How to run a ticketed beer release without a queue meltdown

Why does a single-door-count release produce a queue?

Because everybody arrives at the start. A release advertised as "doors at 7" is read by guests as "be there at 6:45 or miss out", especially when the beer is limited and they suspect it will go. You have not sold ninety admissions spread over an evening; you have sold ninety simultaneous arrivals to a bar that can pour maybe forty in the first half hour.

The visible symptom is the queue. The expensive symptoms are the ones behind it: guests who leave, staff who cannot check ID properly under pressure, and a first-hour service standard that decides whether anyone comes back.

How many spots should a slot hold?

Size the slot to your pour rate, not to how many bodies fit in the room. Take the number of servers who will actually be pouring, multiply by the pours one server completes in fifteen minutes at a release pace, and multiply by four for the hour. That number, minus a little headroom for the people already in the room drinking, is your slot size.

Most small taprooms land between twenty and forty per slot. If the maths gives you sixty, check whether it assumed a server who never talks to anyone — at a release, the conversation about the beer is the product, and it is slow.

• Count pouring servers, not total staff on shift.
• Time one real pour-and-explain during a normal week and use that number.
• Leave 15–20% headroom for walk-ups and guests who overstay their slot.
• Keep the last slot smaller — it absorbs the drift from every slot before it.

When should the on-sale go live?

For a low-tier vertical like brewery events, the median lead time bindro models is ten days. That is your planning horizon: an on-sale three weeks out mostly buys anxiety, and an on-sale two days out gives your regulars no chance to arrange a lift home.

Announce the release first and put the on-sale at a stated time a few days later. A stated on-sale time concentrates demand into a moment you have staffed, which is far easier to handle than a slow trickle you have to keep watching.

What actually happens when two people buy the last spot at once?

One of them gets it and the other is told, before any money moves. Inventory in bindro is only ever changed under a row lock inside the transaction that confirms the order, and there is a database-level constraint behind that as a backstop rather than as the mechanism. The practical consequence for a taproom is the one that matters: you cannot sell ninety-one spots to a ninety-spot release, and no guest is charged for inventory that was gone.

How should the door be run on the night?

One person scanning, one person checking ID, and do not combine them. The scan answers "is this spot admitted"; the ID check answers "is this person 21", and the second question is the one with a licence attached to it. Bindro records the ID confirmation against the scan rather than the order, precisely because the booker is often not the drinker.

Door devices carry a signed manifest and scan with no connection, so a cellar bar or a railway arch with no signal is a normal operating condition rather than an incident. Re-entry is on for this vertical, so a guest who steps out to the beer garden scans back in on the same code.

• Scanner and ID checker are two people, always.
• Brief door staff that a guest without ID is refused and not refunded — it is on the booking page, so it is not a surprise.
• Door staff see a name and an admit/refuse; they do not see prices or other bookings.
• Sell the walk-ins through the same door checkout, so the room count stays true.