Amazon Project Analytics Case Study: Data-Driven Insights

Introduction

Overview of the Organisation and Project

• Amazon is a globally recognised e-commerce platform with its market reaching in numerous products and services, customer responsiveness, and a data-driven nature.
• Amazon is facing intense competition in the Indian market from rivals such as Flipkart and Reliance JioMart (Das et al., 2023).
• Amazon pays extensive attention to collecting project analytics and customer data as a basis to establish pricing, product classification, and promotion policies in order to remain competitive.

In this study, the project under analysis one is to explore the data on the customer product reviews on computer and electronic accessories that are available on Amazon India. The data contains information on more than 1,400 individual product entries, including such details as the original price, discounted price, discount percentage, customer ratings, number of ratings, and text of reviews. Thereon, this data is used to come up with an assessment of how analytics can aid pricing optimisation, customer satisfaction, orders, and product refinement.

At Native Assignment Help, students can explore academic guidance on project analytics, business intelligence, and data-driven decision-making through practical case studies. Our Assignment Helper UK service supports students in understanding analytical methods, interpreting datasets, and presenting evidence-based recommendations in line with UK academic requirements.

Justification for Dataset and Case Study Selection

• The choice of this dataset is related to the fact that it matches project analytics goals quite well.
• It presents highly detailed, real-world customer interaction information that can be deployed to monetise descriptive and predictive behaviours (Schuurman, 2025).
• A great variety of product categories and usage of structured (pricing, ratings) and unstructured (review text) data means that a wide range of analytics methods can be applied.

With Amazon's known use of artificial intelligence (AI) and machine learning (ML) in their recommendation systems, pricing algorithms, and sentiment analysis systems, this dataset offers an opportunity to replicate and critically analyse the same analytical methods. Besides all this, the results may be used to develop wider strategies of an organisation, including product bundling, designing a discount policy, and review management.

Application of Analytics

Descriptive Analytics

discounted_price	actual_price
Mean	3125.311	Mean	5444.991
Standard Error	181.4303	Standard Error	284.1211
Median	799	Median	1650
Mode	199	Mode	999
Standard Deviation	6944.304	Standard Deviation	10874.83
Sample Variance	48223364	Sample Variance	1.18E+08
Kurtosis	25.64359	Kurtosis	29.72308
Skewness	4.452395	Skewness	4.559879
Range	77951	Range	139861
Minimum	39	Minimum	39
Maximum	77990	Maximum	139900
Sum	4578580	Sum	7976911
Count	1465	Count	1465

The descriptive analytics was applied to summarise and interpret the pricing structure of the products that were listed on Amazon. For discounted_price and actual_price values, the central tendency and dispersion were calculated as indicated in the summary table. The average actual price of all products is around 5,445 INR. Still, the average price after discount is 3,125 INR, which indicates the aggressive nature of the price and promotional practice of Amazon in the Indian e-commerce market, where it competes against companies like Snap, dealer discount products like Galaxy J models of Samsung and others.

The difference between the mean and the median value of the discounted price (1190.25) is large, indicating a positively skewed distribution shape, and the values of both actual price skewness (4.55) and discounted price skewness (4.45) are very high. This means there is a generation of goods at the low end of the price and a long tail of high-priced goods. The absolute value of the standard deviation in actual price (10,874) and discount price (6,944) indicates the huge deviation that the broad range of products can cause, from cheap cable to top-range electronics. Such insights can give a good base of how to unfold the prices and division of customers that Amazon may target (budgetary and top-quality buyers).

Diagnostic Analytics

• To determine the cause of the different customer responses, diagnostic analytics was implemented.
• Correlations between the percentage of discount, price, rating, and number of reviews were analysed, in order to find possible factors that lead to customer engagement.
• For example, products with larger discounts typically received more reviews, indicating a positive correlation between price cuts and customer activity.

This implies that the effects of promotional pricing may extend beyond sales levels to influence the level of user responses, ultimately affecting the products that are displayed on the platform. In addition, the use of mean ratings, accompanied by the price list, revealed customers' views on the value of money or whether low-priced items were judged more strictly. This exploration was made possible using tools like pivot tables, scatter plots, and conditional formatting in Excel, which helped me to describe patterns of behaviour in review and product performance across the category.

Predictive Analytics

• Predictive analytics offers the opportunity to forecast future outcomes based on past trends (Oluwatoyin et al., 2024).
• Rating count, discount percentage, and product category can be identified as important variables in this project, out of which one can construct models that will profile the future success of a product.
• An example of the analysis presented in a linear regression model would be the determination of whether increased discounts will always culminate in increased levels of engagement and whether certain items are reviewed more positively than others.

Even though the use of more sophisticated tools such as the decision trees, or support vector machines was not implemented in this analysis, the structure of the dataset justifies the possibilities of using the more accurate tools in the future. These models would help the category managers at Amazon predict the outcomes of pricing strategies as well as help them understand the products that are supposed to be successful when launched.

Prescriptive Analytics

• Prescriptive analytics requires prescribing dialectical courses of action to be undertaken depending on predictive observations (Marjuni et al., 2022).
• Based on the identified patterns in the form of a descriptive and predictive analysis, one can offer optimal pricing models of Amazon product listings.
• As an illustration, when it is discovered that a 999 price with 45 per cent discount offers a high number of ratings and positive comments, this price-discount can be proposed to similar products.

The Solver tool in Excel can be used to conduct a what-if analysis and optimisation, such as ensuring a minimum profit margin is reached or a desirable number of reviews per week. A strategic use of analytics may allow to the improvement of strategic decision-making through the provision of evidence-based recommendations based on consumer behaviour.

Machine Learning and Artificial Intelligence

• Machine learning and artificial intelligence is already present within Amazon, in more limited forms, via recommendation engines, dynamic pricing, and fraud detection functions (Fanti, Guarascio and Moggi, 2022).
• These technologies are not directly incorporated by this project, although the status of the dataset favours the possibility of such applications.
• More accurate marketing and stock choices were possible with clustering being done on products based on price, rating, and the level of discount, utilising clustering algorithms, e.g. K-means.

Furthermore, the review content could be analysed using natural language processing techniques and sentiment analysis. This would enable Amazon to automatically identify positive or negative trends in customer feedback and provide real-time corrections on what is offered, its description, or even the price. When combined with business intelligence tools, these advanced analytics methods can significantly enhance the value of customer data.

Strategic Impact

• The use of analytics in this Amazon project gives significant strategic opportunities that also align with the core goals of the organisation
• In terms of the satisfaction of customers, efficiency of operations, and leadership within the market.
• Through deliberate study of price data and customer reviews Amazon is able to make decisions that directly impact on the nature of competition.

The information gained during the development of the descriptive and diagnosing analytics is used to improve the categorisation of products and find pricing strategies that can improve engagement with customers. This plays a vital role in a very competitive e-commerce market, in which the customer experience and prices are used to distinguish.

• The analytical process also enhances Amazon's strategic working capacity, as managers can refine design tactics to predict customer behaviour.
• By knowing the correlation between the discount rates and review numbers
• The rating of a product, Amazon is able to determine the future of success of a product listing (Vollero, Sardanelli and Siano, 2021).

This will help in the optimisation of inventory, more relevant targeted marketing and it will enhance the conversion rate that are all geared towards the objective of Amazon that is to maximize customer value at scale.

• Prescriptive analytics helps make evidence-based decisions in a strategic way, recommending the best prices that strike a balance between profits and customer satisfaction (Ridwan, 2025).
• These suggestions can be integrated into the general marketing tactic and financial formulas.
• With time, this analytical model can be incorporated in Amazon dynamic price engine to make necessary adjustments in real-time depending on the market conditions, consumer demand and competitive pressure.

This strategic advantage depends on the significance of the work of data scientists and analytics teams.

• Due to the continuous updates of models with real-time data, Amazon will be able to operate with high agility and act ahead of trends.
• Besides, utilising the big data is not only about the improvement of pricing and product strategies (Aljumah, Nuseir and Alam, 2021).
• It allows making personalised recommendations, targeted advertisements, and customer retention policies, which are vital to the expansion in the long perspective.

By incorporating analytics into strategic practices, Amazon constructs a data-driven culture that is perpetually leading to innovation and efficient practices.

Digital Tools & Ethics

• In this project, a few digital tools were used or mentioned to complete some analytical tasks.
• Microsoft Excel was used as a basis of data cleaning, statistical summarisation and what-if scenario analysis (Sneha Gathani et al., 2025).
• Tasks like pivot tables, regression mercy, and Solver were used in descriptive, diagnostic, and prescriptive modelling.
• Such tools do not require special knowledge and are rather common, which is why they are well-suited to both exploratory and strategic-level analytics.

In a more developed case, dynamic dashboards with product performance visualised with separate categories and pricing segments according to Tableau or Power BI platforms could be created.

• In addition to Excel, the fact that machine learning libraries are available for Python, in particularly scikit-learn and natural language processing libraries
• It provides a huge opportunity to extend analytics capabilities to a bigger scale.
• Such tools also allow automating complex analysis activities and processing of large amounts of unstructured data, including those about customer reviews, which would take a long time to analyze manually.

Digital capabilities will also enter into the product development and customer service processes as Amazon continues to develop its AI infrastructure.

• On the one hand, there are significant advantages associated with the utilization of digital tools
• On the other hand, the practice implies ethical concerns.
• The utilization of customer data, especially review content and identifications of users should be suspended by the privacy rules, including the General Data Protection Regulation (GDPR) (Quach et al., 2022).

This ensures the anonymity of personal data and the security of its storage, thereby maintaining customer confidence. In addition, when implementing machine learning models, matters of algorithmic bias, transparency, and accountability should be considered. As an illustration, consider a situation in which the models that determine the sentiment of reviews remain biased to a particular category of products; unfair promotion or demotion of a product might thus occur within the platform.

• Amazon also needs to address the ethical concerns of playing with the customer's minds with the help of automatic pricing or recommendation engines (Nunan and Di Domenico, 2022).
• Analytics should be used ethically and focus more on transparency and fairness, where the customers are not subject to manipulation by personalised algorithms.
• It is possible to minimize these risks by developing strong governance structures and audit processes that would facilitate the adequate use of digital tools.

Therefore, ethical responsibility and digital literacy should always be intertwined in order to reach the potential of analytics with sustainable and trustworthy effect.

Conclusion & Recommendations

• Project analytics In this project, the readers have seen how project analytics could be concisely used against actual e-commerce data, where strategic decision support and operational excellence could be pushed.
• Through the analysis of 1,425 product entries of Amazon India, this idea has been able to come up with insights in the realms of descriptive, diagnostic, predictive and prescriptive analytics.
• These insights indicate the high weight of the discount approach in customer interaction with the reference to the connection between price cut and the volume of reviews.

Descriptive analysis indicated a skewed distribution of prices with high variability that shows the existence of both the budget and premium segments of products. The role of pricing and discount levels in the context of customer ratings and review behaviour became further explained through diagnostic analytics.

• This application of predictive analytics has a good chance of enabling Amazon product managers to predict the success of new listings.
• The model can be developed in terms of discount percent, category of the product and historical rating, so that the response of customers can be predicted to set the price on strategic grounds.
• The development of such optimum values of discounts at which profit margin levels on one hand (Nukala and Rao, 2021).

The presented integration of digital technology, such as Excel, and possible expansion towards Power BI and Python-based analytics displays the area of increasing significance of digital literacy within contemporary project dimensions. Furthermore, both moral aspects of utilising customer-generated information should be identified. It is imperative to ensure transparency, data privacy and fairness in the analytics models in order to retain customer trust and comply with the regulations.

• It is suggested, based on the results of this analysis, that Amazon should keep applying the structured data on customer reviews so as to provide x-ray pricing and promotions.
• Generation of automated predictive models, which combine the level of discounts, review metrics, and sentiment scores are going to be another factor that will enhance the competitive edge of Amazon.
• Moreover, sentiment analysis on the contents of the reviews may offer more qualitative details, allowing the company to respond to the products level issues even before they are asked.

Long-term strategic capability will also be sustained by investing in training employees on more advanced analytics tools, as well as training employees on how to use data ethically.

References

Aljumah, A.I., Nuseir, M.T. and Alam, Md.M. (2021). Traditional marketing analytics, big data analytics and big data system quality and the success of new product development. Business Process Management Journal, 27(4), pp.1108–1125. doi:https://doi.org/10.1108/bpmj-11-2020-0527.

Das, B., Santosh, M., Prof, R. and Marketing Operations Manager (2023). INTERNAL GUIDE EXTERNAL GUIDE Summer Internship Report 2021. [online] Available at: http://biitm.dspaces.org/bitstream/123456789/505/1/Susmita%20Panda%20SIP.pdf [Accessed 23 May 2025].

Fanti, L., Guarascio, D. and Moggi, M. (2022). From Heron of Alexandria to Amazon’s Alexa: a stylized history of AI and its impact on business models, organization and work. Journal of Industrial and Business Economics, 49(3), pp.409–440. doi:https://doi.org/10.1007/s40812-022-00222-4.

Marjuni, Adly , R. and Ulwani, F. (2022). Analysis of The Best Strategy for Great Education Policy Using Prescriptive Analytics (Indonesian School Experience). Journal of Social Transformation and Regional Development, [online] 4(2), pp.51–58. Available at: https://penerbit.uthm.edu.my/ojs/index.php/jstard/article/view/13103 [Accessed 29 Jun. 2025].

Nukala, V.B. and Rao, S.S.P. (2021). Role of debt-to-equity Ratio in Project Investment valuation, Assessing Risk and Return in Capital Markets. Future Business Journal, [online] 7(1), pp.1–23. Available at: https://fbj.springeropen.com/articles/10.1186/s43093-021-00058-9.

Nunan, D. and Di Domenico, M. (2022). Value creation in an algorithmic world: Towards an ethics of dynamic pricing. Journal of Business Research, 150, pp.451–460. doi:https://doi.org/10.1016/j.jbusres.2022.06.032.

Oluwatoyin , Adaga, M., Efe Egieya, N.Z., Kuzankah, S., Abdul, Abrahams, O. and Farayola, A. (2024). Advancements in predictive analytics: A philosophical and practical overview. World Journal Of Advanced Research and Reviews, 21(3), pp.240–252. doi:https://doi.org/10.30574/wjarr.2024.21.3.2706.

Quach, S., Thaichon, P., Martin, K.D., Weaven, S. and Palmatier, R.W. (2022). Digital technologies: Tensions in Privacy and Data. Journal of the Academy of Marketing Science, 50(1), pp.1299–1323. doi:https://doi.org/10.1007/s11747-022-00845-y.

Ridwan, I. (2025). Optimizing Enterprise Decision-Making through Causal Machine Learning Models and Real-Time Business Intelligence Integration. International Journal of Advance Research Publication and Reviews, [online] 02, pp.67–88. doi:https://doi.org/10.5281/zenodo.15386602.

Schuurman, W. (2025). ‘Beyond Transactions : How B2B Platforms in Construction Progressively Create and Monetise Value through Data-Driven Value Propositions’ - University of Twente Student Theses. Utwente.nl. [online] doi:https://purl.utwente.nl/essays/105215.

Sneha Gathani, Liu, Z., Haas, P.J. and Çağatay Demiralp (2025). What-if Analysis for Business Professionals: Current Practices and Future Opportunities. Springer, pp.1–17. doi:https://doi.org/10.1145/3706598.3713672.

Vollero, A., Sardanelli, D. and Siano, A. (2021). Exploring the Role of the Amazon Effect on Customer expectations: an Analysis of User‐generated Content in Consumer Electronics Retailing. Journal of Consumer Behaviour, 22(5), pp.1–12. doi:https://doi.org/10.1002/cb.1969.