On this page
There is a version of pour control that works and a version that poisons a bar, and they use identical numbers. The difference is entirely in how the finding is framed and who it is addressed to.
The version that poisons goes like this. Management installs a measurement. The output is a ranking of staff by how heavy they pour. That ranking is raised in a one to one. The bartender concludes, correctly, that they are now being watched, and starts optimising for the measurement rather than for the drink.
What people do when they are measured badly
Nobody responds to surveillance by simply complying. They respond by managing the signal.
Pours get careful when a manager is on the floor and revert when they are not. Comps get logged as something else. Waste stops being reported, because reporting it is what makes you look bad. The measurement improves and the underlying behaviour does not, which is worse than not measuring, because now there is a number giving false comfort.
This is not cynicism about bar staff. It is what any competent person does when the thing being measured is them rather than the work.
The reframe that makes it survivable
The unit of the conversation has to be the drink, not the bartender.
"The house pour on this spirit drifted about ten percent heavy last week, mostly Friday and Saturday after ten. That is roughly this much money. Here is the jigger, here is why the recipe is balanced where it is, we will look again next week." That is the same information, addressed to the bar rather than to a person, with a cause and a remedy attached.
It works for a practical reason as well as a cultural one. Most drift is not one person being careless. It is a busy period rewarding speed, a new starter learning from whoever was next to them, a bottle without a pourer, a recipe nobody agreed on. Framing it as an individual failing gets you the wrong diagnosis as well as the wrong reaction.
Three rules that keep it fair
Fair and visible, not personal and private
If the standard exists, everyone should be able to see where the bar currently sits against it. A number discussed only behind closed doors reads as a case being built. The same number on the wall reads as a shared target.
Explain the balance, not just the volume
A pour standard presented as a cost control measure is a rule imposed to protect the owner's margin, and it will be followed exactly as far as it is enforced. The same standard presented as the point at which the drink is correct is a professional standard, and professionals hold those without supervision.
Both things are true at once. Lead with the one that survives a busy Saturday.
Close the loop quickly enough to be believed
Feedback that arrives a month later cannot be connected to anything anyone remembers doing. It reads as an accusation about a period rather than a correction about a shift. Inside the week, people can still recall the nights, which is what makes the conversation concrete rather than defensive.
Where the measurement should sit
The useful cut is by product and by shift, not by person. Which spirits drifted, on which nights, in which sections. If a specific individual is genuinely the cause, that pattern will be obvious from the shift data without anyone having built a leaderboard to find it, and it can be handled as the exception it is.
That ordering matters. A system that starts from the product and occasionally lands on a person is a quality programme. A system that starts from the person is a monitoring programme, and bar teams can tell the difference within a week.
The honest limit
None of this eliminates drift. Free pouring varies, busy services vary, and a bar that demands perfect consistency from people will lose the people. The realistic goal is a band you have chosen rather than a band that chose itself.
If you want to see where your own band sits, by product and by shift rather than by name, that is what a CoreTAP demo shows.