The Effect of Price on E-Commerce Platforms: Statistical Evaluation
of Amazon's Pricing Strategy
Xinguo Luo
a
Institute of Problem Solving, University of Toronto, 27 King’s College Cir, Toronto, Canada
Keywords: Amazon, Online Retailing, Sales, Customer Rating, Pricing Strategy.
Abstract: This present study is meant to explain whether Amazon’s pricing strategies, like providing discounts or not,
have some benefits for product sales and consumer ratings (i.e., customers’ satisfaction with the products).
The original dataset was downloaded from the Kaggle platform, and it contains sales records of about 1.4
million products during September 2023. However, in order to fit this present study better, a simplified dataset
with 166,481 samples was filtered from the 1.4 million samples of the original dataset according to the average
monthly consumption of Americans. The method of this study was simple linear regression, which can be
used to understand the association between two variables. Then, two linear models were generated with Excel.
After all, the results showed that the p-values are statistically significant for both discounts and sales and
discount and customer ratings. In more detail, discounted products have better sales than regular goods, and
they have better customer evaluation.
1 INTRODUCTION
1.1 Background
With the growth of the Internet and electronic devices
in recent years, online retail platforms have
experienced significant development in many aspects.
More and more people like to browse and buy what
they want through e-commerce platforms. According
to Jap et al. (2022), in 2020, some famous online
marketplaces such as Amazon, Taobao, and T-mall
earned $2.7 trillion in global sales, which held the
majority of global online sales of that year. Besides
selling things, Amazon has entered even more
different areas these years, such as education. All of
this shows that e-commerce platforms like Amazon
have already become a substantial part of the global
economy.
1.2 Literature Review
Studies indicate that many people visit shopping
websites several times a week and shop at online
retail stores at least once a month (Menon, 2023).
Each platform has different advantages, and it is
a
https://orcid.org/0009-0001-3926-6151
important to build suitable and customized sales
strategies to develop an online sales platform (Kwak
et al., 2019; Zhang et al., 2023). For example, some
products sold on Amazon.com always remain at a low
price or with a high discount rate (Reimers &
Waldfogel, 2017).
Pricing is an important but complex part of retail,
especially for online retailing. Unlike traditional
offline stores, people can see the price of products
more directly when shopping online and may be more
sensitive to price. Online retail prices relate to many
aspects, such as customer satisfaction, sales, and
profit (Nataraja et al., 2017). In this case, although the
price of goods may change frequently according to
market conditions, many e-commerce platforms still
implement computer algorithms to set prices
automatically to increase profits (Aparicio et al.,
2024). Some companies even rely on high sales to
stay afloat at extremely low prices (Sussman, 2019).
Consequently, this study will mainly focus on
Amazon's pricing strategy and its potential influence.
It seems that lower and discounted prices can help
increase customer satisfaction and sales.
Luo, X.
The Effect of Price on E-Commerce Platforms: Statistical Evaluation of Amazon’s Pricing Strategy.
DOI: 10.5220/0012925300004508
Paper published under CC license (CC BY-NC-ND 4.0)
In Proceedings of the 1st International Conference on Engineering Management, Information Technology and Intelligence (EMITI 2024), pages 239-242
ISBN: 978-989-758-713-9
Proceedings Copyright © 2024 by SCITEPRESS – Science and Technology Publications, Lda.
239
2 METHOD
2.1 Data
The original dataset found on the Kaggle website
includes extensive information about Amazon's
product sales in September 2023. It contains the
records of 1,426,337 Amazon products, each with a
product ID, a title, links to product images, official
links to the products, rating, number of reviews,
present price, primary price, category, bestseller
status, and number of products sold on Amazon in
September 2023 (Asaniczka, 2023). In these 1.4
million samples, the price of the products ranges
widely from $0.01 to $19,700. However, according to
the U.S. Bureau of Labour Reports (2023), the 2022
annual expenditures for all consumers in the United
States averaged $72,967, and only 15.2% of that
spending is likely to be related to online shopping. In
other words, Amazon consumption will be at most
$924.25 per month for most American consumers.
The products sold for over $924.25 seem less familiar
for daily purchases and were removed from the
dataset. Also, products under $50 are deleted since
the discount rate is too small. In addition, because
some of the product information is incomplete, those
products will not be analyzed.
Therefore, the simplified dataset has 166,481
products in 270 categories. The price interval
becomes $50 to $924.25, with some discounts and
some not. Besides, product ratings range from 0 to 5
stars, with an average rate of 3.8 (SD = 1.51), which
shows that most customers are satisfied with the
goods. The top seller in September 2023 was a
refrigerator filter that sold 40,000 pieces in a month
and had never been a best seller before.
2.2 Design
2.2.1 Pricing Strategy and Sales
A linear regression model will be generated through
Excel. The independent variable is the pricing
strategy, which determines whether the product is
discounted. It is a dummy variable that was created
for non-quantitative classification. If the present price
is greater or equal to the original listed price, it is
considered to have no discount (i.e., discount equals
0). Otherwise, if the current price is smaller than the
primary one, a discount was implemented and could
be represented by 1. The response variable is the
number of sales in September 2023 collected from
Amazon.com. This model aims to find the
relationship between price discounting activity and
sales volume.
The original formula for this linear regression
model is:
𝑦
∗
(1)
where 𝑦1 refers to the number of sales, 𝑥1 refers
to whether there is a discount or not, 𝛽1 shows how
the sales will change with the discount, 𝛽0where 𝑦
refers to the number of sales, 𝑥
refers to whether
there is a discount or not, 𝛽
shows how the sales will
change with the discount, 𝛽
is the number of sales
when there is not any account, and 𝜖
is the influence
of other confounding factors.
2.2.2 Pricing Strategy and Customer Rating
This study analysed whether discounts impact
customers’ satisfaction with the products they buy.
Whether there is a discount is an independent
variable, and the response variable is the stars of the
products. To be more specific, the stars are the scores
the products got that aim to evaluate customers'
satisfaction with the products. The statistical method
used is also the linear regression model:
𝑦
∗
(2)
𝑦
is the customer rating of products ranging from
0 to 5, 𝑥
refers to whether there is a discount or not,
𝛽
explains how the rating will change with the
discount, 𝛽
means the scores the products got when
there is not any account, and 𝜖
is the potential
random factors.
3 RESULTS
3.1 Pricing Strategy and Sales
Table 1: The linear model of pricing strategy and sales.
Coefficients t Stat P-value
Intercept 48.46824613 44.03809732 0
Discount 56.7219767 26.77102313 1.5201E-157
Table 2: The regression statistics of pricing strategy and
sales.
Multiple R .06547152
R Square .00428652
Adjusted R Square .00428054
Standard Erro
r
383.728542
Observations 166481
EMITI 2024 - International Conference on Engineering Management, Information Technology and Intelligence
240
Table 3: ANOVA test of pricing strategy and sales.
df SS MS F
Significa
nce F
Regressi
on
1
10553053
6.3
10553053
6.3
716.6876
787
1.5201E-
157
Residual
16647
9
24513632
187
147247.5
939
Total
16648
0
24619162
723
According to the regression model in Table 1, some
statistical relationships between the price strategy and
the sales quantity demonstrated that:
𝑦
. .
(3)
Table 2 exhibits that the R-squared value for the
linear regression is about 0.004, suggesting that the
model could only explain a small portion of the sales
variability and that there are many other factors
affecting sales the model did not include. However,
the p-value in Table 1 is much smaller than the
significant level, p < .001, representing that the
influence of discount for sales is statistically
significant. Since the coefficient is greater than 0,
there is a positive association between the pricing
strategy and sales, which means the sales will
increase as the discount gets bigger. Also, the model
is a good fit for the data because Table 3 represents
the significance of the F-statistic is less than 0.05.
3.2 Pricing Strategy and Customer
Rating
Table 4: The regression statistics of pricing strategy and
starts of products.
Multiple R .06547152
R Square .00428652
Adjusted R Square .00428054
Standard Erro
r
383.728542
Observations 166481
Table 5: ANOVA test of pricing strategy and starts of
products.
df SS MS F
Significa
nce F
Regressi
on
1
10553053
6.3
10553053
6.3
716.6876
787
1.5201E-
157
Residual
16647
9
24513632
187
147247.5
939
Total
16648
0
24619162
723
Table 6: The linear model of pricing strategy and starts of
products.
Coefficients t Stat P-value
Intercept 48.46824613 44.03809732 0
Discount 56.7219767 26.77102313 1.5201E-157
According to the tables above, the linear regression
model of pricing strategy and customer rating shows
a formula:
𝑦
. .
(4)
The statistical data in Table 6 exhibits a statistical
significance between the variables, p < .001. Because
of the positive coefficient, the product rating would
increase with the increase in the discount. Although
the R-square in Table 4 indicates that only 1.299% of
the variable could be explained by the model, the F-
statistic result (i.e., significance F) of the ANOVA
test in Table 5 is less than 0.05, indicating that the
model can significantly fit the collected data.
Based on the analysis of two linear regression
models above, because the two p-values in Table 1
and Table 6 are less than 0.01, there is a statistically
significant association between the pricing strategy
and the sales, as well as the pricing strategy and rating.
According to the coefficients, the higher the discount,
the higher the sales will increase significantly, and the
customer's evaluation of the product will also
increase. A suitable pricing strategy is vital for
Amazon to operate better.
4 DISCUSSION
True to the original assumption, the discount and low-
price strategy can improve product sales and ratings.
Firstly, discounts can easily stimulate people's
consumption. On the one hand, lower prices would
make consumers more willing to order the target item.
On the other hand, excitement may push consumers
to make some purchases that are not in their primary
plan, especially when discounts are high (Kim &
Tanford, 2021). Secondly, because the price is low,
the consumers’ tolerance for the product has become
higher. Although the goods may have some flaws,
people will also feel satisfied because of the relatively
lower price.
A proper promotion can bring many a lot of
benefits to a company. Amazon understood this idea
early on and provided a practical example. In the field
of online book retailing, Amazon has beaten many
competitors with the low prices they set (Reimers &
Waldfogel, 2017). Amazon has never raised book
The Effect of Price on E-Commerce Platforms: Statistical Evaluation of Amazon’s Pricing Strategy
241
prices in the past two decades, even though few
competitors have left (Reimers & Waldfogel, 2017).
While most companies trying to compete with low
prices have gone bankrupt, Amazon has already
passed the deficit stage (Sussman, 2019) and
increased revenue every year (Reimers & Waldfogel,
2017). Amazon has found a pricing strategy that
works best for it.
However, although the results may match
Amazon’s actual situation, this present study has
some limitations. Firstly, the dataset used in this study
only includes Amazon's sales in the United States
over a one-month period, which means it may not be
a good interpretation for other countries or times. For
example, some American holidays in September, like
Labor Day, may become a factor that leads people to
buy more items than usual for celebrating. Secondly,
by refining the limited dataset, the study only
discussed the influence of pricing strategies through
the data collected but did not pay much attention to
the products or the consumers themselves. Just as
people are influenced by shopping values and price
awareness, transactional tendencies, and coupons
when shopping at the mall (Khare et al., 2014), so are
they when shopping online. There are many
environmental and personal factors that may affect
the relationship between price, rating, and sales, but
it is hard to collect and define them.
5 CONCLUSION
In conclusion, the preliminary analysis has been
solved with the available variables: a promotion
strategy positively affects Amazon’s online retail
store, both for sales and customer satisfaction with the
products. Nevertheless, more potential variables are
involved in the interaction in real life, and the pricing
strategy may have a broader effect. As people
continue to do more studies in wider fields and
connect them with each other, more personal factors
can be added to the discussion, allowing people to
understand the pricing strategy and the customers
better.
REFERENCES
Aparicio, D., Metzman, Z., & Rigobon, R. (2024). The
pricing strategies of online grocery retailers.
Quantitative Marketing and Economics, 22(1), 1–21.
https://doi.org/10.1007/s11129-023-09273-w
Asaniczka. (2023). Amazon Products Dataset 2023 (1.4M
Products) [Data set]. Kaggle. https://doi.org/
10.34740/KAGGLE/DS/3798081
Jap, S. D., Gibson, W., & Zmuda, D. (2022). Winning the
new channel war on Amazon and third-party platforms.
Business Horizons, 65(3), 365–377. https://doi.org/
10.1016/j.bushor.2021.04.003
Khare, A., Achtani, D., & Khattar, M. (2014). Influence of
price perception and shopping motives on Indian
consumers’ attitude towards retailer promotions in
malls. Asia Pacific Journal of Marketing and
Logistics, 26(2), 272–295.
https://doi.org/10.1108/APJML-09-2013-0097
Kwak, J., Zhang, Y., & Yu, J. (2019). Legitimacy building
and e-commerce platform development in China: The
experience of Alibaba. Technological Forecasting &
Social Change, 139, 115–124. https://doi.org/1
0.1016/j.techfore.2018.06.038
Menon, S. (2023). Amazon Consumer Behaviour Dataset
[Data set]. Kaggle. https://www.kaggle.com/datasets/
swathiunnikrishnan/amazon-consumer-behaviour-data
set
Myers, S., Paulin, G. D., & Thiel, K. (2023). Consumer
expenditures in 2022. U.S. Bureau of Labor Statistics.
https://www.bls.gov/opub/reports/consumer-expenditu
res/2022/home.htm
Natarajan, T., Balasubramanian, S. A., & Kasilingam, D. L.
(2017). Understanding the intention to use mobile
shopping applications and its influence on price
sensitivity. Journal of Retailing and Consumer
Services, 37, 8–22. https://doi.org/10.1016/j.jret
conser.2017.02.010
Reimers, I., & Waldfogel, J. (2017). Throwing the Books at
Them: Amazon’s Puzzling Long Run Pricing
Strategy. Southern Economic Journal, 83(4), 869–885.
https://doi.org/10.1002/soej.12205
Sussman, S. (2019). Prime predator: Amazon and the
rationale of below average variable cost pricing
strategies among negative-cash flow firms. Journal of
Antitrust Enforcement, 7(2), 203–219. https://doi.org/
10.1093/jaenfo/jnz002
Zhang, T., Li, G., & Tayi, G. K. (2023). A strategic analysis
of virtual showrooms deployment in online retail
platforms. Omega (Oxford), 117, 102824-.
https://doi.org/ 10.1016/j.omega.2022.102824
EMITI 2024 - International Conference on Engineering Management, Information Technology and Intelligence
242
