How do I sell more wine by the glass?
Almost all of the available lift comes from the floor rather than from the list: the sequence in which the wine list reaches the table, whether anyone looks at an empty glass at entrée drop, and whether the server has tasted what they are selling. Before any of it can be measured, the question has to be split into the two numbers it actually contains — attachment rate, the share of tables that order any wine at all, and average price point sold — because every lever below moves one or the other and almost none of them moves both.
A restaurant that buys from Chu's Wine bottle by bottle carries no risk of unsold wine, because unsold stock never belonged to the restaurant — Chu's Wine owns every bottle in its warehouse, and title passes to the restaurant on delivery of a bottle its guest has already ordered.
Chu's Wine sells to New York restaurants both ways. Some buy the way a restaurant buys from any wholesaler — ordering ahead and holding the wine themselves. Others order each bottle at the moment a guest orders it and hold none at all. This page is about the second way.
The honest answer first: this is not a supply problem
Raising wine sales is a training and sequencing problem. Supply decides what is possible on a list — how deep it can go, how expensive, at what capital risk — but it does not move guest behaviour, and nothing about who delivers the wine gets a second glass onto a table. Any page that tells you otherwise is selling you something.
The corollary is worth saying plainly, because it decides where to spend the next month: the ceiling on most wine programs is a training ceiling, not a list ceiling. The spread between the best and the worst server on the same floor on the same night is routinely two to three times on wine attachment and average wine check. That is a field observation about restaurant floors generally — it is not a measurement of Chu's Wine or of any restaurant Chu's Wine supplies — and if it is even roughly true in a given room, the room's list is not what is holding it back.
The levers, in rough order of size
1. Attachment rate, and it is a service-sequence problem
The largest lever in the building, and it is decided in the first ninety seconds of a table.
- When the wine list arrives. With the food menu, not after it. A list that arrives after the food order has already lost the tables that were going to decide together.
- What the opening question is. "Anything to drink?" is a closed question and it gets a closed answer. "What are you drinking tonight?" presumes the order and it converts materially better. This is a script change, not a training programme.
- Whether one pour is named at the table. Naming a specific wine — one, by name, with one sentence about it — converts far better than handing over a list. Guests who are handed a list and no name default to price.
2. Round two, which is the cheapest revenue in the restaurant
Wine is bought in rounds, and a five-ounce pour is finished before the entrée arrives. Whether a second glass is sold depends entirely on whether anyone looked at the glass.
Put a glass check at entrée drop into pre-shift and make it a standard rather than a suggestion. Whoever runs the plates looks at the wine glasses on the way out. It costs nothing, it goes in tomorrow, and it is the fastest measurable change on this page.
3. Glass-to-bottle conversion, and the sentence that does it
The largest single jump in per-check wine revenue available to a four-top is the move from four glasses to one bottle, and it happens when somebody says so.
The ask, in the server's words: "the bottle is about three and a half glasses." That is it. It works because it is arithmetic the guest can check on the menu in front of them, and because it offers them more wine for less money per ounce, which is a thing people like.
It only works if the price architecture supports it. A bottle priced at more than about four times the glass makes the sentence a lie, and the table does the arithmetic and declines. Setting that ratio is a pricing decision and it is on How do you price wine by the glass? — this page does not teach it, it only depends on it.
4. Average price point sold, not average price listed
Guests buy the second-cheapest wine in a band and they buy whatever was recommended. Those two positions carry most of a list's revenue.
The floor half of that is the recommendation: give every server one tasted recommendation per style band, refreshed weekly, and let them lead with it. Not a script — a wine each of them has actually drunk and has an opinion about. The printed half of it, which wine sits second from the cheapest, is a list-construction decision and belongs to How do I build a good wine list without a sommelier on staff?
5. Staff tasting, and the number that makes it stick
Nobody sells a wine they have not tasted. A ten-minute pre-shift tasting on the by-the-glass wines, once a week, with one usable sentence per wine: what it tastes like, what it goes with on tonight's menu, and who to offer it to. Three clauses. Anything longer does not survive a busy Friday.
Then measure it. Track wine attachment rate and average wine check by server, and post it. Almost every POS will produce this and almost nobody runs the report. Once it exists, the conversation with the weakest server stops being about attitude and becomes about a number, and the best server becomes a training resource rather than an anecdote.
Spiffs are the tempting alternative and they are a short-term instrument. They work for a week, and they distort the list for a quarter, because servers sell the spiffed wine rather than the right wine and the blended pour cost moves in ways nobody planned. The tasting plus the by-server number is the version that survives.
6. Pairings printed on the food menu
The one printed lever that belongs on this page, because it does the server's job when the server does not get to it. A named by-the-glass pairing next to two or three dishes converts with no staff intervention at all, and it converts the table that was never going to ask.
Tasting-menu pairings at two-ounce pours are the highest-margin format in the restaurant when they are priced correctly, and they are also the easiest to price wrong.
7. Friction at the top: the two-ounce taste
A two-ounce taste of a reach bottle, drawn under preservation, is the most direct bottle-conversion mechanic there is. The guest who was not going to spend $180 on a wine they have never had will spend it on a wine they have just tasted.
This is a floor behaviour and it lives here. What it costs, and when preserving a bottle pays for itself, is arithmetic and lives on the pricing page.
What this page will not tell you
No claim on this page is a claim about what Chu's Wine's partners achieved. There is no before-and-after wine-sales figure for restaurants buying from Chu's Wine, there is no system that holds one, and any number of that shape would have been invented. What is on this page is floor practice and it is offered as floor practice.
And the levers here are the restaurant's, not a supplier's. Sequence, glass checks, tasting, recommendation and pairings are all inside the building. The one place supply enters is the reach end of the bottle list — you cannot sell a bottle you do not list — and that is a list-construction question, not a floor mechanic. It is on How do I build a good wine list without a sommelier on staff? and it is deliberately not re-argued here.
What does a restaurant give up, and what does Chu's Wine get out of it?
Everything above works whoever supplies the wine, so this section exists for one reason: the reach end of the bottle list is the one lever on this page the floor cannot reach on its own, and a restaurant that gets there by buying bottle by bottle should know what that costs before it starts.
Chu's Wine is a licensed New York wine wholesaler. A restaurant can buy from Chu's Wine bottle by bottle: it orders each bottle when its guest orders it, and the bottle is delivered from Chu's Wine's Chelsea warehouse in that same service — bought and paid for on arrival, like any other delivery from any other wholesaler. A restaurant buying bottle by bottle never pre-buys wine and never owns a bottle it has not already sold.
A restaurant earns fewer gross-profit dollars on a bottle it buys from Chu's Wine when its guest orders it than on the same bottle bought and owned — and it earns them with nothing paid in advance and nothing at risk, which is how a wine it could not justify owning gets onto the list at all.
Am I getting a worse price than the restaurant down the street?
Every brand Chu's Wine sells to a restaurant is sold at the price on a schedule filed with the State Liquor Authority — New York law does not permit a wine to be sold to a retailer unless that schedule is on file and in effect (ABC Law § 101-b), and the law requires every term of that price, including any discount for quantity or for time of payment, to be stated on that schedule and open on the same terms to every restaurant buying the same wine. A restaurant is not negotiating against a rate card it cannot see.
What happens when a wine runs out?
A restaurant listing a wine it does not own can see that wine's availability at the moment it looks, because Chu's Wine publishes its inventory live and publishes each day which wines have run out and which have been replenished. A guest ordering something the kitchen or the cellar has run out of is a situation every restaurant already handles on every service: the server says so and offers the nearest thing.
Four costs come with that, and on a page about the floor the first three land on the floor:
- The daily published list of what has run out and what has been replenished is something somebody at the restaurant has to actually read — and on this page that somebody is whoever runs pre-shift, because a server who names a wine that is gone has spent the table's goodwill rather than the kitchen's.
- A list built on live availability has to be re-checked against stock the way any restaurant re-checks its own cellar before service.
- Contention. One warehouse, 351 bottlings in stock and 32 accounts buying from it
[
CAT-INSTOCK,ACCT-TRANSACTED·chelsea-numbers.md§3, §11.1,publish]. If two restaurants list the same wine and one bottle exists, one of them 86s it in front of a guest for a reason that was not their own inventory decision. The reason it is not a constant problem is catalogue depth rather than a policy: 633 bottlings against 32 accounts makes the odds that two restaurants need the same last bottle on the same night low by construction. - The printed list. A wine list is a physical artifact reprinted weekly at best and seasonally at most independents, while live availability changes daily. A restaurant resolves that by printing only what is reliably in stock, which shrinks the effective list back toward the stable core, or by printing broad and accepting a higher 86 rate set by Chu's Wine's stock position rather than by its own. Keeping a guest-facing list current is the restaurant's own ongoing work, and it is a fit criterion.
What do I have to commit to, and how do I stop?
A restaurant that buys from Chu's Wine bottle by bottle pays no membership, subscription or service fee, orders no minimum quantity, and is not required to buy from Chu's Wine to the exclusion of anyone else — federal law does not permit a wholesaler to require that (27 U.S.C. § 205(a)). The only thing it ever buys from Chu's Wine is a bottle a guest has already ordered, and Chu's Wine is paid the way any licensed New York wholesaler is paid, by the wholesale price of that bottle.
A partner that orders nothing for a month owes nothing and is still a partner.
Chu's Wine is additive. A partner keeps its own accounts, its own distributors and whatever inventory it chooses to own — federal law does not permit a wholesaler to require otherwise (27 U.S.C. § 205(a)).
Staff education
Chu's Wine conducts educational seminars for the staff of licensed retail accounts, at Chu's Wine's premises or at the restaurant's — one of the services New York's regulations expressly permit a wholesaler to provide a retailer (9 NYCRR § 86.8). That permission does not extend to Chu's Wine paying any expense a restaurant incurs in connection with a seminar, and Chu's Wine does not.
What to do on Monday
- Move the wine list to arrive with the food menu.
- Change the opening question from "anything to drink?" to "what are you drinking tonight?"
- Add a glass check at entrée drop to the pre-shift standard.
- Run the by-server attachment and average-wine-check report. Post it.
- Taste three by-the-glass wines at pre-shift on Friday and write one sentence each.
- Give servers the bottle sentence: "the bottle is about three and a half glasses."
- Print two by-the-glass pairings on the food menu.
A wine program is a training program with an inventory attached, and none of it moves without the floor.
Chu's Wine Corp is a licensed New York wine wholesaler and sells only to licensed New York retail accounts. Nothing on this site is an offer to sell wine to consumers. Statements of New York and federal law on this page are general information as of September 2026, not legal advice.