Project Scenario

Welcome to this segment, where you will be introduced to the detailed project scenario and the
required data to make decisions in the described scenario.

So, let’s get started!

Scenario: You are a supply chain manager at a company called "TechHub Electronics," which
specializes in producing and distributing electronic gadgets. Your company offers a range of
products, including smartphones, laptops, and smartwatches. The market for these products is
highly competitive, with rapidly changing consumer preferences and technological advancements.

Now, let’s look at the previous year’s demand data for the company's various products.

Data Provided:

The table below shows the monthly units sold for each product category in the past year (Jan-Dec).

Months Smartphones Laptop Smartwatches

January 1500 800 300

February 1600 820 310

March 1800 900 320

April 1900 920 330

May 1700 850 310

June 2000 1000 350

July 2100 1050 360

August 2200 1100 370

September 1900 950 340

October 1800 900 320

November 2000 1000 350

December 2300 1150 380

This data is sensitive to recent changes. Based on this data, you will need to make certain decisions
for the company. So, let’s proceed to the next segment to see the key tasks and decisions you need
to make.
Project Tasks

Now that you are aware of the scenario and are equipped with the necessary data let’s go through
the five key tasks you need to perform as the supply chain manager for the company.

Task 1: Demand Forecasting Method

a. State and justify your choice of demand forecasting method if you had to forecast the demand for
the next six months (a short-term forecast). Also, explain at least one assumption made. Remember
that the data of demand provided is sensitive to changes.
(Word Limit: 200 words)

Task 2: Demand Forecast Calculation

a. Using the monthly demand data provided for the previous year, forecast the demand for each
product for the next six months using a suitable forecasting method (e.g., moving average, weighted
moving average, simple exponential smoothing).

Note: For any forecasting calculation purposes, you can consider α=0.3

Task 3: Calculating Forecast Errors

a. Consider the following table as the actual demand data for the company’s products after six
months.

Months Smartphones Laptop Smartwatches

January 1600 755 259

February 1649 865 292

March 1569 877 250

April 1750 923 214

May 1920 1000 200

June 2040 1100 196

Based on the demand forecasts you prepared in Task 1 and the actual data provided in the table
above, calculate the forecast errors for the six months.

Task 4: Metrics and Analysis

a. Based on the forecast and forecast errors you calculated in the previous tasks, calculate the
following key performance metrics:

1. Mean absolute percentage error (MAPE)

2. Bias

Task 5: Recommendations
 Analyze the forecast errors and metrics calculated so far and propose corrective actions in your
approach and strategy. Provide recommendations for improving the supply chain planning process
to handle demand fluctuations better, reduce lead times, and enhance overall supply chain
performance. Provide at least three recommendations (at least three; one for each product
category).
(Word Limit: 500 words)

The assessment and grading criteria are provided in the next segment.

Assessment and Grading Criteria

You will be graded on the following parameters in their respective weightage. Please go through
these detailed grading criteria carefully before working on your project submission.

Marks
Graded
Description (Total = 40
Component
marks)

 1 mark will be awarded for making a suitable choice of quantitative

forecasting method for calculating the demand forecasts for the next
six months.
Task 1  3 marks will be awarded for a well-structured explanation of why the 4 marks
chosen forecasting method is suitable. The explanation should
highlight the strengths, limitations, and assumptions of the chosen
method to demonstrate a thoughtful decision-making process.

 1 mark will be awarded for clearly stating the formula used in the
selected forecasting method.
Task 2
 9 marks will be awarded for accurately calculating the demand 10 marks
forecasts for each product type for the next six months based on the
selected quantitative forecasting method.

 1 mark will be awarded for clearly stating the formula used for
Task 3 calculating forecast errors.
10 marks
 9 marks will be awarded for accurately calculating the forecast errors
based on the provided ‘actuals’ data for the six months.

 5 marks will be awarded for correctly calculating the Mean absolute

Task 4 percentage error (MAPE) 10 marks
 5 marks will be awarded for correctly calculating the Bias

 6 marks will be awarded for a clear and sound explanation of how

your recommendations (at least three; one for each product
Task 5 6 marks
category) will help maintain a smooth flow of the operational activity
and not lead to a stockout or overstock situation.
Please note as you prepare to submit your projects, we want to emphasize the importance of
academic integrity and originality. Please be informed that all project submissions will undergo a
thorough plagiarism check. Plagiarism, including copying content from external sources without
proper attribution, undermines the authenticity of your work and the learning experience. We
encourage you to approach your projects with creativity, critical thinking, and a genuine
commitment to learning. When using external resources, remember to cite them appropriately to
give credit to the original authors.
Backup:

SES Strength/limitation:

SES is well-suited for short-term forecasting, especially when there are minimal trends or
seasonality. It adapts relatively quickly to short-term fluctuations in the data. However, reliability
diminishes as forecasting horizon extends beyond the near future, leading to greater uncertainty and
potential deviations from actual demand.

Cannot effectively deal with data exhibiting strong trends or consistent seasonal patterns, leading to
inaccurate forecasts.

1. SES often performs well for short-term forecasting horizons, especially when the data
exhibits no significant trends or seasonality. This makes it useful for planning inventory
levels, production schedules, or short-term marketing campaigns.

Strengths of Simple Exponential Smoothing (SES):

1. Simplicity and Ease of Use:

 One of the primary strengths of SES is its simplicity. It's easy to understand and
implement, making it accessible even for users with limited forecasting expertise.

2. Efficiency in Computing:

 SES doesn't require extensive computational resources. It's a lightweight method

that can be quickly applied, making it suitable for scenarios where efficiency and
speed are important.

3. Adaptability to Short-Term Changes:

 SES is well-suited for short-term forecasting, especially when there are minimal
trends or seasonality. It adapts relatively quickly to short-term fluctuations in the
data.

4. Real-time Adjustments:

 SES provides real-time adjustments to forecasts as new data becomes available. This
responsiveness to recent data is advantageous in situations where conditions can
change rapidly.

5. Flexibility with Smoothing Parameter (�α):

 The smoothing parameter (�α) allows users to control the level of responsiveness
to recent observations. This flexibility is valuable for customizing the forecasting
method based on the nature of the data.

6. Interpretability:

 The results of SES are easy to interpret. The forecasted values are influenced by
recent data, and the smoothing parameter provides a clear indication of the weight
given to these recent trends.

Limitations of Simple Exponential Smoothing (SES):

 1. Assumption of Stationarity:

 SES assumes that the time series data is stationary, meaning that the statistical
properties of the data do not change over time. In reality, demand patterns may
evolve, and trends or seasonality may emerge, violating the stationarity assumption.

2. Inability to Handle Seasonality:

 SES is not suitable for datasets with strong seasonal patterns, as it tends to place
equal weight on recent observations and may not capture recurring seasonal
variations effectively.

3. Lack of Trend and Seasonality Modeling:

 SES does not explicitly model trends or seasonality. It is designed for scenarios
where these components are minimal. In the presence of significant trends or
seasonality, more advanced methods may be required.

4. Dependency on Smoothing Parameter (�α):

 The performance of SES is sensitive to the choice of the smoothing parameter (�α).
Selecting an inappropriate value for �α can lead to suboptimal forecasts, and
determining the optimal value may require trial and error.

5. Equal Weighting for All Observations:

 SES assigns exponentially decreasing weights to past observations, but the weight
decreases exponentially without considering the time gap between observations. All
observations receive equal weighting based on their distance from the current
period.

6. Vulnerability to Noisy Data:

 SES can be sensitive to outliers and noisy data. It may place too much emphasis on
recent noisy observations, leading to inaccurate forecasts. Robustness to outliers is a
limitation when dealing with datasets that include irregular or extreme values.

7. Limited Forecasting Horizon:

 SES is typically used for short-term forecasting. Extending the forecasting horizon
too far into the future may result in less accurate predictions.

In summary, while SES has strengths such as simplicity, efficiency, and adaptability to short-term
changes, its limitations include assumptions about stationarity, the inability to handle seasonality
effectively, and sensitivity to parameter choices. Analysts should carefully assess the characteristics
of the data and the specific requirements of the forecasting task before choosing SES or consider
more advanced forecasting techniques for complex and dynamic demand patterns.

Strengths of Simple Exponential Smoothing (SES):

 Simplicity and ease of use: Requires minimal data preparation and only one parameter to
tune, making it accessible even for users without a strong statistical background.

 Adaptability to changes: Places more weight on recent observations, allowing it to adjust

quickly to shifts in demand patterns.
  Efficiency and minimal resource requirements: Computationally efficient and requires fewer
resources compared to more complex methods.

 Effectiveness for short-term forecasts: Performs well for short-term predictions, especially
when data has no significant trends or seasonality.

 Transparency and interpretability: Calculations are straightforward and easy to understand,

allowing users to interpret results and gain insights into the forecasting process.

Limitations of Simple Exponential Smoothing:

 Inability to capture trends and seasonality: Cannot effectively deal with data exhibiting
strong trends or consistent seasonal patterns, leading to inaccurate forecasts.

 Limited long-term accuracy: Reliability diminishes as forecasting horizon extends beyond the
near future, leading to greater uncertainty and potential deviations from actual demand.

 Single parameter dependence: Relies heavily on a single smoothing parameter (alpha),

whose choice can be challenging and impacts forecast accuracy significantly.

 Sensitivity to outliers: Outliers can disproportionately influence the smoothed level and
subsequent forecasts due to the exponential weighting scheme.

 Limited explanatory power: Focuses primarily on generating forecasts without providing

much insight into the underlying factors driving demand.

In summary, SES is a valuable tool for short-term forecasting of relatively stable demand data due to
its simplicity, adaptability, efficiency, and effectiveness. However, its limitations in capturing complex
patterns and providing long-term accuracy necessitate choosing it strategically and considering
alternative methods for more demanding forecasting needs.

1. Smartphones: The MAPE of smartphones is relatively low, indicating that the forecast is
generally accurate. However, the bias is negative for the product line, meaning that the
forecasts are underestimating the demand. This could lead to stockouts and lost sales.

 MAPE (Mean Absolute Percentage Error): 6.67%

 Bias: -6.67%

Recommendation: Improve Forecasting Accuracy with Collaborative Planning

 Enhance collaboration between sales, marketing, and supply chain teams to

incorporate market intelligence, sales forecasts, and customer feedback into the
forecasting process for smartphones. By leveraging cross-functional inputs, TechHub
Electronics can develop more accurate demand forecasts that reflect market
dynamics and consumer preferences. Additionally, consider implementing advanced
forecasting models and demand planning software to improve forecasting accuracy
further.
 2. Laptops: The MAPE of product line is relatively low, indicating that the forecast is generally
accurate. However, the bias is negative for the product line, meaning that the forecasts are
underestimating demand. This could lead to stockouts and lost sales.

 MAPE: 8.39%

 Bias: -8.39%

Recommendation: Implement Agile Supply Chain Practices

 Adopt agile supply chain practices such as lean manufacturing principles, flexible
production scheduling, and supplier diversification to reduce lead times and
enhance responsiveness to demand changes for laptops. By streamlining production
processes and optimizing inventory levels, TechHub Electronics can minimize the
impact of forecast errors and fluctuations in demand. Additionally, invest in
technologies like real-time monitoring and data analytics to identify bottlenecks and
optimize production efficiency.

3. Smartwatches: The MAPE for smartwatches is much higher than for the other two product
lines, indicating that the forecasts are much less accurate. The bias is also positive, meaning
that the forecasts are overestimating the demand. This could result in excess inventory
carrying costs.

 MAPE: -32.06% (Negative MAPE indicates over-forecasting)

 Bias: 32.06%

Recommendation: Utilize Demand Sensing and Segmentation

 Implement demand sensing techniques and customer segmentation strategies to

better understand demand patterns and tailor inventory management strategies for
smartwatches. Leverage data analytics and machine learning algorithms to identify
demand signals and segment customers based on purchasing behaviour,
preferences, and demographics. By aligning inventory levels with specific customer
segments and demand signals, TechHub Electronics can reduce over-forecasting,
minimize stockouts, and improve supply chain efficiency for smartwatches.

Conclusion: By implementing these recommendations tailored to each product category, TechHub

Electronics can improve forecast accuracy, reduce lead times, and enhance overall supply chain
performance, enabling better handling of demand fluctuations across smartphones, laptops, and
smartwatches.

● I first need to calculate the absolute percentage error (APE) for each month and product
line. The APE is calculated by dividing the absolute value of the difference between the
actual demand and the forecasted demand by the actual demand, and then multiplying by
100, I then calculate the MAPE for each product line by averaging the APEs for that product
line.
○ MAPE Smartphones: (6.25%+7.22%+2.68%+2.99%+11.54%+12.31%)/6= 7.16%
○ MAPE Laptops: (5.96%+6.82%+4.88%+6.83%+14.30%+18.19%)/6= 9.50%.
○ MAPE Smartwatches: (15.83%+3.77%+23.24%+47.04%+56.63%+65.45%)/6=
35.33%.