixty bottles of the same white will turn a wine guests never choose into a Star.
• Leave comps and staff drinks out. They aren't sales. If your till records them as zero-price items, filter them before you copy the numbers. A staff drink is a decision, and it belongs on its own line.

Glasses or bottles: run two matrices

Run wine list engineering twice: once for the wines you sell by the glass, and once for the wines you sell by the bottle. Even when it's the same wine from the same case, a glass and a bottle are two different products.

The guest decides which one they want before they decide which wine. Someone ordering a glass is drinking alone, or wants something different with each course, or doesn't want to commit to 75 cl. A table ordering a bottle is choosing one wine to share for the whole meal. The glass list competes with the other glasses, and the bottle list competes with the other bottles. A guest who wants a glass of Lugana isn't weighing it against a bottle of Barolo.

The numbers behave differently too. A glass earns a few euros and a bottle earns twenty or more. A glass carries the cost of whatever gets poured away when an open bottle goes off, and a bottle carries none. If you put both on one chart, every glass lands in the low-margin half, every bottle lands in the high-margin half, and the two lines that split the matrix end up somewhere that describes neither.

So a wine sold both ways gets two rows, one on each matrix, each with its own price, cost per serving, units sold and box. Don't be surprised when the boxes disagree. The Lugana can be a Star by the glass, because it's the obvious aperitivo and the pour is priced well, and a Plowhorse by the bottle at the same time. That's two separate decisions: keep the glass as it is, and work on the bottle price.

Before you open Excel: three documents

You need three things on the desk before you start, and gathering them is usually the slowest part of the job.

1. The wine list as guests see it, with the current prices.
2. Your latest supplier invoices, so you have today's cost per bottle before VAT.
3. The till report for your chosen period, showing units sold per item. Most POS systems export it to Excel or CSV. If you can't find the export, your POS provider or your commercialista usually can in a few minutes.

Download the Excel template

The file does all the maths for you: the 10% VAT, the margin per wine, the two lines and the box each wine falls into. There's a tab for bottles and a tab for glasses, and the matrix chart draws itself. It opens with the ten-wine example from this guide, so you can see how it works before you type over it with your own list.

Download the Excel template

Draw the two lines that split the list

The four boxes come from two lines: one for popularity and one for margin.

The popularity line is 70% of the average units sold. If every wine on the list sold equally, each one would sell the average. Menu engineering calls a wine popular when it reaches at least 70% of that fair share. The 70% is there because sales on any list are lopsided: a handful of wines do most of the volume, and with the plain average as the line, almost everything else would count as unpopular.

The margin line is the average margin per sale. Total margin divided by total units. It's weighted by sales on purpose, because it should tell you what a typical sale earns you, not what a typical line on the price list says.

In the example below, the ten wines sold 1,540 bottles in a year, an average of 154 each, so the popularity line sits at 107.8 bottles. Together they left €38,706 of margin, so the margin line sits at €25.13 per bottle.

A worked example: ten bottles, one year

Here's an illustrative bottle list for a mid-sized Milan restaurant, measured over twelve months. Prices include 10% VAT, costs are before VAT.

Wine	Menu price	Cost	Sold	Margin per bottle	Total margin	Margin %	Box
Lugana DOC	34	10	410	20.91	8,573	68%	Plowhorse
Valdobbiadene Prosecco Superiore DOCG	38	8	380	26.55	10,087	77%	Star
Chianti Classico DOCG	42	13	260	25.18	6,547	66%	Star
Montepulciano d'Abruzzo	30	9	150	18.27	2,741	67%	Plowhorse
Sancerre	50	16	95	29.45	2,798	65%	Puzzle
Nero d'Avola, Sicilia DOC	28	6	70	19.45	1,362	76%	Dog
Champagne brut	85	31	60	46.27	2,776	60%	Puzzle
Timorasso, Colli Tortonesi	48	15	45	28.64	1,289	66%	Puzzle
Barolo DOCG	95	36	40	50.36	2,015	58%	Puzzle
Cannonau di Sardegna	30	10	30	17.27	518	63%	Dog

A few things jump out. The Lugana is the best seller on the list and earns less per bottle than either Star. The Nero d'Avola has one of the best percentages on the page and still lands among the Dogs. And the Chianti Classico is a Star by five cents, so a single supplier increase would move it into the Plowhorses without anyone noticing.

The four boxes and what to do with each wine

The box names come from food menu engineering, and they've stuck because they're easy to remember. What changes on a wine list is what you do about each one.

Margin per sale →

Puzzles Low sales, high margin. Make them visible. Stars High sales, high margin. Protect them. Dogs Low sales, low margin. Sell through, then cut. Plowhorses High sales, low margin. Reprice or recost.

Units sold →

Stars: high sales, high margin

These are the wines that pay for the list. Your job is to protect them. Never let a Star run out, because every night it's missing, the table that wanted it most likely orders something that earns you less. Keep its price and its place on the page steady, and check the supplier price every time you reorder, since a Star with a creeping cost quietly turns into a Plowhorse.

In the example, the Prosecco Superiore and the Chianti Classico are Stars. Watch the Chianti first.

Plowhorses: high sales, low margin

Guests like these wines and they earn too little. Cutting them would be a mistake, since they're often the reason a table orders a bottle at all. You have two levers instead.

The first is a small price rise. In my experience, a popular wine absorbs a modest increase far better than an unpopular one absorbs any change. Raise the Lugana from €34 to €36 and, at the same 410 bottles, you add €745 of margin a year from a single line.

The second is the cost. Renegotiate with the supplier, or replace the wine with one of the same style that costs you less, so guests get the same experience at a better margin for you.

Plowhorses also tend to gather at the cheap end of the list for a reason. A guest who doesn't want to look cheap often skips the entry bottles and orders the one just above them, so the wine in that slot sells on its position as much as its merits. In the example, the Lugana sits exactly there, just above the €28 to €30 entry wines. Whatever you put in that slot will sell, so make sure it earns.

Puzzles: low sales, high margin

Each sale of a Puzzle is worth a lot, and too few people buy it. Puzzles are usually one of two things: your most expensive bottles, or good wines your guests don't understand.

For the second kind, the fix is visibility. Give the wine a plain description of what it tastes like and which dish it suits, and a better spot on the page. Then open a bottle for the floor team, so they can recommend it with a straight face. A Timorasso sells badly when the guest has never heard of the grape, and much better when the line tells them it's the white to order with the fish.

For the expensive bottles, look at the price. With one flat multiplier, your top bottles end up looking expensive next to the shop price a guest can find on their phone in seconds. A regressive markup, where the multiplier gets smaller as the cost goes up, earns fewer euros per bottle and gives the bottle a real chance to sell. If the Barolo dropped from €95 to €80, each bottle would leave €36.73 instead of €50.36. If that sold 70 bottles a year instead of 40, it would earn €2,571 instead of €2,015.

Change one thing at a time and measure again after a month, so you know which change did the work.

Dogs: low sales, low margin

These wines earn little and sell little, and they hold cash on a shelf while they do it. Bottles that sit for 30 to 90 days tie up money you've already paid out, and trade sources put the spoilage risk at up to around 8% of wine inventory. The usual answer is to sell through what you have, by the glass or as a staff recommendation, and not to reorder.

Before you cut, ask whether the wine has a job. It might be the only passito for the dessert menu, or the one a regular asks for every Friday. A Dog with a job stays, at a price that makes sense. The rest can go, and your list gets shorter and easier for a guest to read.

In the example, the Cannonau and the Nero d'Avola are Dogs, and the Nero d'Avola proves the point about percentages: 76% margin, and still not worth its place unless it has a job.

How often to run it

Once a list goes live, it tends to stay untouched. As one wine-trade writer put it, "once busy season starts, most wine programs freeze for months." Meanwhile supplier prices keep moving, and a Star can slide into the Plowhorses between two reorders.

A rhythm that works for most venues:

• Once a year, run the full matrix on twelve months of sales.
• Every season, run a quick check on the lines that sell most, since that's where small moves add up to real money.
• After any change, such as a new price, a new wine by the glass or a supplier increase, re-run the lines it touches a month later.

The first time takes an afternoon, mostly spent gathering the three documents. After that the sheet is built, and an update is a matter of pasting in a new till report.

What the matrix can't tell you

The chart tells you which wines to look at. It can't tell you what to replace a Dog with, or why a good wine stays a Puzzle. Those answers come from two places a scatter chart never sees: whether each wine sits beside a dish that actually leaves your kitchen, and whether a guest can understand the line when they read it.

The best result I can show you came from exactly those two. A 40-seat bistro I worked with roughly tripled the wine it sold after we rebuilt its selection around the food menu and replaced wine names the guests couldn't read with plain descriptions that matched the dishes. Neither change would have shown up as a dot on this chart, although both would have moved the dots once the next till report came in.

And if you'd rather do it yourself, you now have everything you need. Export the till report this week, pull the latest invoices, and fill in the template. Then start with your best-selling wine, because if it turns out to be a Plowhorse, that one line is probably worth more than everything else on the chart.