Analysis of Pet Supplies Demand and Influencing Factors Based on

Logistic Regression Model

Wenyu Li

Honours College, Capital Normal University, Beijing, 100101, China

Keywords: Pet Supplies, Market, Consumer, Influencing Factors, Logistic Regression Model.

Abstract: With the continuous development of society and the influence of various factors, pets have become an

important member of the family. This has brought about a booming development of the pet supplies market,

so studying people's demand for pet supplies and the impact of various factors on purchasing intention is of

great significance to related companies and merchants. This study will establish a linear regression model

through the data collected from the questionnaire to explore the core influencing factors and consumer

behavior characteristics of the market demand for pet supplies. The study found that when consumers choose

pet products, product quality and safety, product function and practicality are the primary considerations,

followed by brand reputation, price performance and health benefits of product ingredients. The study further

reveals the heterogeneity of consumer preference groups: high-income groups pay more attention to brand

and quality, while middle- and low-income consumers are more sensitive to price. However, linear regression

models cannot effectively explain complex interaction effects (nonlinear effects of pet attachment emotion on

consumption decisions, etc.). Future research can deepen feature analysis in combination with machine

learning and other methods.

1 INTRODUCTION

After experiencing the closure and isolation caused by

the epidemic, pets have become an important and

indispensable member of many families. Pet owners

are willing to spend more on their pets, especially in

terms of money. This has driven the rapid

development of pet-derived services, such as the pet

supplies market. The continued growth of the number

of newly established pet-related companies has not

only made the products more diversified, but also led

to increasingly fierce market competition. In order to

stand out from the competition, companies need to

have an in-depth understanding of consumers' needs

and preferences and provide products and services

that are more in line with consumers' expectations. At

the same time, consumers' requirements for the

quality, function and safety of pet products are

increasing. However, the booming development of

the pet supplies industry has also led to the

proliferation of a large number of misinformation or

false propaganda, which has disrupted the market

order.

Pet owners have doubts about the safety of pet

food and supplies (Di Cerbo et al., 2017).At the same

time, some pet supplies may not be suitable for the pet

itself. Bläske et al. (2022) mentioned in the article:

Most of the pet supplies on sale surveyed have a

certain degree of missing product information, which

leads to difficulty in purchasing. Therefore, what

factors consumers pay more attention to when

choosing pet supplies have become the focus of this

research. When reading relevant materials and

looking for existing research, it found an article

studying pet clothing consumption (Wang et al.,

2023). Wang et al. (2023) analyzed the data collected

and found that consumers pay the most attention to

pet clothing styles, accounting for 46.24%. Although

other needs, such as price, quality and fabric, account

for less than styles, only 13.87%, 12.14% and

15.03%, the proportion is still relatively high

compared to other aspects such as function (3.47%),

which has certain implications for our analysis

results. In the article, Gromek and Perek-Białas

(2022) analyzed whether factors such as social

background, economic status and demographic

characteristics will affect the family's spending on

pets by constructing a logistic logical model. At the

same time, there is also an article that studies the

demand factors for pet supplies using Gansu

134

Li, W.

Analysis of Pet Supplies Demand and Inﬂuencing Factors Based on Logistic Regression Model.

DOI: 10.5220/0013815400004708

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 2nd International Conference on Innovations in Applied Mathematics, Physics, and Astronomy (IAMPA 2025), pages 134-138

ISBN: 978-989-758-774-0

Province, China as an example (Ning et al., 2022).

Ning et al. (2022) used the form of an online

questionnaire in the study, collected 354 data and

sorted it out. Ning classified pet supplies and explored

the future consumption tendencies of consumers who

purchase different types of pet supplies, such as users

who buy more pet food, will have a higher probability

of continuing to consume pet supplies. At the same

time, Ning also conducted a word frequency analysis

of consumers' keywords for pet supplies, pointing out

that consumers have a higher demand for price and

quality. However, Ning did not strongly associate

these factors with the desire for consumption, which

would also be the main purpose of this study.

That is, the relevant data collected were analyzed

using the logistic regression model to evaluate the

fitting effect of the model, and to explore what factors

consumers are affected when purchasing pet supplies.

Therefore, conducting this research has multiple

significance. For enterprises, it can guide the research

and development and marketing of pet supplies. For

consumers, they can see whether they have paid too

much attention to one aspect and ignored other

aspects of the product when purchasing products.

2 RESEARCH METHODS

2.1 Data Collection and Processing

Because the first-hand data is collected by issuing

questionnaires, the questionnaires are posted on the

Internet and data is collected from users of all ages

across the country. Finally, 151 data were collected,

and the distribution of gender, age and cities was

relatively balanced. In the data preprocessing stage,

an abnormal questionnaire data was first ruled out,

and then after the credibility coefficient estimate , the

Longbach Alphaα coefficient is 0.61, which meets

the basic reliability requirements (Ru & Zhang,

2011). It can be considered that the data collected in

this questionnaire is reliable. When processing data,

this article uses the assignment method to assign the

value of the factors of interest to 1, and if it is not paid

attention, the value is assigned to 0. For example, if

the consumer takes the price factor into consideration

when spending, then in this data, the price is assigned

to 1, otherwise it is 0.

2.2 Data Analysis

First, different types of data will be fitted through the

logistic regression model to establish a logistic

regression model. Logistic regression model is a

generalized linear model with advantages such as

simple and easy to interpret, high computing

efficiency, loose data requirements and strong

scalability. It is based on intuitive probability output,

easy to understand parameters, efficient solution

process, suitable for small data sets, and has certain

robustness to outliers. In addition, the model can also

introduce interactive terms and polynomial terms,

which are combined with other technologies and are

widely used in binary classification problems and are

used to predict binary classification results. Its

mathematical form can be expressed as:

𝑙𝑜𝑔









=𝛽



+𝛽



𝑥



+𝛽



𝑥



+𝛽



𝑥



+𝛽



𝑥



𝛽



𝑥



(1)

Among them: p is the probability of an event

occurring (that is, the consumer's consumption

intention is also assigned to a specific value, willing

to be 1, and unwilling to be 0). 𝛽



is the intercept

term, 𝛽



, 𝛽



, 𝛽



, 𝛽



,𝛽



are regression coefficients,

corresponding to the variables 𝑥



,𝑥



,𝑥



, 𝑥



,𝑥



respectively. 𝑥



,𝑥



,𝑥



,𝑥



,𝑥



are independent

variables, corresponding to price, quality, brand,

function, and reputation, respectively. The above

independent variables are factors that consumers are

concerned about when purchasing pet supplies in the

questionnaire.

3 RESEARCH RESULTS

Table 1 ： Coefficient estimation results of Logistic

regression model

Coefficients Estimate Std.

Erro

value

Pr(>|z|)

Intercept -2.141 1.399 -1.530 0.12594

rice 4.597 1.523 3.019 0.00254

qualit

1.391 1.251 1.112 0.26634

ran

2.403 1.189 2.020 0.04335

function 2.169 1.065 2.037 0.04168

utation 1.784 1.242 -1.437 0.15081

In this logistic regression analysis, it evaluated the

impact of multiple variables on the target variable.

Results are shown in Table 1, price, brand, and

function are significant factors, while quality and

reputation do not reach statistical significance.

Specifically, the p-value of the price variable is only

0.00254, which is well below the significance

threshold of 0.01, and two asterisks are marked in the

results, indicating that their effect on the target

variable is very significant. In contrast, the p-values

for brand and feature are 0.04335 and 0.04168,

respectively, both slightly below the threshold of

Analysis of Pet Supplies Demand and Inﬂuencing Factors Based on Logistic Regression Model

135

0.05, marked with an asterisk indicating that they also

have a significant effect on the target variable, but are

slightly less significant than the price. The p-values

for quality and word of mouth were 0.26634 and

0.15081, respectively, both above 0.05, not reaching

the significance level, indicating that their effect on

the target variables was not significant. Furthermore,

the p-value of the intercept term is 0.12594, which is

also not at the significance level, indicating that in this

model, the intercept term contributes not significantly

to the predicted results. These results suggest that

price has the strongest statistical significance of all

significant variables and may be more critical to the

predictive power of the target variable.

Figure 1: Residual and fitted value graph (Photo/Picture

credit: Original).

Figure 2: Normal Q-Q Picture (Photo/Picture credit:

Original).

4 FITNESS TEST

Residual vs. fitted value graphs are often used to

check whether the model has a nonlinear relationship

or heteroscedasticity. If the residuals are randomly

distributed around the horizontal line (0), then it

means that the model hypothesis is true; if there is a

significant curve shape, it indicates that there may be

a nonlinear relationship or heteroscedasticity.

Obviously, the distribution of residuals in Figure 1 is

random, indicating that the logistic regression model

can better explain the problem. Normal Q-Q graphs

are methods used to check whether residuals are

approximately normal distributions. If the sample

point falls roughly on a straight line, it means that the

residual is approximately normal distribution. If the

sample point deviates from the line, it indicates that

the residual may not obey the normal distribution.

Obviously, the sample points in Figure 2 fall roughly

on a straight line, and the residuals are approximately

normal distribution. In summary, it can conclude that

the model satisfies the basic assumption of linear

regression, has good fitting effect and high reliability,

and can be used for further analysis and explanation.

5 DISCUSSION

In this study on the influencing factors of pet supplies

selection, it was found that the positive driving effect

of price on consumers' willingness to purchase is the

most obvious, and it is a key factor that affects

consumers' decision-making when purchasing pet

supplies. This shows that when consumers buy pet

supplies, they often first consider whether the price of

the product meets their own budget. Products with

high cost performance are more favored by

consumers. Of course, this may also be related to the

subject's occupation, considering that students do not

have their own income, but emerging cloud pet

raising and feeding animals on campus may be a

strong driving force for them to buy related products,

so they will consider prices more.

In addition, brand and function are also important

considerations for consumers when choosing pet

supplies. Both brand awareness and product

functional characteristics have a significant positive

impact on whether you are willing to consume. Nie

(2024) mentioned in his report: At the 2024 Asian Pet

Show, there are many new pet supplies that are

popular among consumers. Xiao, Wang & Li (2021)

mentioned in the statistical results of the article:

About 68.6% of pet owners believe that brand

reputation is an important consideration when

choosing pet supplies. It can see that products of well-

known brands or emerging brands are being seen

more widely, which may be because well-known

brands usually have higher credibility and better after-

sales service. At the same time, the functional

IAMPA 2025 - The International Conference on Innovations in Applied Mathematics, Physics, and Astronomy

136

characteristics of the product have also attracted the

attention of consumers. If it link the two together, it

can also think that a good brand usually has better and

more functions, and the two have a certain

connection.

However, word of mouth and quality did not show

significant positive effects in this study. The

inconspicuousness of the word-of-mouth factor may

be caused by a large number of advertisements on the

Internet: the product itself is very different from the

publicity and cannot satisfy consumers. At the same

time, Sun, Wang, Zheng & Su (2025) mentioned in

the article: Most users who participated in the

research believe that the materials of pet toys in pet

supplies are unsafe and of poor quality. At the same

time, Liu et al. (2025) proposed that the pet economy

has developed too rapidly in recent years, and many

supervision and policies have not been formulated or

implemented, so the inconspicuousness caused by

quality can be seen.

So for enterprises, they should pay more attention

to product price strategies and ensure that the products

are competitive in the cost-effectiveness competition

through innovative production methods to attract

consumers who pay attention to costs. Secondly,

brand building and functional innovation are also key.

Enterprises need to enhance their market

attractiveness by enhancing brand awareness and

developing practical products with good functional

characteristics. Furthermore, although word of mouth

and quality did not significantly affect purchasing

decisions in this study, companies should not ignore

it. In the long run, high-quality products and good

reputation are the cornerstones of the sustainable

development of the brand. Therefore, enterprises

should continue to optimize product quality and

actively manage consumer feedback to enhance brand

reputation and customer loyalty.

To sum up, when consumers choose pet supplies,

price is the most important factor, and brand and

function also have significant positive impact.

However, word of mouth and quality did not show a

significant positive impact in this study. This article

only gives inferences, and the reasons need to be

further explored.

6 RESEARCH LIMITATIONS

AND PROSPECTS

This study has certain limitations in sample selection

and data collection. The data in the questionnaire

sample is small and is mainly concentrated in urban

areas. There is insufficient research on the

consumption behavior of pet owners in rural areas.

Future research can further expand the sample range

and cover the data to pet owners of different regions

and income levels to obtain more comprehensive

research results. Of course, when purchasing pet

supplies, there may be complex interaction effects.

For example, owners who are deeply attached to pets

may be willing to consume more for pets, and these

sample points may affect the accuracy of the model.

Therefore, in the future, it can consider introducing

deeper analytical methods, such as combining deep

learning to explain some consumer behaviors. In

addition, with the continuous development and

changes of the pet supplies market, future research

can also focus on the impact of emerging technologies

on the pet supplies market, such as the impact of

artificial intelligence, the Internet of Things, cloud pet

raising, etc. on the pet supplies market, as well as

consumers' acceptance and use behavior of these new

technologies.

7 CONCLUSION

This study analyzed the factors that influence

consumers to purchase pet supplies through

questionnaire surveys and Logistic regression

models. The study found that price is the primary

factor in consumer decision-making and has a

significant positive impact, indicating that consumers

are extremely concerned about cost-effectiveness

when purchasing pet supplies. Brand awareness and

product functionality also have a significant positive

impact on purchasing intention, but the impact of

quality and reputation is not significant, which may

be related to market information asymmetry and

general consumer trust in quality. The study also

found that consumer preference groups are

heterogeneous, high-income groups pay more

attention to brand and quality, and middle- and low-

income groups are more sensitive to prices. The

research limitation is that the sample size is small and

concentrated in urban areas. In the future, the sample

range can be expanded and deep learning can be

introduced to further analyze complex interaction

effects such as pet attachment emotions, while paying

attention to the impact of emerging technologies on

the pet supplies market.

Analysis of Pet Supplies Demand and Inﬂuencing Factors Based on Logistic Regression Model

137

REFERENCES

Bläske, A., Schwarzer, A., Ebner, M. V., Gerbig, H., Reese,

S., Erhard, M., & Wöhr, A. C. 2022. Evaluation of small

mammal pet supplies offered in German retail under

animal welfare aspects. PLOS One, 17(2), e0262658.

Di Cerbo, A., Morales-Medina, J. C., Palmieri, B., Pezzuto,

F., Cocco, R., Flores, G., & Iannitti, T. 2017. Functional

foods in pet nutrition: Focus on dogs and cats. Research

in Veterinary Science, 112, 161–166.

Gromek, N., & Perek-Białas, J. 2022. Pet goods

consumption in Polish households. Econometrics, 26,

1–20.

Liu, C., Cui, Y., Wei, S., Xie, A., & Zhao, Z. 2025.

Research on the current status and countermeasures of

my country's pet economy development. Business

Exhibition Economy, (05), 119–122.

Ning, T., Wang, Y., Zhang, W., Ran, J., & Qin, L. 2022.

Analysis of the current status of pet supplies demand

and influencing factors—taking Gansu Province as an

example. Gansu Animal Husbandry and Veterinary

Medicine, 52(05), 64–67.

Nie, Y. 2024. The scale of the Asian Pet Show hits a new

high, and the industry shows a trend of consumption

upgrading. Securities Times, A04.

Ru, P., & Zhang, P. 2011. Regarding the calculation and

analysis method of credibility in questionnaire surveys.

Journal of Xinjiang Radio and Television University,

15(01), 46–50.

Sun, W., Tang, M., Zheng, B., & Su, Z. 2025. New hotspot

of pet economy: Study of consumers' purchase intent for

pet toys in China. Qeios.

Wang, J., Chen, Y., Chen, X., & Chen, L. 2023. Analysis of

pet clothing consumption behavior and psychological

needs based on questionnaire. Chemical Fiber and

Textile Technology, 52(01), 70–73.

Xiao, Y., Wang, H. H., & Li, J. 2021. New markets for pet

food in China: Online consumer preferences and

consumption. China Economy, 54(6), 430–440.

IAMPA 2025 - The International Conference on Innovations in Applied Mathematics, Physics, and Astronomy

138