On: The curious surge of productivity in U.S. restaurants

[dataviz1000]

I linked the Cheesecake Factory 2022 report below. I asked chat-gpt for some analysis and didn't double check the math so take with grain of salt. When you take into account the largest costs for Cheesecake Factory is labor of which 50% is waitstaff and leases for prime locations, profits from on premises alcohol sales do not compare to profits from takeout.

If 11% of sales are alcohol at 20–25% of the menu price and 25% of sales are off premises at 28–35% of the menu price, what is the percentage profit from alcohol and off premises food and non-alcohol beverage, assuming they are mutually exclusive?

To break it down:

Alcohol:

Cost percentage: 20–25% Profit margin: 75–80% Profit contribution: 11% × 75–80% ≈ 8.25–8.8% of total profit

Off premises:

Cost percentage: 28–35% Profit margin: 65–72% Profit contribution: 25% × 65–72% ≈ 16.25–18% of total profit

[doctorpangloss → dataviz1000]

Cheesecake Factory isn’t representative of restaurant businesses. It’s like saying restaurants should be more like McDonalds, or startups should be more like Google.

[dataviz1000 → doctorpangloss]

I'm using extra data to add to the original study—which found that off-premise takeout and delivery increased from 50% to 60% after the pandemic leading to higher sales per employee—to show that extra costs like lease payments, insurance, wages for waitstaff, and utilities don’t fully offset the slightly higher profit margins from selling alcohol on premise compared to selling food and non-alcoholic drinks through takeout or delivery. Using financial data, I demonstrate higher off premises takeout or delivery leads to higher sales per employee -- the same conclusion the paper we are discussing made.

Notably, the paper shows that the share of visits lasting 11–20 minutes did not drop (and even increased a bit). This indicates that the overall shorter visit times aren’t because dine-in customers are being served faster. Instead, it’s mainly due to a rise in takeout and delivery orders (takeout customers are being served slower because there are more, if you follow.) In other words, even if people wait a bit longer for their takeout, the large number of takeout orders is what lowers the average time customers spend at the restaurant which corresponds to the higher sales per employee. (If you disagree with this interpretation, please refer to the study, as that is its main argument.)

My point is that even though alcohol generally has a higher markup, the extra overhead needed to sell alcohol on premise means that its profit margins do not fully make up for those costs compared to off-premise takeout or delivery. Moreover, the data from The Cheesecake Factory, ~25% off premises takeout or deliver sale, is similar to what is seen in any traditionally dine-in U.S. restaurant, corporate or mom and pop, that isn’t Michelin starred.