Forecasting Sales at Ska Brewing Company
Case Study 1
EBTM 350, Spring 2020
Background[footnoteRef:1] [1: This case is adapted from Drake, M. (2014). The applied business analytics casebook: Applications in supply chain management, operations management, and operations research. Upper Saddle River: Pearson]
Ska Brewing Company is a purveyor of fine craft beers located in Durango, Colorado. With its flagships Pinstripe Red Ale and True Blonde Ale, medal winning Buster Nut Brown Ale and Steel Toe Stout, and seasonal Mexican Logger and Euphoria Pale Ale, Ska has enjoyed double-digit growth for more than a decade with no signs of slowing down. Learn more about Ska by visiting its website: http://www.skabrewing.com.
In the early ‘90s founders/owners Dave and Bill were dissatisfied with watered-down corporate beer and decided to take matters into their own hands, literally. They begin brewing their own beer in their basement, much to the delight of everyone who knew them. Eventually, it became clear that they might be able to make a living doing what they loved to do, and they founded Ska Brewing Company in 1995 with third owner/founder Matt. Through hard work and a laser-like focus on brewing great beer, Ska continued to grow, and in 2008 the company moved into its $4.8 million, 24,000-square-foot world headquarters. In 2012, Ska brewed more than 25,000 barrels of beer (1 barrel = 2 standard kegs = 252 pints = 4,032 ounces), with sales exceeding $6.5 million.
Ska was not alone in its success. Durango, a town with fewer than 20,000 people, has four long-term successful breweries/brewpubs, a brand new brewpub that opened in 2012, and another one in the works. Rather than considering these other breweries as competition, Ska has worked together with them (as well as others across the state of Colorado) to brew specialty beers for festivals and other occasions; Ska also contract brews beer for Steamworks Brewing Company (using its recipes) because Steamworks has exceeded its own brewing capacity. Owner Dave calls this unique relationship “coopitition.” Steamworks and Ska are just examples, however.
The craft brewing industry has seen phenomenal growth during the last three decades across the United States and in other countries as well. According to the Brewers Association, the craft-brewing renaissance started in the late 1970s and saw periods of incredible growth during the 1990s. Historically, before Prohibition, small breweries were everywhere across the United States. The 18th Amendment caused most of the small breweries to go out of business, and only the larger breweries survived until the 21st Amendment repealed Prohibitions 13 years later. It took several decades for smaller breweries to begin the resurgence that we see today.
Our concern is more specific: Will the growth and success at Ska continue? Can Ska anticipate how much beer it will produce, and what sales will be so that the company can plan wisely for the future? In fact, current plans are to increase brewing capacity yet again - a cost investment with potentially high returns. Is this a good decision or not? This is where you come in.
Mission
Despite its success, Ska is still relatively a small operation. The company has one main numbers person, accountant Erik. In a nutshell, Erik would like to predict Ska’s sales dollars and barrels sold for the current year, 2013. He has done some of this work on his own, but he would like you to confirm (or refute) his forecasts, and to do so in much greater detail, as Erik is too busy. Ska’s total barrels (BBLS) sold and sales ($$$) over the previous 13 years can be found in the Ska Annual Data file. Please note that this is actual (not phony textbook) data.
Even a cursory glance of the information in the table shows that both the number of barrels and sales are increasing annually at a pretty good rate. In fact, both values have shown tenfold growth between the years 2000 and 2012. What will these two numbers look like at the end of 2013? Ideally you learned that when forecasting real data, there is no “one-size-fits-all” approach. Ahead you will try several approaches and then combine them to make a final prediction.
Your task is not only to forecast these two values for 2013, but to give Erik, Dave, Bill, and Matt a better picture of what is happening with their business overall. To do so, you will be asked to produce several graphs and perform relevant analyses. You will first be asked to learn a little more about the brewing industry in general to give you a better idea of the current status of craft brewing. Your final report should be thorough, professional, and accurate. Good luck!
Analysis
Part A
2.1 Use Microsoft Excel to draw scatter plots of both year versus barrels and year versus sales. (Hint: You might want to change the year range from 2000-2012 to 1-13 to simplify the equations of the curves that Excel will eventually fit to the data.) Consider the barrels data first; then repeat for the sales data.
2.1.1 Have Excel fit a linear trendline to the data and determine the equations of the line and the value. Interpret the slope of the line and the coefficient of determination. Is the linear trend line a good fit for the data - explain?
2.1.2 Now have Excel fit an exponential curve to the data and again determine the equations of the line and the value. Is this a good fit - explain?
2.1.3 Using the best fit equation, plug in 14 (for 2013) to get your first forecast. Does it seem reasonable, too low, or too high?
2.1.4 Make sure you solve questions 2.1.1 through 2.1.3 for barrels as the response, and then again with sales as the response.
2.2 Now draw a scatter diagram of barrels versus sales. Fit a linear trendline to the data and interpret both the slope of the line (hint: 1 barrel = 2 kegs) and the coefficient of determination. Can you reasonably conclude the more beer Ska produces, the more money it makes?
2.3 Reconsider the graphs from question 2.1. Consider the last four points on each graph from 2009 to 2012. Ignoring the rest the data, do those four points appear to have an (obvious) pattern? Explain.
2.3.1 Using only the last four years’ data, fit a line for both barrels and for sales. Interpret both the slope and r2 value for each line.
2.3.2 Plug 5 (for the next period which is 2013) into each equation to get your second forecast for barrels and sales in 2013. How confident are you with these predictions?
2.4 For both barrels and sales, determine the MAPE for each of your predictions and explain how MAPE can be used to evaluate forecasts.
2.5 Consider one more way to forecast barrels and sales for 2013 before you make your final decision. Determine the percentage growth for both barrels and sales for each year. For example, from 2000 to 2001, barrels increased from 2,595 to 3,025, or a growth rate of (3025-2595)/2595 = 17%. Calculate these rates for years 1 to 12 for both columns of data.
2.5.1 Determine the average growth rates for both barrels and sales.
2.5.2 Considering only sales, draw a scatter plot of year versus sales growth. Do any of the growth rates look like outliers? Discuss. (Hint: Recall that Ska moved into its new world headquarters in 2008, increasing its brewing capacity tremendously.)
2.5.3 As a side note, Erik, the accountant, asked the owners to do a quick, back-of-the-beer-coaster estimate of what growth would be for 2013. Their immediate response was “20%.” Would you say that Dave, Bill, and Matt are guessing, or do they know their business very well?
Part B
Congratulations, you have just completed a very thorough analysis of Ska Brewing Company’s production in barrels and sales figures. At this point, it might be worth reconsidering how accurate forecasts will help Ska. According to Erik, “an accurate sales budget is the root of the entire budgeting process.” In addition, Dave says that accurate forecasts would help “tremendously,” allowing Ska to “increase efficiencies from a production standpoint” and help “make decisions about whether or not Ska could enter any new markets.”
Now it’s time to tie everything together and make you best forecast for 2013 for both barrels and sales. Carefully combine your forecasts. What is your final forecast? How do you justify that prediction?