The Complete Guide to Wine Inventory Management for Restaurants

The cellar and the menu rarely agree

Ask most beverage directors what they have in the cellar and you'll get a confident answer. Ask them to prove it, bottle by bottle, and the confidence drains away. The list says one thing. The shelf says another. Somewhere between the last delivery, three covers of a by-the-glass pour, a transfer to the private dining room, and a breakage nobody logged, the number drifted.

This is the quiet tax on a wine program: you stop trusting your own stock. And once you stop trusting it, you over-order to be safe, you "86" wines mid-service, and you make pricing decisions on numbers you can't defend.

Good wine inventory management isn't about counting harder. It's about modeling stock in a way that survives contact with a busy service. This guide covers the difference between counts and lots, why movement history matters, where POS and spreadsheets fall short, and what a system you can actually trust looks like.

Counts vs. lots: the distinction that changes everything

Most inventory tools treat wine like any other SKU: one line, one number. You have twelve bottles of the Barolo. That's a count.

A count answers one question, how many?, and even then only at the moment you last looked. It can't tell you what you paid, when it arrived, which delivery it came from, or whether the twelve bottles on the shelf are the same twelve you bought at the same price.

A lot is different. A lot is a specific batch of a wine, tied to its acquisition: the cost you paid, the date it landed, the supplier, the conditions of purchase. When you buy the same Barolo twice at different prices, that's two lots, not one count that quietly averages your cost basis into mush.

Why does this matter for a restaurant rather than a collector?

Cost basis is real, not blended. If your last allocation of that Barolo cost 40% more than the case you bought last spring, lot-based tracking keeps those costs distinct. Your margin math stays honest.
You can value what you hold. Knowing what you paid, lot by lot, is the foundation for understanding what your cellar is worth, at cost, at selling price, or against the market.
Movement becomes traceable. When stock leaves, it leaves a specific lot. That's what makes an audit trail possible.

Vinius models inventory as lots and movements rather than a simple count list. The result is inventory with memory, it knows not just how much you have, but the story of how it got there.

Movements: the verbs of your cellar

If lots are the nouns of inventory, movements are the verbs. Every meaningful change to stock is a movement:

Purchases, stock arrives, a lot is created or extended.
Sales, a bottle goes out the door at the bottle price.
Pours, a by-the-glass serving draws down a bottle in millilitres, not whole units.
Transfers, stock moves between sites, rooms, or storage units.
Adjustments, breakage, corkage, tasting, staff training, spoilage. The honest losses.

The point of tracking movements is not bureaucracy. It's that every change has a cause, and when you can see the cause, variance stops being a mystery. A bottle didn't "disappear." It was poured, or transferred, or broken, and the record says so.

This is also where by-the-glass programs live or die. A pour is a movement measured in millilitres. If your system can't represent a partial draw on a bottle, it can't reconcile a glass program against bottle stock, and the leak becomes invisible. (We go deep on this in our piece on by-the-glass margin.)

The audit trail: what changed, when, and why

Put lots and movements together and you get the thing every serious operator eventually wants and almost never has: an audit trail.

An audit trail answers questions you currently answer with a shrug:

Why is variance high on the Champagne this month?
Who transferred six bottles to the second venue, and when?
Did that case of Sancerre actually arrive, or did we pay for a delivery that never landed?
What did we lose to breakage last quarter, and is that normal?

A trail turns inventory from a snapshot into a ledger. Snapshots lie by omission, they show you a number without the history that explains it. A ledger shows you the sequence of events. When your monthly count is off, you don't start a witch hunt; you read the record.

For multi-person teams this is non-negotiable. The moment more than one person can touch stock, "I think someone moved it" is not an answer. The system has to know.

Why POS and spreadsheets fall short

Two tools usually try to do this job. Both were built for something else.

The POS knows what sold, not what remains

A point-of-sale system is an excellent record of transactions. It knows what rang through the till, at what price, at what time. That's genuinely useful data.

But a POS reasons in counts and sales, not lots and movements. It doesn't know your cost basis per delivery. It doesn't track a transfer to another site, a tasting pour for staff, or a breakage. It depletes a count when something sells and calls it a day. It is, fundamentally, a sell-side record, not a cellar system of record.

That gap is the subject of its own article: your POS isn't a cellar system of record. The short version: the POS closes the sale; it doesn't close the loop on what physically remains in your cellar at lot level.

Spreadsheets fail at the boring, important parts

Spreadsheets are flexible, free, and familiar. They're also where wine programs go to lose money slowly.

Capability	Spreadsheet	What you actually need
Multi-user editing	Conflicts, overwrites, "final_v3_FINAL.xlsx"	Concurrent, attributed changes
Audit trail	None, last edit wins	Every movement logged with cause
Lot-level cost basis	Manual, error-prone	Native, per acquisition
Sync with guest-facing list	Copy-paste, instantly stale	Generated from live stock
mL-level pour tracking	Effectively impossible	Built in

A spreadsheet can hold numbers. It cannot enforce discipline, preserve history, or stay in sync with the wine list a guest is reading. The day you publish a list from a spreadsheet is the day that list starts going stale.

What good looks like

A wine inventory system you can trust has a recognisable shape. You should be able to:

See stock by location. Site, room, storage unit. Not one global pile, but where things actually are, which matters the moment you run more than one space. (More on this in multi-venue wine pricing.)
Read the movement history of any lot. Every purchase, pour, transfer, and adjustment, in order, with cause.
Trust the cost basis. Lot-aware costs, not a blended average that hides your real margins.
Track pours in millilitres, so the by-the-glass program reconciles against bottle stock instead of leaking quietly.
Generate the guest-facing list from live data, so the menu and the cellar never drift apart. Stale lists are a symptom of inventory that lives somewhere the menu can't see. Vinius generates wine cards, branded PDFs and digital displays, directly from current stock.
Turn low stock into action. A reorder threshold that becomes a supplier-ready order, with the history to match.

This is the difference between inventory as a chore you do monthly and inventory as a live operating layer. The chore version tells you, once a month, how wrong you were. The live version tells you, continuously, what you have, so the menu is honest, the margins are real, and "86" moments stop ambushing you mid-service.

Getting there without a fire drill

You don't have to boil the ocean. A sane rollout looks like this:

Week one: import your existing inventory, set your storage locations, and get a baseline. Even the act of importing surfaces the gaps you've been carrying.
Early days: formalise movements. Pours, transfers, adjustments, get the team logging the verbs, not just counting the nouns.
Then: wire up reorder thresholds and connect the POS so sales deplete stock automatically and the loop closes.

The goal isn't a perfect cellar on day one. It's a cellar that, from day one, remembers what happens to it.

Vinius is built around exactly this model, lot-based inventory with full movement history, because everything else in a wine program (pricing, lists, reordering, valuation) is only as trustworthy as the stock data underneath it.

The takeaway

Counting harder won't fix a wine program. Modeling better will. Treat inventory as lots (what you hold, and what it cost) and movements (every change, with its cause), and you get an audit trail that answers the questions a count never could. POS systems track what sold; spreadsheets track nothing reliably. What good looks like is inventory with memory, live, located, lot-aware, and honest enough to price and publish from. Start by importing what you have and logging what moves. The trust follows.