StarBucks

Group Project Report

SIE 462
Spring 2018
Table of Contents
(1) Introduction / Background 2

(2) Problem Description 2

(3) Current Practice and Related Issues 3

(4) Solutions 3

(5) Discussion and Conclusion 9

6) Bibliography 9

SIE 462 Final Report

Group 11 Page 1
(1) Introduction / Background

Company History
The first Starbucks was founded in 1971 in Seattle, Washington’s Pike Place Market and
has grown into the massive empire it is today through its mission to “inspire and nurture the
human spirit - one person, one cup, and one neighborhood at a time”. This first Starbucks only
served high quality beans - no coffee. By 1982, Starbucks expanded and began providing coffee
to fine restaurants and espresso bars. Inspired by his travels to Italy, director of retail Howard
Schultz mimicked the coffeehouse culture in Milan at the Starbucks in Seattle.
By 1989 there were 55 stores total and that number doubled by 1991. The first drive-thru
location opened in 1994. In the following five year, the company began serving the Frappuccino,
extended into grocery channels, launded Starbucks.com, and acquired Tazo Tea. By 2000, there
were over 3,500 stores worldwide. As of June 28, 2015, total stores have reached the 22,000s,
with stores in Panama, Colombia, Brazil, Egypt, and many more.

Company Overview
Starbucks prides itself in its quick, meaningful service and superior product. Starbucks
Company has been recognized by ​Fortune Magazine ​as one of the “World’s Most Admired
Companies” and by ​Forbes ​as one of the “World’s Most Valuable Brands” for several years
running. People walk into Starbucks and expect more than coffee. Starbucks around the globe
serves as neighborhood gathering places, and is a part of many daily routines. Coffee drinkers
are welcomed to expand their Starbucks experience with a selection of premium teas, pastries,
and other goodies.

(2) Problem Description

The forecasting method for the demand of hot and cold teas does not take into account
the seasonality aspect of the demand. The demand for hot teas is increased during the colder,
winter months, while demand for cold teas is higher during the warmer summer season. These
dynamic demands change throughout the year and create a problem for forecasting that doesn't
consider this seasonal change.
Further, the aggregate tea demand is also inaccurate due to hot and cold tea forecast
issues. By updating the forecasts for hot and cold teas individually, the aggregate tea demand
forecast will be more accurate and allow for more useful inventory analysis and workforce
planning.
By using Winter’s method to recalculate seasonal tea demands, this project will optimize
the tea forecasts and create more accurate planning and preparation for tea sales. This will lead to
shorter lines and increased customer satisfaction.

SIE 462 Final Report

Group 11 Page 2
(3) Current Practice and Related Issues
With the mission of trying to supply and run over 27 thousand stores worldwide,
Starbucks has to take a lot of variables into consideration, such as customers’ profiles and
demand, store location, and order management. For our project we will be focusing on the
customer’s demand to achieve a more accurate demand forecast of both hot and iced teas to help
the company better predict changes and future events. These data will help optimize the
operation process in the Starbucks store, which is the one located at the University of Arizona
bookstore, as well as increase customer satisfaction. In addition, the collected data will help us
put together an analysis on how many workers will be needed given the forecasted demand.

(4) Solutions
The team first wanted to make actual forecasts for the demand for the rest of the 2018
year because this is something that Starbucks currently lacks. Having a forecasted demand based
on previous data will allow for Starbucks to make informed decisions about employee count and
inventory levels. Because the demand information the team received was specific to Starbucks’
teas (hot and cold), this is where our demand forecasting was focused as well. Because hot and
cold teas are both seasonal drinks (fluctuate with the changing temperatures throughout the year),
the team used Winter’s Method to forecast demand. Because over the past few years the overall
temperature has been getting slightly hotter, Winter’s method with trend was used for this
problem. In this case, N=4 because of the four seasons of the year that the demand fluctuates
during. The data the team received for hot teas is shown below.

Figure 1: Problem Data

Years 1 and 2 are previous Starbucks data and Year 3 is the current 2018 year. The data
is given in number of cartons of tea Starbucks uses to make their teas, rather than the specific
individual customer ordered drink count. The first step is determining the year one and year two
demand averages, V​1 and V​2​. From here, the initial slope estimate (G​0​) and value of series at time
0 (S​0​) can be found. These calculations are shown below.
SIE 462 Final Report
Group 11 Page 3
 Equation 1: G​0

Equation 2: S​0

Figure 2: Initial Calculations

The next step is to initialize the seasonal factors (c​t​) using the following formula.

Equation 3: C​t

Figure 3: Initialize Seasonal Factors Calculations

These values are used to determine the average seasonal factors. Because, for example,
C​-7 and C​-3 both refer to the first quarter, then the average seasonal factor for quarter one is the
average of those values.

SIE 462 Final Report

Group 11 Page 4
 Equation 4: C​(-N-1)

Figure 4: Average Seasonal Factors Calculation

Then the team normalized the seasonal factors. We also ensured our math was correct by
performing the check that the sum of the C values should be equal to N (4).

Equation 5: C​j

Figure 5: Norming Factor Calculations

The team then made assumptions that ɑ = 0.2, β = 0.15, and ɣ = 0.1. From here we were able to
determine S​1​, G​1​, and c​1​.

SIE 462 Final Report

Group 11 Page 5
 Figure 6: S​1,​ G​1,​ C​1,​ Calculations

From here, the team forecasted the demand for hot teas for the remaining three quarters
of this year.

Figure 7: Forecasts

The team used the same method and formulas to make demand forecasts for the cold teas
for the remainder of 2018. All of the calculations are shown below.

Figure 8: Cold Tea Demand

SIE 462 Final Report

Group 11 Page 6
Figure 9: Jan-Mar Year 3 Demand

SIE 462 Final Report

Group 11 Page 7
 Figure 10: All Calculations

Figure 11: Tea Totals

Once the team forecasted the tea demand for the rest of the year, we could then determine
the work force needed to support this demand. Because there are about 8 tea servings per carton,
the team decided to assume that a barista can use up one tea carton per shift to make individual
drinks. This analysis was performed as if a barista was dedicated only to making teas for
customers and they aren’t cross trained. If a barista can produce enough tea to use up one tea
carton per shift, then a barista can use up 168 cartons per quarter. This value is used in the
units/worker section of the table. The rest of the assumptions and work is shown below.

SIE 462 Final Report

Group 11 Page 8
Assumptions:
● Current number of employees = 1/shift dedicated to teas
● Assumed: 8 teas consumed/carton
● Units produced/workers (in terms of tea cartons) = 1/shift = 14/week = 56/month
= 168/quarter

Figure 12: Summary

(5) Discussion and Conclusion

Based on our collected data and the calculations made, the team can now optimize the
Starbucks operation in trying to predict and fulfill the demands of cold and hot teas during
seasonal changes. In detail, we applied Winter’s method to achieve the forecasted demand of
both hot and cold teas. As the team predicted, the forecasts show that the hot tea demand peaks
in the winter months (October-December) with 227 tea cartons being consumed. The iced teas
are more popular in general throughout the year, but show a spike in demand in the summer
months (April-September) with 564 cartons being consumed. Overall, the demand seasonality of
the hot and iced teas seemed to somewhat balance out when combined because their seasonalities
are flipped.
Furthermore, we applied some calculations to find the best number of baristas needed to
operate the store after the seasonality change. Based on our results, we recommend that the
University of Arizona Bookstore Starbucks should have one extra person working to help fulfill
the demands for teas. Because baristas are cross trained at Starbucks to make all the drinks on the
menu, having two designated people total to make the teas for customers isn’t necessarily
plausible. Therefore, Starbucks can take this data and mold it to fit their employee structure
according to the percentage of their sales that consists of their teas compared to the rest of their
drink menu. Adding these baristas to help account for the teas alone will help decrease the long
lines at the University of Arizona bookstore location and in turn increase customer satisfaction.
If Starbucks applies this same ideology to the rest of their menu and determines the employee
count needed for each section of their drinks, they’d be able to optimize their whole workforce
and decrease their lines and wait times dramatically.

SIE 462 Final Report

Group 11 Page 9
6) Bibliography

“Company Information.” Starbucks, Starbucks Coffee Company,

www.starbucks.com/about-us/company-information​.

SIE 462 Final Report

Group 11 Page 10