Sunday, March 8, 2009

Tavern Sales

Recently, the Brewers Association released its 2008 Craft Brewing Statistics and the National Restaurant Association (NRA) released its 2009 Restaurant Industry Forecast. Last week, I updated RePublic's business plan to reflect the new information. Diving back into the numbers got me thinking about whether or not my sales assumptions are reasonable.

Every year (I think), the NRA conducts a nationwide survey of restaurants and publishes the results in a document called the Restaurant Industry Operations Report. In the most recent edition, the median sales per seat for restaurants with average checks of $10-$14.99 was $9,859. I've been using that figure as the basis of RePublic's expected revenue, but one thing about the number bothers me: the median food sales reported by the NRA was 85.4% of total sales, but I'm only expecting food to account for 50% of RePublic's total sales (the NRA considers non-alcoholic beverages to be food, so 'beverage' equals 'alcohol only' throughout this discussion). Common sense suggests that bars without food will typically have far lower sales than $9,859 per seat. Unfortunately, I don't know of any published data that describes the sales of establishments with less than 70% food. I also suspect that sales per seat will drop for restaurants with no alcohol, which should cause a bar/restaurant sales prediction to look something like this:

I know that one former tavern in town sold about 50% food and generated around $5,500 per seat in annual sales. The place wasn't open seven days a week, though, and I wouldn't describe their operation as "busy". Returning to industry-wide data, the NRA provides five data points that I feel are useful (all for restaurants with average customer checks under $15):

-Median sales per seat (checks of $10-$14.99) = $9,859/yr
-Median food sales (checks of $10-$14.99) = 85.4%
-Median sales per seat, no booze = $8,345/yr
-Maximum reported beverage sales = 27.9%
-Maximum reported beverage sales per seat = $2,500/yr

Assuming the 27.9% and $2,500/yr figures are from the same establishment, the total sales per seat would be $8,961. That allows us to describe three scenarios based on NRA data:

-Sales per seat with 100% food = $8,345/yr.
-Sales per seat with 85.4% food = $9,859/yr.
-Sales per seat with 72.1% food = $8,961/yr.

Using that data, a linear graph can be drawn that looks like this:

It's not pretty, but at least it acknowledges that total revenue will depend on the food-to-beverage sales ratio. The result is that RePublic's projected sales per seat would drop from $9,859/yr to $7,469/yr. I'm aware that every market is different, but the raw NRA data should be in the ballpark. Comments from bar/restaurant owners on how I'm interpreting that data, or on sales per seat vs. food/beverage mix in general, would be most welcome!


justabrewer said...

Joe, isn't the profit margin on beer much higher than the profit margin on food? Especially when you are making the beer. I think profit per seat per year would be a more useful statistic.


Joe said...

Hey Tony, no, I'd much rather know sales per seat because profit per seat is based on a lot of factors that will change from business to business. It would be like replacing original gravity with alcohol content in a beer recipe. The NRA reports operating expenses as percentages of revenue as well, so I'd have to figure out revenue per seat either way. Calculating profit from revenue is a lot easier than calculating revenue from profit. If you want to see how I calculate profit, go to and download the 'Financial Model - Brewpub' spreadsheet.