Cutting Staff Turnover by 73 Percent Across Three Venues

The cost nobody had priced

Every operator knows turnover is expensive. Very few have a number for it, and the ones who do usually have a number that is far too small, because they are counting the recruitment cost and stopping there.

The recruitment cost is the cheapest part. The expensive parts are the six weeks the new hire is slower than the person they replaced, the higher void and comp rate while they learn the system, the overpouring while they learn the pour, the senior staff pulled off their own work to train, and the guest experience during all of it. Then a proportion of those hires leave inside ninety days and the whole cycle restarts.

A venue that turns over its bar team twice a year never gets far enough up the learning curve for anything to compound. It is permanently paying to train and permanently operating at the bottom of the curve.

Why people were actually leaving

The assumption going in was pay. It usually is, and it is usually not the whole story.

What the group found when they looked properly was that departures clustered in the first ninety days, and clustered around specific shifts and specific stations. That is not what a pay problem looks like. A pay problem produces steady attrition across tenure as people find better offers. This looked like people arriving, having a bad early experience, and leaving before they ever got comfortable.

Digging into it, the pattern was recognizable to anyone who has worked in a busy bar. New staff were being put on the hardest station on the busiest night because that was when the venue was short. They were being trained by whoever happened to be on, which meant the quality of their training was luck. They were making mistakes that were entirely predictable given the situation, being corrected in front of guests, and concluding they were not cut out for it.

None of that is a compensation problem. You cannot pay somebody enough to enjoy being set up to fail.

What LearnTAP changed

The training was rebuilt around how these venues actually run rather than around generic hospitality material.

• Standardized what had been informal. The pour, the close, the way an order gets corrected without a void. Previously this varied by whoever trained you, which meant a new hire's competence was determined by their trainer's shift rotation.
• Sequenced the stations. New staff stopped being dropped onto the hardest position on the busiest night. The first ninety days got a shape.
• Made the standard explicit. A lot of correction in bars is a manager reacting to a deviation from a standard that was never actually stated. Writing it down converts most of that friction into a straightforward answerable question.
• Connected training to the operational data. When CoreTAP flagged a high void rate on a particular person, the response could be a specific piece of training rather than a general conversation about being careful.

That last point is the one that matters most and is easiest to miss. Most training is delivered on a schedule to everyone. Training that responds to a measured, individual gap is a different intervention entirely, and it is far less insulting to the person receiving it.

Why turnover moved

Turnover fell 73 percent across the three venues. The mechanism was not that the work became easier or better paid. It was that new staff stopped failing in their first month.

People rarely leave a job because it is hard. They leave because they feel incompetent at it and cannot see that changing. A new bartender who is clearly improving each week will tolerate a great deal. One who is being corrected constantly for things nobody explained will leave for a job that pays the same, because at least there they will not feel stupid.

The compounding part

Lower turnover is worth more than the recruitment saving, because a team that stays gets good.

A bartender in their second year pours more accurately, upsells more naturally, handles a difficult table without a manager, and trains the next hire properly. None of that is available to a venue that never keeps anyone past month four. The group's variance improved and its guest feedback improved, and neither was targeted directly. They were downstream of having an experienced team for the first time.

What we are not claiming

The 73 percent figure belongs to these three venues over this period, and it followed a specific, sustained change in how people were trained and scheduled in their first ninety days. Installing training software and continuing to put new hires on the hardest station on a Saturday would not reproduce it.

Turnover is also genuinely a pay problem in some venues. If a venue is paying below market, this is not the case study for it.

Related

• Cost control and payback, where a lot of turnover cost actually shows up.
• Enterprise and multi-location, on moving a practice that works between venues.
• All case studies.