How Many Wines Should a Restaurant Wine List Carry

How many wines should a restaurant list carry? There is no right number, and the number is not the right question. A list of two hundred wines, sixty of which never sell, is worth less than a list of forty that all turn. The only useful question is what share of your list is actually working.

Why the question is usually asked wrong

Operators ask us regularly how many wines they should have. There is no answer. And in looking for that figure, they miss the question that actually decides whether a list is profitable: what proportion of your references are working for you.

Long lists are usually built by accumulation, one cuvée after another, with nobody ever taking anything off. Each addition looked harmless. The result is not.

So ask it differently. Not how many wines, but how many wines are ordered at least once a month.

What a reference really costs

Every line on your list ties up money in the cellar. That is the obvious cost. There are three others, less visible.

• Rotation cost. A bottle that sits for eighteen months takes up space, ages unsupervised, and may well sell below its value when you finally want rid of it. The lists we see commonly hold 12 to 18% of their value in references that no longer move.
• Inventory cost. Counting two hundred references takes longer than counting eighty, and you pay that difference every single time. Most restaurants end up counting less often, which makes worse the exact problem they were trying to avoid.
• Attention cost. Your team cannot know two hundred wines. They will know twenty and sell those. The rest exists on paper and nowhere else.

The ceiling nobody measures: your dining room

Here is the figure that counts, and nobody calculates it.

A trained server recommends about fifteen wines with conviction. That is what they retain, what they have tasted, what they can talk about. Beyond that they stop recommending. They start reading aloud.

This does not mean you should carry fifteen references. It means that past a certain size your list needs structure to stay usable: legible sections, a sort by style rather than by region alone, a filter or a virtual sommelier if the list is digital.

A well-structured list of a hundred and twenty navigates better than a disorganised list of sixty. Size is not the problem. The absence of a path through it is.

Glass and bottle have different ceilings

This distinction changes everything, and it is routinely forgotten.

Your by-the-glass selection has a physical ceiling. An open bottle degrades, so each glass reference has to turn fast enough not to be lost. That ceiling depends on your volume, not on your ambition.

Take a concrete case. An open bottle holds three days in good conditions, four with a preservation system. On a 750 ml bottle poured at 5 oz, you need to move five or six pours in three days to lose nothing — about two a day. A restaurant selling forty glasses a week can carry six or seven references by the glass. The same room offering twelve will pour half of them down the drain and never notice, because the loss disappears into food cost.

Do this calculation with your own numbers before adding a glass reference. It takes two minutes.

Your bottle list has no such constraint. An unopened bottle waits. It costs you tied-up capital, not outright loss. So the two halves of your list are sized separately.

A list that is right in December is wrong in July

Most operators size the list once and then let it live. That is a mistake.

Your volumes change with the season, your guests change with it, and what they ask for changes most of all. A restaurant with a terrace can double its by-the-glass service in summer while its bottle sales slow. The right number of references is not a number, it is a range that moves.

Two adjustments a year is enough. One before the high season, one after. Take out what has not turned, bring in what suits the season ahead, and hold the total roughly stable rather than letting the list grow at every pass.

Three measurements that settle it

No general rule will replace these three figures, and you can have all of them in an evening.

• The dormant share. How many of your references have not been ordered in six months? If it is more than a quarter, your list is too long for your restaurant, whatever its size.
• Style coverage. Can a guest looking for a crisp white, a light red or a dessert wine find one? A list can be long and still have gaping holes. This is common, because accumulation follows buying opportunities rather than what guests ask for.
• Inventory time. How many hours does a full count cost you? If the answer makes you put it off, the list is above what you can hold.

These three measure different things. A list can be short and badly covered, or long and turning perfectly. The number alone says nothing.

If your sales sit in a POS, the dormant share comes straight out of it. If not, an evening with six months of order sheets is enough to spot the references that never appear. The exercise is not pleasant. It is short, and it almost always surprises whoever does it for the first time, because we remember very poorly what we do not sell.

If your list is already too long

Taking a reference off is harder than putting one on. Nobody enjoys admitting a purchase was a mistake.

Work in small waves rather than one clear-out. Take off five dormant references, move them by the glass or as a suggestion, and watch what happens. Nobody asks for them.

A reference coming off has three possible destinations, and it is better to choose than to let it happen. By the glass, to clear it fast even at a thinner margin. As the suggestion of the day, if it deserves one last defence. Or to staff, which has the advantage of getting your team tasting and turns a loss into training.

And set yourself one rule before the next purchase: one in, one out. Most over-long lists did not get that way by choice, but by the absence of a rule.

How a digital list helps you hold the line

A digital wine list lets you switch a reference off without deleting it. The decision becomes reversible, and therefore easier to take, and the history stays available if you want it back later. Stock adjusts straight from the tablet on the floor, at the moment the bottle is opened, which makes the dormant share measurable instead of estimated.