What Is Draft Drift, and Why Does One Location Always Outperform the Others?
Draft drift is the performance gap between locations in the same group. Same beer, same equipment, same training, different results. It is almost never one bad manager. It is four things missing at once: no shared baseline, no way to compare sites, no root cause visibility, and no proof of what changed.
The four failures
No shared baseline. Each venue defines normal by its own history. A location running 12 percent variance for three years calls that normal because it has never seen 4 percent. Nobody is lying. They are measuring against themselves.
No way to compare. Two venues report waste differently, count kegs on different days, and use different POS categories. The numbers arrive at the office in incompatible shapes, so the group compares revenue instead and calls it a day.
No root cause visibility. You can see that Location 4 is down. You cannot see whether it is warm glycol, a pressure setting, a line logged as cleaned that was not, an untrained pour, or an unrung round. Four different problems, four different fixes, one identical symptom on the P&L.
No proof of what changed. Somebody visits, adjusts something, and the number improves the next month. Nobody can say which action did it, so it cannot be repeated at the other eleven sites.
Why the portfolio average hides it
Roll twelve locations into one number and drift disappears. A group averaging 8 percent variance can be six sites at 4 percent and six at 12, and the average will look stable for years while half the portfolio bleeds.
The spread is the signal. The average is the thing that stopped you from seeing it.
Draft drift is not a manager problem. It is a language problem.
Walk a drift conversation through a hospitality group and count the number of different numbers in the room.
The GM argues from keg counts and what the bar team told him. The Director of F&B argues from a monthly variance report that arrives three weeks late. The CFO argues from cost of goods as a percentage of revenue, which moves for a dozen reasons that have nothing to do with draft. Three people, three sources, three definitions of the same problem. The meeting ends in an opinion contest, and the person with the strongest relationship wins.
None of them are wrong. They have never had a number all three could stand on.
Shared, current pour data changes what the meeting is about. When every location measures the same thing at the same interval, the argument stops being whether Location 4 has a problem and starts being which of four causes it is. That is a faster conversation and a less political one. It also makes the GM an ally instead of a defendant, because for the first time he can show that his site's gap came from equipment and not from his team.
Draft efficiency is not one person's job. It is the keg room, the bar, the schedule, and the P&L, and it only improves when those four are looking at the same reading.
What closing the gap looks like
At Station 7, a Martin City Brewing Company location, unsold poured ounces ran $3,705.80 over seven days. After calibration, the same venue pouring the same volume ran $63.91 over seven days. $3,641 recovered in a week.
The recovery matters. What matters more for a group is that the cause was identified, the fix was recorded, and it can now be checked at every other location in the portfolio. One site's fix becomes the standard, not an anecdote.
The practical order
Establish one baseline. Same metric, same interval, every location. Poured ounces reconciled against sold ounces is the only version that survives scrutiny.
Rank the portfolio. Not to punish the bottom. To find the spread.
Attribute the gap. Temperature, pressure, pour behavior, or POS accuracy. Variance you cannot attribute is a blind spot, not a number.
Prove the fix. Record what changed and what moved after, then apply it across the group.
Drift is not the cost of running multiple locations. It is the cost of running them without a shared baseline.
PourScore gives every location in a group the same score on the same data, reconciled against POS. Schedule a demo.
