Key Drivers of Consumer Purchase Decisions in Short Video

Platform Promotions

Haifeng Guo

UTS Business School, University of Technology Sydney, Sydney, 2007, Australia

Keywords: Digital Marketing, Data-Driven Marketing Decisions, Social Media Analytics, Behavioral Data Mining,

Business Analytics.

Abstract: With the rapid development of short video platforms, self-media marketing has emerged as a central strategy

in digital brand promotion. Although existing studies suggest that the number of comments on short video

advertisements may positively impact product sales, the role of comment content remains inconclusive. To

investigate this further, this study integrates survey responses with web-scraped datasets and applies

regression analysis to assess the relationship between user engagement and brand sales performance. The

results demonstrate a significant positive correlation between comment volume and sales figures on both

Douyin and Taobao, even when comments do not explicitly mention the brand. This may be due to the

influence of non-brand-related factors, including influencer appearance or emotional expression. Furthermore,

the findings reveal that high-income consumers tend to show stronger purchasing power and reduced price

sensitivity, yet their overall participation in short video e-commerce remains relatively low, highlighting

untapped market potential.

1 INTRODUCTION

With the widespread adoption of mobile internet and

the rapid rise of short video platforms, digital

marketing has become a vital strategic tool for

interaction between businesses and consumers

(Ratchford et al., 2022). Platforms such as Douyin and

TikTok have reshaped the way brand value is

communicated by offering highly interactive and

immersive environments, which in turn subtly

influence user attitudes and behaviours (Li & Xia,

2022; Tiago & Veríssimo, 2014). As consumers

increasingly rely on digital content to guide their

purchasing decisions, the effectiveness of short video

advertising has emerged as a significant topic within

the field of marketing communications.

Previous research has highlighted the importance

of advertising content characteristics, the role of key

opinion leaders (KOLs), and social interaction in

driving user engagement and purchase intention

(Zhang & Zhang, 2024; Chen & Liao, 2021).

However, the way in which these factors interact

across different platform environments remains

underexplored. In addition, the dynamic nature of

promotional content and algorithmic recommendation

systems poses challenges in evaluating the long-term

impact of brand communication (Lamberton &

Stephen, 2016).

This study focuses on the relationship between

media richness, KOL characteristics, interactivity, and

consumer purchase behaviour in the context of short

video advertising. By combining regression analysis

with visualisation tools, the study aims to identify the

variables most closely associated with sales

performance and user response, offering strategic

insights for marketers seeking to optimise digital

advertising campaigns across varied social media

ecosystems.

2 METHOD

This study investigates the key factors through which

short video platform promotions influence consumer

purchase decisions. Given the complex interplay of

subjective and objective variables involved in such

decisions, a mixed-methods approach was adopted,

combining questionnaire-based surveys with web data

extraction for enhanced reliability.

The questionnaire was designed with reference to

Wang and Li (2023), whose framework explored how

642

Guo, H.

Key Drivers of Consumer Purchase Decisions in Short Video Platform Promotions.

DOI: 10.5220/0013998000004916

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

In Proceedings of the 2nd International Conference on Public Relations and Media Communication (PRMC 2025), pages 642-647

ISBN: 978-989-758-778-8

social and technological affordances shape consumer

responses to embedded advertising. Additionally, Liu

and Chen (2024) demonstrated that short-form user-

generated video content (UGC) is significantly more

persuasive in driving purchase intent than traditional

image- or text-based formats. Drawing on these

insights, the survey focused on key constructs such as

frequency of interaction, preference for UGC, brand

identification, and actual purchase intent. The final

questionnaire comprised three sections: demographic

profiling, behavioural variables, and attitudinal

measures. It included closed-ended questions,

multiple-choice items, and a set of 5-point Likert scale

statements aimed at quantifying key psychological

and behavioural dimensions. In order to verify internal

consistency, both reliability and validity assessments

were undertaken to ensure the instrument’s robustness

before large-scale deployment.

The survey was disseminated via major social

platforms including WeChat, Xiaohongshu, and

Douyin, targeting users who regularly engage with

fashion-related content. Respondents were required

to follow a minimum of five fashion influencers to be

eligible. A total of 200 valid responses were collected

over the course of one month. The majority of

participants were aged 19 to 25, lived in tier-one

cities, and reported monthly income ranging from

RMB 10,000 to 20,000—an income bracket consistent

with high digital consumption behaviour in China’s

urban youth segment. In parallel, web-scraped data

were gathered using a Python-based crawler, which

retrieved promotional video statistics—namely likes

and comments—from 30 fashion-related KOLs on the

Douyin platform. To supplement this, sales data

corresponding to these KOLs ’ promoted brands

were collected from both Douyin and Taobao,

enabling a comparative assessment of platform-

specific performance and the influence of consumer

interaction metrics. Data collection occurred between

February 2024 and February 2025.

Following preprocessing in Python, the dataset

was cleaned and analysed using Microsoft Excel,

with regression analysis applied to explore variable

relationships. This study complied fully with research

ethics guidelines: all participants gave informed

consent, remained anonymous, and the collected data

were used exclusively for academic research.

3 RESULTS

In total, 200 valid questionnaires were collected and

analysed for this research. Among all city tiers, users

who followed 11–20 fashion influencers accounted for

the largest proportion, representing 46% of the sample

(see Figure 1). In terms of income distribution, 38% of

respondents reported monthly earnings in the RMB

10,000 – 20,000 range, which corresponds with the

sample's urban composition and higher engagement

with fashion-related digital content. The survey

findings indicate that 96% of respondents believe that

positive user comments and favourable brand-related

reviews increase their sense of brand identification.

This underscores the critical role of social proof in

shaping consumer perceptions, suggesting that peer

feedback and community sentiment have a significant

influence on individual attitudes toward fashion

brands.

To assess the correlation between user

engagement metrics and brand-related comment

content, linear regression analysis was conducted on

data from 30 key opinion leaders (KOLs) active on the

Douyin platform. The results revealed a low

explanatory power, with an R ² value of 0.073,

indicating that likes and comments together explained

only 7.3% of the variation in the dependent variable

(see Table 1). The overall model failed to reach

statistical significance (Significance F = 0.357), and

the individual p-values for likes and comments were

0.247 and 0.945 respectively, both well above the 0.05

significance threshold (see Tables 2 and 3). These

findings suggest that there is no statistically significant

linear relationship between the volume of engagement

and whether the comments contain brand-related

content. As shown in Figure 2, the age group of 36–45

has the highest proportion of consumers purchasing

high-end fashion, reaching 67%. In contrast, the 19–

25 age group demonstrates the strongest preference

for fast fashion, with 46% of respondents opting for it.

This suggests that consumers aged 36–45 are more

inclined towards premium fashion brands, whereas

those under 25 tend to favour fast fashion options. As

shown in Figure 3, 75% of individuals with a monthly

income between 3,000 and 6,000 opt for fast fashion,

while 40% of those earning over 20,000 tend to choose

high-end fashion. This pattern closely aligns with

variations in income levels across age groups, further

highlighting the significant role of economic capacity

in shaping consumer purchasing behaviour. Finally,

regression models examining the relationship between

comment volume and platform-based sales

performance demonstrated a clear positive correlation.

The model for Douyin yielded an R² of 0.700 and a

regression coefficient of 10.042, while the Taobao

model recorded an R² of 0.565 and a coefficient of

8.645 (see Tables 4). Both results were statistically

significant; however, the Douyin model demonstrated

a stronger fit, indicating that in-platform comment

Key Drivers of Consumer Purchase Decisions in Short Video Platform Promotions

643

engagement is a more effective driver of sales

performance within the short video ecosystem.

Alt Text for the figure: A bar chart showing that Tier 1 users

follow more KOLs than Tier 3 users, mainly in the 11–20

range.

Figure 1 Number Of KOLs Followed By Users From

Different City Tiers

Table 1. Regression Statistics between Comments

and

Brand-Related Comments 1

Regression Statistics

Multiple R 0.271

R Square 0.073

Adjusted R Square 0.005

Standard Erro

14.382

Observations 30

Table 2. Regression Statistics between Comments and

Brand-Related Comments 2

Significa

nce F

Regress

ion

2442.3

221.1

1.0

0.357

Residua

5584.

206.8

Total 2

6026.

Table 3. Regression Statistics between Comments and Brand-Related Comments 3

Coefficients Standard

Error

t Stat P-

value

Lower

95%

Upper

95%

Lower

95.0%

Upper

95.0%

Intercept 13.454 7.803 1.724 0.096 -2.557 29.465 -2.557 29.465

Total Likes

on Promo

Videos

3.46E-05 2.92E-05 1.183 0.247 -2.5E-

9.46E-

-2.5E-

9.46E-

Total

Comments

on Promo

Videos

-0.000 0.002 -0.070 0.945 -0.005 0.004 -0.005 0.004

PRMC 2025 - International Conference on Public Relations and Media Communication

644

Alt Text for the figure: A stacked bar chart illustrating fashion style preferences by age group, where users aged 36–45

prefer high-end fashion and those under 25 favour fast fashion.

Figure 2. Distribution Of Fashion Style Preferences Across Age Groups

Alt Text for the figure: A stacked bar chart showing fashion style choices across different income brackets, indicating that

higher-income groups tend to prefer luxury or designer brands.

Figure 3. Distribution Of Fashion Style Preferences Across Income Groups

Key Drivers of Consumer Purchase Decisions in Short Video Platform Promotions

645

Table 4. Regression Analysis of Comment Volume and Sales Performance on Two Platforms

Metric Douyin Taobao

Multiple R 0.837 0.752

R Square 0.700 0.565

Adjusted R Square 0.690 0.550

Standard Error 3355.254 3873.493

Observations 30 30

F-statistic 65.437 36.387

Significance F 8.29E-09 1.68E-06

Intercept -890.733 469.061

Total Comments Coefficient 10.042 8.645

Intercept P-value 0.564 0.792

Total Comments P-value 8.29E-09 1.68E-06

Intercept Lower 95% -4013.96 -3136.57

Intercept Upper 95% 2232.496 4074.69

Total Comments Lower 95% 7.499 5.709

Total Comments Upper 95% 12.585 11.581

4 DISCUSSION

This study found that there was no significant

relationship between the number of brand-related

comments and likes and the extent to which brand

names were mentioned in user comments. However,

the number of comments was positively and

significantly associated with sales performance

across both in-platform (Douyin) and external

(Taobao) channels. One possible explanation for this

phenomenon is that the KOLs selected for promotion

were not fashion-specific influencers but rather

beauty-oriented figures whose content tended to

focus more on personal image, emotional expression,

or general interaction, rather than on brand-specific

messaging. This likely introduced noise into the

dataset, thereby weakening the explanatory power of

the regression analysis. Although the presence of

brand mentions in comments could not be effectively

predicted by the volume of interaction, further

regression analysis revealed that promotional

activities—despite not being within a highly vertical

content domain—still had a significant and positive

effect on brand sales, both within and outside the

platform. These findings align with the research

results of Li and Xia (2022), who underscored the

critical role of interactivity in advertising

effectiveness, and also support Ge et al.’s (2021)

argument that, in the context of short video

advertising, “ engagement outweighs content. ”

Based on these insights, it is recommended that

brands guide users to focus their comments and

feedback more directly on the product itself, thereby

enhancing the brand relevance of the comment

content. Strengthening this relevance can improve the

predictive value of comments as indicators of sales

performance. Prior studies have shown that the

characteristics of opinion leaders — such as

professionalism, interactivity, and credibility — can

significantly boost consumer trust and engagement,

which in turn drive purchase intention (Lin & Huang,

2020; Feng, Sun, & Tang, 2025). Enhancing brand-

focused engagement is therefore instrumental in

strengthening brand associations and improving

conversion efficiency. In addition, the research

highlights that high-income consumers tend to be less

sensitive to price fluctuations during the decision-

making process (Wang, 2023). This demographic is

more inclined to pay a premium for high-quality,

trusted brands rather than being influenced solely by

price. However, the penetration of this consumer

segment in short video e-commerce remains

relatively limited. As such, brands should further

refine their audience segmentation strategies,

optimise content design, and adopt personalised

targeting approaches to more effectively reach these

high-potential but currently underserved consumer

groups.

5 CONCLUSION

This study finds that although the number of

comments is not significantly linearly associated with

whether the content of those comments refers to the

brand, it is nonetheless positively and significantly

PRMC 2025 - International Conference on Public Relations and Media Communication

646

correlated with sales performance across both in-

platform and external channels. This correlation was

particularly strong on Douyin, where the model

demonstrated a better fit. These results suggest that

the volume of comments may influence sales

regardless of whether the comments themselves

explicitly mention the brand. Moreover, the effect of

promotion is not limited to the platform on which the

content appears, indicating spill-over impact. The

research also highlights that high-income consumers

are relatively insensitive to price when purchasing

high-end fashion products, revealing a segment with

strong purchasing power that brands should

strategically target. These findings contribute to the

existing literature by addressing the underexplored

relationship between engagement metrics in short

video advertising and sales performance, offering

both theoretical insights and empirical evidence to

inform more refined content strategies and high-value

customer targeting in the context of digital marketing.

Nevertheless, this study is not without limitations.

The sample size was relatively small, and the analysis

did not account for the emotional tone or semantic

structure of user comments. Additionally,

multidimensional characteristics of KOLs— such as

gender, professional background, or content

specialisation—were not considered. Future research

could adopt natural language processing techniques

to explore sentiment and textual features of user

comments in greater depth. Expanding the sample

scope and incorporating diverse KOL characteristics

would further enhance the robustness and explanatory

power of the model. Moreover, future investigations

could explore the potential influence of virtual KOLs

and platform algorithms on consumer purchase

intentions.

REFERENCES

Chen, M.-H., & Liao, W.-Y. 2021. The influence of KOL

attributes on advertising effectiveness and purchase

intention—Using YouTubers as an example. Journal of

Marketing Research and Case Studies, 2021: Article ID

237451.

https://chihleeir.lib.chihlee.edu.tw/bitstream/31099330

0Q/3127/2

Feng, Y., Tang, L., & Sun, X. 2025. The influence of short

video marketing on consumers' purchase intention: The

mediating role of perceived trust. E-Commerce Letters,

14(2): 644–653. https://www.hanspub.org/journal/ecl

Gao, Y., & Zhang, X. 2023. Uncovering heterogeneous

prestige effect in luxury consumption: The role of

income. Journal of Business Research, 157: 113594.

https://doi.org/10.1016/j.jbusres.2023.113594

Ge, J., Wang, Y., & Jin, L. 2021. Effect of short video ads

on sales through social media: The role of

advertisement content generators. International Journal

of Advertising, 40(5): 710–735.

https://doi.org/10.1080/02650487.2020.1848986

Huang, L., & Xu, Y. 2024. Research on the Influence of

KOL Characteristics on Brand Identity and Purchase

Intention. Open Journal of Business and Management,

12(2): 415–426. Available at:

https://www.hanspub.org/journal/paperinformation.asp

x?paperid=94883

Lamberton, C., & Stephen, A. T. 2016. A thematic

exploration of digital, social media, and mobile

marketing: Research evolution from 2000 to 2015 and

an agenda for future inquiry. Journal of Marketing,

80(6): 146–172. https://doi.org/10.1509/jm.15.0415

Li, W., & Xia, Z. 2022. A study on the impact of social

interaction and media richness on the effectiveness of

short video advertising. Modern Business, (33): 84–86.

https://www.hanspub.org/journal/paperinformation?pa

perid=55996

Lin, Y., & Huang, Y. 2020. The impact of influencer

characteristics on consumer trust and engagement: The

mediating role of trust. Open Journal of Business and

Management, 8(5): 2134–2147.

https://www.hanspub.org/journal/paperinformation?pa

perid=37317

Ratchford, B., Soysal, G., Zentner, A., & Gauri, D. K. 2022.

Online and offline retailing: What we know and

directions for future research. Journal of Retailing,

98(1): 152–177.

https://doi.org/10.1016/j.jretai.2022.01.003

Tiago, M. T. P. M. B., & Veríssimo, J. M. C. 2014. Digital

marketing and social media: Why bother? Business

Horizons, 57(6): 703–708.

https://doi.org/10.1016/j.bushor.2014.07.002

Wang, C. 2023. Income level and consumer price

sensitivity: Evidence from digital retail platforms.

Economics Journal, 11(4): 89–102. https://economics-

journal.com/index.php/ej/article/view/2009/1910

Zhang, J., & Zhang, Y. 2024. The impact of content

characteristics of short-form video ads on consumer

purchase intention. Journal of Business Research. (In

press). https://doi.org/10.1016/j.jbusres.2024.11456

Key Drivers of Consumer Purchase Decisions in Short Video Platform Promotions

647