Unveiling Patterns in Forecasting Errors: A Case Study of 3PL Logistics in Pharmaceutical and Appliance Sectors

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript has improved since I saw it for the first time. It is generally OK.
I have a few minor comments:
- some tables use incorrect decimal symbols, commas instead of points (Table 7,8,9).
- in l.389,392,395 authors say that a more significant trend component entails a more significant seasonal component etc. This is not correct, as this should not be a statement on significant. Indeed, a larger trend component entails a larger seasonal component etc. This is correct on l.395, where authors use "greater".
- In this context, I would also avoid "significant" on l.329, "important" etc. could be a good alternative and less confusing with a statistical evaluation that is not implied here.

Author Response

Dear Reviewer,

Thank you for your review and helpful comments. We have addressed all your suggestions and made the necessary changes to the manuscript. Your feedback has been very useful in improving the paper.

1. The decimal separator has been corrected in all tables. Dot instead of comma.
2. The wording in lines 389-395 and 329 has been improved. Statements of statistical significance should be more formally justified.

Reviewer 2 Report

Comments and Suggestions for Authors

The article submitted for review presents an analysis of forecast errors in the case of using ARIMA models for time series. The editing layer of the text and all tables and titles of chapters and subchapters should definitely be improved (different colours, inadequate numbering). In the Conclusion, I propose adding a paragraph stating whether the conducted study fits the results obtained in the literature or proves something new.

Author Response

Dear Reviewer,

Thank you for your review and helpful comments. We have addressed all your suggestions and made the necessary changes to the manuscript. Your feedback has been very useful in improving the paper.

1. Formatting has been improved (color, decimal separators...)
2. As suggested, the work was supplemented with a paragraph referring to how the results fit into the studied literature on the subject.

Reviewer 3 Report

Comments and Suggestions for Authors

Thank you for a good piece of research. My comments are as follows:

Title: Emphasize the study’s novelty and insights, e.g., "Unveiling Patterns in Forecasting Errors: A Case Study of 3PL Logistics in Pharmaceutical and Appliance Sectors."

Line 8-11: "The research focused on ten distribution channels served by a 3PL logistics operator utilizing the Google Cloud AI forecasting tool as part of the Google Cloud AI service." The sentence is overly technical. Consider simplifying: "This study examines forecasting errors across ten distribution channels managed by a 3PL operator using Google Cloud AI forecasting."

"Statistical tests for various channels show significant differences in forecast error groups in some cases, suggesting that the forecasting tool may perform more accurately for certain channels than others."  The phrase "in some cases" is vague. Specify which channels or types of channels.

Line 25-27: "The findings underscore the need for a tailored forecasting approach for each channel..."  This is an important insight. Briefly mention how tailoring forecasting tools can impact operational efficiency.

Line 35-38: "Accurate forecasting plays a pivotal role in this process, enabling organizations to make informed decisions, optimize inventory levels, and meet customer demands effectively." This statement is generic. Enhance by adding specific challenges faced by 3PL operators, e.g., "Accurate forecasting is crucial for mitigating supply chain disruptions and managing inventory costs in dynamic markets."

Line 45-46: "Understanding the hidden patterns and behaviors within forecast errors for each channel is essential to improving the predictive capabilities of these models." The term "hidden patterns" is ambiguous. Specify the types of patterns, such as seasonality or trends.

Line 47-51: "The purpose of this study is to conduct a comprehensive analysis of time series forecast errors generated by a 3PL logistics operator across ten distribution channels it manages." This is clear but lacks emphasis on novelty. Highlight why focusing on time series errors in 3PL is unique or impactful.

Line 59-63: "Forecasts are critical inputs for decision-making in procurement, production, delivery, and inventory management." This is repetitive from the introduction. Instead, provide examples of specific forecast types (e.g., demand, inventory, or transportation forecasts).

Line 76-80: "Popular algorithms for demand forecasting in logistics flows include ARIMA-based models, machine learning, and neural networks." This list is useful but lacks depth. Briefly explain why these models are popular and how they address specific logistics challenges.

Line 103-106: "The centralization concept posits that a logistics operator, equipped with the necessary attributes, can assume centralized forecasting functions in a distribution network."  This point is significant but needs elaboration. Discuss specific benefits (e.g., reduced forecasting errors, streamlined operations) and challenges (e.g., reliance on data integration).

Line 150-156: "By employing these metrics, the study evaluates the performance of the forecasting system and identifies areas for potential improvement, especially in the context of optimizing the tool's calibration." Provide specific examples of how these metrics (e.g., MAE, MAPE) guide improvement.

Line 185-190: "This is a logistics company specializing in providing services related to the distribution and warehousing of goods for various enterprises." This is generic. Include specific services offered by the company and their relevance to forecasting.

Line 217-223: "Forecasts in this context were created with a 30-day horizon, with daily data updates in daily granularity." Explain why a 30-day horizon is significant for the study and how it aligns with industry practices.

Line 275-278: "The visual analysis of time series forecast errors involves plotting these errors on a timeline." This description is generic. Include details on key patterns observed, such as cyclicality or anomalies, to make the analysis more specific.

Line 341-348: "The analysis of the randomness of forecast errors indicates that each of the analyzed series can be considered random..." Clarify what "randomness" implies for forecasting model performance and why it matters for operational strategies.

Line 454-466: "High values for the mean, standard deviation, coefficient of variation, and skewness suggest variability of errors relative to the mean." While true, this statement needs more actionable insights. Explain how these findings can guide adjustments to the forecasting tool.

Line 475-479: "The second hypothesis was not positively verified. However, the authors suggest there would be a high chance of its verification if detailed insights into the models used for forecasts were available." This is speculative. Recommend a more specific approach, such as collaborating with software providers to enhance model transparency.

Line 529-533: "Additionally, the lack of information about the forecasting models, calibration parameters, and input data can limit the full understanding of error sources." Discuss specific strategies to overcome these limitations, such as using interpretable machine learning models or conducting sensitivity analyses.

Line 538-540: "Forecast error analysis can inspire further research on specific channels, product types, or seasonality." This is vague. Suggest concrete future research areas, such as the impact of external factors (e.g., market shocks, policy changes) on forecasting accuracy.

Line 559-563: "The results of the forecast error analysis clearly demonstrated the critical role of error analysis in improving forecasting models." This is repetitive. Emphasize key actionable outcomes, such as specific improvements in operational planning or resource allocation.

Line 566-568: "The analysis represents a significant contribution to the theory and practice of demand forecasting by logistics operators in distribution networks." Strengthen this statement by summarizing practical contributions, such as cost savings or enhanced decision-making frameworks.
-Overuse of technical terms without adequate explanation.
-Limited connection between statistical results and practical logistics implications.
-Lack of concrete recommendations based on findings.

Author Response

Dear Reviewer,

Thank you for your thorough review and constructive comments. We have carefully incorporated all your suggestions and made the necessary changes to the manuscript. Your feedback has been invaluable in improving the quality of our work, and we especially appreciate your proposal to revise the title, which we found particularly insightful.

1. The title has been modified
2. The sentence "Statistical tests for various channels show significant differences in forecast error groups in some cases, suggesting that the forecasting tool may perform more accurately for certain channels than others." was removed because the precise results were presented later in the abstract

Suggestions for changing phrases in the text reference were accepted. Lines 8-10, 25-27, 35-38, 45-46, 47-51, 59-63, 76-80, 103-106, 150-156, 185-190, 217-223, 275-278, 341-348, 454-466, 475-479, 529-533, 538-540, 559-563, 566-568.

Reviewer 4 Report

Comments and Suggestions for Authors

The paper deals with Analyzing Forecasting Errors in 3PL Logistics through a case study. The paper is good, however before publishing the following should be addressed first.

·        Introduction should be extended

·        At the end of the introduction a methodology paragraph should be added (what each section consists of)

·        A figure explaining the methodology of the paper should be added

·        Is it possible to conduct sensitivity analysis or model validation and compare the obtained results with results obtained after implementing similar methods?

·        Theoretical and Managerial Implications should be added

·        Comparison with similar models should be added (explaining why this method is better than others)

 

·        The paper generally lacks a focus on sustainability, so it is necessary to better align the paper with the journal's thematic scope.

Author Response

Dear Reviewer,

Thank you for your review and helpful comments. We have addressed your suggestions and made the necessary changes to the manuscript. Your feedback has been very useful in improving the paper.

---

„·        Introduction should be extended“

The introduction has been expanded.

„·        At the end of the introduction a methodology paragraph should be added (what each section consists of)“

The paragraph has been added.

„·        A figure explaining the methodology of the paper should be added“

A figure illustrating the analytical methodology for detecting patterns in error series has been added. The methodology is described in detail in a separate section, including figures.

„·        Is it possible to conduct sensitivity analysis or model validation and compare the obtained results with results obtained after implementing similar methods?“ &

„·        Comparison with similar models should be added (explaining why this method is better than others)“

Sensitivity analysis and model validation are not feasible in our study due to the nature of the data and the forecasting system employed. Our analysis relies on secondary data derived from the forecasting tool used by the logistics operator. This tool operates as a black box, utilizing artificial intelligence mechanisms, and the adjustments or updates made to the system pertain to different time periods. Therefore, it is not possible to conduct research within the same time frame using a recalibrated forecasting system.

Consequently, a direct comparison of results with similar forecasting methods is also not viable. However, we find the reviewer’s comments inspiring for future research. In particular, we see potential in generating diverse forecasting scenarios using different methods and conducting ex post analyses to compare the forecast quality under various system settings. This approach could provide valuable insights and address the limitations mentioned above.

• Theoretical and Managerial Implications should be added

It has been added to discussion as a new section.

The paper generally lacks a focus on sustainability, so it is necessary to better align the paper with the journal's thematic scope.

A section has been added in discussion.

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

Thank you for the revision. The manuscript has been substantially improved, and I have no reservations at this stage. 

Good luck with your publication.

Reviewer 4 Report

Comments and Suggestions for Authors

Everything is ok.