TABLE OF CONTENT

CASE 1: FORECASTING TECHNIQUES................................................................................................................... 2

 Introduction................................................................................................................................................ 2
 Analysis:......................................................................................................................................................... 3
 Improvements............................................................................................................................................ 6
CASE 2: QUALITY MANAGEMENT.......................................................................................................................... 7
Reference........................................................................................................................................................................... 9
Appendices..................................................................................................................................................................... 11

ABBREVIATION LIST

ABBREVIATION

Honey Baking Company HBC

Moving Average MA

Weighted Moving Average WMA

Exponential Smoothing ES

Linear Trend LT

Kitty Tools KT
 CASE 1: FORECASTING TECHNIQUES

Introduction

Forecasting is a critical organizational function, anticipating future events with the

purpose of effective planning and decision making (Lawless 1990). Various studies
(Gronroos 1996, Mentzer-&-Moon 2004; Smith 2010) found connection between
forecasting and operational performance. In HBC, demand forecasts are conducted to
estimate production quantities and minimize the daily amount of stock-outs and
leftovers (Mou 2018).

There are two primary categories: qualitative and quantitative methodologies. When
data is scare, qualitative techniques are employed to make judgments (Chambers 1971).
Meanwhile, statistical techniques are adopted when data is available with clear and
consistent trends (Macnulty 1979). Furthermore, statistical is superior to judgemental
methods regarding accuracy and timeliness dimensions (Hogarth-&-Makridakis 1981).
Based on the given four-week data, this paper analyze four quantitative techniques’
performances (MA, WMA, ES, and LT) to select the optimal method and recommend
some improvements.

Analysis:
Table 1 shows the daily Week 5 forecast using MA and WMA (Appendix A):

MA offers a smoothed trendline that is less likely to fluctuate due to slight, temporary
swings (Maverick-2021), however, it is slower to react to rapid changes; making it
preferable for longer time operations. Meanwhile, WMA’s greater sensitivity suffers
from being more susceptible to false signals (George-1965); hence it is commonly
utilized over shorter time periods.
Table 2 reveals results of using ES formula and LT equation (Appendix A):

ES, a widely used business technique (Stevenson-2018) is ideal for production and
inventory as it considers most recent data. By neglecting ups and downs, ES shows
underlying phenomenon; however, gives projections lagging behind the actual trend
(Gardner-2006). Simply said, ES is best used in the absence of outliners (e.g. cyclical or
seasonal variations).

LT imposes a line of best fit to time series historical data (Monczka-et-al., 2016). LT
shows more accuracy and relevance to short-term changes method as it only uses
historical data without additional inputs (ES’s smoothing constant). However, the
limitation of LT is outlier sensitivity.
Increased accuracy in production and inventory management results in lower safety
stocks (Bala-2012); therefore, the manager must carefully assess the accuracy of each
forecasting method. To select the most accurate method, error measures (MAD, MSE,
MAPE) will be employed. Lower values explain more precise result.

Regarding MA and WMA, due to the lack of data, HBC is unable to validate the accuracy
of predicting outcomes, hence they appear to be ineffective. More historical data is
required to assist in forecasting to get reliable findings, which is costly and time-
consuming to obtain (Chambers, Mullick-&-Smith 1971).

Figures 1 and 2 exhibit that LT has lower MAD, MSE, and MAPE compared to ES;
therefore, it is recommended that HBC should select LT to forecast demand.
From Figure 3, LT trendline follows the real demand more closely compare to ES,
implying that LT delivers a more accurate prediction.

Aside from the accuracy, analysis cost, relevance, and historical data availability,
managers must consider industry characteristics to decide on the best appropriate
method (Stevenson-2018). Doughnuts are perishable food with one-day shelf-life,
therefore, it is advised to apply short-term planning approaches that are economical,
simple to compute, and react to demand changes while remaining accurate enough to
avoid backorder costs. After weighing benefits and drawbacks, the optimal option for
HBC is Linear Trend.

Improvements
The demand fluctuated over the period with seasonal variations, making it harder to
forecast trends (Figure 3). Particularly, the demand always peaks on Sunday and trough
on Thursday. To improve forecasting results of ES and LT techniques, seasonalized
forecast is applied (Calculating steps in Appendix B).
While seasonality forecast improves both ES and LT techniques’ accuracy (Appendix
B), LT follows the actual demand more closely than ES, thus considered the best option
for HBC.

CASE 2: QUALITY MANAGEMENT

Quality is a set of intrinsic features meeting customer expectations (IOS 2005). Since the
20th century, production volumes have been increasingly growing, necessitating
inspection methods into the manufacturing process to enhance and maintain quality
(Walkowitz-1989). O’Neill (2016) also stated that firms prioritizing quality
management have considerable financial performance and survival advantages over
those not engaging in. Before officially launching, KT inspected new product’s quality
with two sample sets. This study summarizes findings and interprets issues of
inspection.

In Set 1, we obtained 15 samples with 20 observations (n = 20) so that A2 = 0.18

(Appendix C). From the given data, the average of ranges is also calculated (R = 0.87).

Apply those values to the formula, we have:

To explain the incapable process’s conclusion, we calculate the Process Capacity:

A process requires a minimum capability index of 1.00 to be capable; however, with an
index of 1.00 to 1.33, it cannot be completely capable (Stevenson-2018). The capability
index is 1.08, indicating that the process is not entirely capable.

Set 2 uses a greater number of samples, which better reflect the population and
minimize the probability of missing out-of-control samples, resulting in more accurate
results (Khoo-2013). According to Haridy (2017), the fewer the number of observations,
the more accurate the control charts will be. Furthermore, significant observations
leads to less accurate standard deviation (Oakland-2019). Stevenson (2017) also noted
that with lower n, the sample distribution will be wider, implying standard deviation
charts will be more efficient and less dispersed.

The trend in Set 2 is described more clearly. Although the data is still under control,
data’s extremities are more apparent. Set 2 displays a cyclic pattern, which Set 1 does
not.
If the problem can be detected and solved, the standard deviation would be smaller,
resulting in higher process capability index. From Set 2, we calculate the average of
ranges (R = 0.40). Since n = 5, A2 = 0.58 (Appendix C). Apply the two similar formula
used for Set 1, we have:

The capability index of Set 2 is 1.48, greater than 1.33, therefore, the process is capable.

According to Fugard-&-Potts (2015), sample size should be determined by the amount

of data being analyzed. Large sample sizes are essential for large-scale studies (Adcock-
2002) as the sampling distribution is normally distributed, resulting in more accurate
population findings. Meanwhile, small sample sizes with high standard deviation might
lead to observed values not always representative of actual outcomes (Faber-2014),
prompting inspectors to accept false results. Furthermore, the inspection can waste
resources and lead to possible controlling problems, which costs higher to fix (Fugard
2015). However, regarding small population, small sample sizes are appropriate since
they are cost-effective and require little time or effort to execute while providing
reliable information (Grous-2013).

KT limited the budget for the process and replacing an oven would be costly, indicating
using more samples is not optimal and a small sample size is more preferable in this
situation.
Reference

Adcock, CJ 2002, ‘Sample size determination: a review’, Journal of Royal Statistical

Society, vol. 46, no. 2, pp. 261-283.

Bala, P.K. 2012, ‘Improving inventory performance with clustering based demand
forecasts’, Journal of Modelling in Management, Vol. 7 No. 1, pp. 23-37.

Chambers, JC, Mullick, SK & Smith, DD 1971, ‘How to choose the right forecasting
technique’, Harvard business review, vol. 49, no. 4, p. 45–.

Chase, CW 2013, Demand-Driven Forecasting: A Structured Approach to Forecasting,

John Wiley & Sons, Incorporated, viewed 12 March 2022, ProQuest Central Database.

Faber, J & Fonseca, L 2014, ‘How sample size influences research outcomes’, Dental
Press Journal of Orthodontics, vol. 19, no. 4, pp. 27-29

Fugard, AJ & Potts, HW 2015, ‘Supporting thinking on sample sizes for thematic
analyses: a quantitative tool’, International journal of social research methodology, vol.
18, no. 6, pp. 669–684.

Gardner, ES 2006, ‘Exponential smoothing: The state of the art—Part II’, International
journal of forecasting, vol. 22, no. 4, pp. 637–666.

George, SF 1965, ‘Smoothing, Forecasting and Prediction of Discrete Time Series’,

Operations Research, vol. 13, no. 2, pp. 328–330.

Gronroos, C 1996, ‘Relationship marketing: strategic and tactical implications’,

Management decision, vol. 34, no. 3, pp. 5–14.

Grous, A 2013, Applied Quality Control, John Wiley & Sons, Wiley Online Library
database.

International Organization for Standardization 2005, ‘Quality Management Systems -

Fundamentals and Vocabulary’, International Organization for Standardization, pp. 1-41.

Hogarth, RM & Makridakis, S 1981, ‘Forecasting and Planning: An Evaluation’,

Management science, vol. 27, no. 2, pp. 115–138.

Khoo, MBC, Lee, MH, Teoh, WL, Liew, JY & Teh, SY 2013, ‘The effects of parameter
estimation on minimizing the in-control average sample size for the double sampling X
bar chart’, South African Journal of industrial engineering, vol. 24, no. 3, pp. 58–67.

Lawless, MJ 1990, ‘Effective Sales Forecasting - A Management Tool’, The journal of

business forecasting, vol. 9, no. 1, p. 2–.

Macnulty, CAR 1979, ‘Long range forecasting: From crystal ball to computer’, Long
range planning, vol. 12, no. 4, p. 95–.
Maverick, IB 2021, Advantages and Disadvantages of the Simple Moving Average (SMA)?,
Investopedia, viewed 18 March 2022,
<https://www.investopedia.com/ask/answers/013015/what-are-main-advantages-
and-disadvantages-using-simple-moving-average-sma.asp#:~:text=The%20Bottom
%20Line,might%20be%20old%20or%20stale>.

Mentzer, JT & Moon, MA 2004, Sales Forecasting Management: A Demand Management

Approach, 2nd edn, SAGE Publications, Thousand Oaks.

Monczka, RM, Monczka, RM, Handfield, RB, Giunipero, LC & Patterson, JL 2016,
Purchasing and supply chain management, Sixth edition., Cengage Learning, Boston, MA.

Mou, S, Robb, DJ & DeHoratius, N 2018, ‘Retail store operations: Literature review and
research directions’, European journal of operational research, vol. 265, no. 2, pp. 399–
422.

O’Neill, P, Sohal, A & Teng, CW 2016, ‘Quality management approaches and their impact
on firms‫ ׳‬financial performance – An Australian study’, International journal of
production economics, vol. 171, pp. 381–393.

Sanders, N.R. (1995), "Managing the forecasting function", Industrial Management &
Data Systems, Vol. 95 No. 4, pp. 12-18

Smith, CD & Mentzer, JT 2010, ‘User influence on the relationship between forecast
accuracy, application and logistics performance, Journal of business logistics, vol. 31, no.
1, pp. 159–177.

Stevenson, W 2018, Operations Management, McGraw-Hill Education, ProQuest Ebook

Central database

Walkowitz, DJ 1989, ‘The Fall of the House of Labor: The workplace, the state, and
American labor activism, 1865–1925.’, Journal of Social History, vol. 23, no. 1, pp. 203–
206.
Appendices
Appendix A. Four Techniques - Calculating steps

A.1 Moving Average Formula

∑ A t−i A t−n +… At −2+ A t −1

i=1
F t=MA n = =
n n
where:
F t = Forecast for time period t
MA n = n period moving average
At −1 = Actual value in period t - 1
n = Number of periods in the MA

A.2 Weighted Average Formula

F t=w1 ( A¿¿ t)+ wt −1 ( At −1 ) +…+w t−n ( A t−n )¿

where
w t−1 = Weight for period t - 1, ect.
At −1 = Actual value for period t – 1, ect.

A.3 Exponential Smoothing

F t=F t −1 +α ( At −1−Ft −1)

where
F t = Forecast for period t
F t−1 = Forecast for the previous period
α = Smoothign constant
At −1 = Actual demand or sales from the previous period

A.4 Linear Trend

F t=a+bt
where
F t = Forecast for period t
a = Value for F t at t = 0
b = Slope of the line

Applying this formula, we can calculate a Linear Trend Equation for HBC:
 F t=2094.57+ 22.34 t

Appendix B. Seasonality forecast

Step 1: Computing the relatives of each day of the week

Step 2: Deseasonalize the actual demand by dividing the 24-day data with the relatives

Step 3: Seasonalize forecasting data (using ES and LT) by multiplying them with the
relatives

Deseasonalize the demand and add seasonality to estimate the Week 5 data
using Exponential Smoothing technique
 Deseasonalize the demand and add seasonality to estimate the Week 5 data
using Linear Trend technique

As can be seen from the graph, in comparison to Linear Trend, Sasonalized forecast
results are roughly similar to the demand in four weeks. In other words, seasonalized
strategy improves the accuracy of demand forecasting in this case.
Similarly, seasonalized forecast also enhance the accuracy of ES technique.

Appendix C. Factors for R Chart