1

Impact of Online Sales Channels on Rural Household Income in Pakistan: Adoption

Barriers and Moderating Effects

Abstract
This study investigates the impact of online sales channels on rural household income in
Pakistan, focusing on adoption barriers and the moderating effects of digital literacy and
infrastructure access. Employing Propensity Score Matching (PSM) to mitigate selection bias,
the research analyzes quantitative data collected through structured questionnaires from rural
households in Punjab, Pakistan. Results indicate that online sales channels significantly and
positively influence household income, with households having higher digital literacy and better
internet and transportation connectivity experiencing greater income gains. The findings
underscore the potential of online sales channels as a key driver of rural economic growth while
highlighting barriers that hinder their effectiveness. In addition, the results highlight the need for
targeted policy interventions to maximize the benefits of online sales channels for far-flung
populations. Recommended strategies include educational programs to enhance digital literacy
and investments in infrastructure for greater accessibility. This study contributes to the growing
literature on digital inclusion and rural development, offering valuable insights for policymakers,
development organizations, and digital platforms providers aiming to foster economic growth in
remote areas.
Keywords: Online Sales Channels, Rural Livelihoods, Digital Literacy, Propensity Score
Matching (PSM), E-commerce Adoption, Income Disparity.
 2

1. Introduction
1.1 Background
Due to the progress of digital technology and internet services, there is great potential for
changing the traditional rural economy, and great strides are being made around the globe.
Research indicates that using online sales channels to generate new economic revenues and
stimulate the development of other industries is particularly valid for developing countries like
Pakistan, where a larger proportion of the population is engaged in rural farming and traditional
trading (Hayes, 2017). New media tools provide opportunities for rural producers to rely less on
the local intermediaries and expand their marketing access to urban and even global consumers
(Dwivedi et al., 2016).
Online sales platforms, such as e-shops (e.g., [Link]), social media markets (e.g., Facebook
Marketplace) and applications for agricultural goods, enable rural producers to establish direct
connections with buyers, bypassing intermediaries and gaining greater control over pricing and
sales. This approach facilitates better price determination and generally higher profit margins
(Abebe et al., 2016), particularly for agricultural and small-scale products which significantly
characterize rural production. According to the World Bank report (2016), utilizing online sales
channels to sell agriculture products from rural areas enhances price transparency, bypasses
intermediaries, and improves market efficiency. These channels offer a unique competitive
model, distinct from traditional local markets, which are often constrained by factors such as
geographical distance and associated overhead costs.
In Pakistan, e-commerce and online sales channels have drawn attention due to increased
internet connectivity and mobile phone ownership, even in rural areas (Ali et al., 2016).
Pakistan’s e-commerce sector experienced a 45% growth in 2021, compared to the global
average growth rate of 29% (Wattoo & Iqbal, 2022). This rapid expansion is providing
opportunities for new market structures and supporting the growth of existing sectors, with
projections indicating that this upward trajectory will accelerate in the coming few years (Imtiaz,
Ali, & Kim, 2020).
The purpose of this paper is to explore the impact of online sales channels on household
income in rural Punjab, Pakistan. Specifically, it evaluates whether adopting online sales
channels fosters economic growth by enhancing market connectivity, increasing revenue, and
 3

reducing reliance on intermediaries. Furthermore, the study aims to identify factors that hinder
the use of online sales channels, focusing on technological issues, infrastructural challenges and
cultural influences.
This research addresses a critical knowledge gap by examining the conditions required for the
successful implementation of online sales channels in rural environments. It evaluates both the
benefits and challenges of adoption (Abu-Bader, 2021). The findings offer significant insights
for policymakers, digital platform stakeholders, and non-governmental organizations (NGOs)
aiming to promote digitally-enabled rural economies. By addressing both the economic impacts
and the barriers to adopting digital technologies, this study contributes to the existing knowledge
of how such innovations can help remodel rural economies in developing countries like Pakistan.
However, significant challenges remain in establishing the required internet infrastructure for
supporting these sales channels. Most rural households lack consistent internet access and are
reluctant in adopting new technologies, which are necessary for participating in e-commerce
(Nazir & Roomi, 2021). In addition, sociocultural barriers may pose additional challenges, as
individuals in rural settings often avoid risk due to traditional beliefs and lack experience in
using advanced tools required for online transactions (Hendricks & Mwapwele, 2024).
Overcoming the above stated barriers and promoting web-based selling platforms can
significantly enhance the quality of life in rural areas. Main benefits include increased household
income, reduced prices for goods and services, and higher chances for financial independence
(McCluney et al., 2019). Therefore, the experience of implementing online sales channels in
rural Pakistan can serve as an example for similar regions in other developing countries.
1.2 Research Questions
The following research questions will guide the analysis:
RQ1: How do online sales channels impact the income levels of rural households in
Pakistan?
RQ2: What are the significant barriers to the adoption of online sales channels by rural
households in Pakistan?
Answering these questions will enable the study to provide insights into the impact of new
online sales channels on improving the income and living standards of the rural
populations while identifying the factors that may hinder their adoption. These findings
 4

will assist policymakers and other stakeholders in designing targeted intervention

approaches to promote rural integration of digital innovations (DIGI) and drive overall
economic growth from an econometric perspective.
2. Literature Review
2.1 Online Sales Channels and Rural Livelihoods
Expanding sales channels to market agricultural products has been shown to enhance the rural
household incomes. According to Siegel (2017), the development of online selling platforms
enables rural producers to access broader, more competitive markets, extending beyond
conventional local trading circuits. These platforms help rural producers to get connected with
urban and international clients, thus increasing sales and profit margins. Aker et al. (2016) and
Deichmann et al. (2016) emphasize that digital technologies can address the issue of income
inequality by empowering rural producers. By utilizing the internet, these producers benefit from
better price discovery and reduced reliance on middlemen, resulting in lower transaction costs
and greater value capture. This enhanced value addition plays a significant role in achieving
economic viability in rural economies. Israr (2017) highlights that online sale channels have
transformed market access in many developing countries. For instance, e-commerce linked
through B2B relationships has allowed producers from rural regions in India, Kenya, and parts of
Latin America to penetrate into the urban markets and earn better incomes (Yadav et al., 2021).
Additionally, research carried out by Li et al. (2024) in China revealed that participation in e-
commerce enhanced the developmental resilience of farming households by 9.63% compared to
non-participation, with economic, social, and cultural resilience improving by 9.29%, 9.84%, and
9.92%, respectively. Furthermore, over one-third of the customers of BliBli, a popular online
shopping platform in Indonesia with 3 million users, were from rural areas (Kshetri, 2018).
A key advantage of this online sales model is that it allows rural producers to move beyond
traditional sales methods (Karine, 2021). Farmers can procure seeds and fertilizers from
independent distributors directly and sell products to the final consumer or other dealers without
intermediaries, thus earning higher commissions on the selling price (Dearing & Cox, 2018).
Qualitative studies in Pakistan suggest that similar gains can be achieved through online sales
intermediaries. For instance, existing firms that currently sell their agricultural products to low-
 5

cost channel intermediaries could achieve better returns by selling directly to consumers via the
internet (Tim et al., 2021).
However, the adoption of online sales channels remains limited in many rural settings due to
challenges related to digital technologies, internet access, and cultural barriers. Understanding
these challenges is crucial for designing effective policies and programs to improve the rural
digital commerce (Akinola, 2017).
2.2 Propensity Score Matching (PSM) method in Social and Economic Studies
PSM (Propensity Score Matching (PSM) method) is commonly used in social and economic
studies to match groups based on observed characteristics, particularly in technology adoption
studies to analyze differences between adopters and non-adopters (Puttitanatkul & Pan, 2018).
In rural economic research utilizing cross-section data, PSM has been shown to significantly
reduce selection bias in treatment participation. For instance, in the case of online selling
platforms, treatment participation often depends on indicators such as education level, income, or
technology access (Chen et al., 2023). By matching prospective adopters with similar non-
adopters, researchers can isolate and evaluate the effect of online sales channels adoption on
factors such as the household income, minimizing the influence of confounding factors.
In addition, PSM has been used in Indian and Brazilian settings to assess the impact of digital
platforms on rural economies, revealing that the incomes of adopters are significantly higher than
those of similar non adopters (Rosenbaum & Rubin, 2023). Thus, due to its ability to quantify
the economic impact of online sales channels effectively, PSM is considered suitable to be
applied in this study, particularly as it aligns well with the rural context.
2.3 Moderation Effects on Technology Adoption
The effectiveness of online sales channels in improving rural livelihoods is influenced by several
moderating factors, including digital literacy, infrastructure accessibility, and sociocultural
factors. These variables shape how effectively rural households can adopt and utilize digital
platforms for economic activities (Mitzner et al., 2019).
First, 2.3.1 Digital Literacy
Rresearch depicts that individuals with higher digital literacy levels are more likely to use
technologies effectively. A survey revealed that participants’ digital literacy enhanced not only
 6

their confidence in managing various technological devices but also their ability to effectively
make actual e-commerce purchases (Lee & Acharya, 2022).

Second, 2.3.2 Infrastructure Accessibility

Ananother critical factor is the availability of infrastructure that supports online access, such
as internet connections and transport (Estriegana et al., 2019). For instance, studies in Southeast
Asia demonstrate that a lack of access to adequate infrastructure significantly hampers the
functionality of digital platforms, particularly in rural regions. Without the availability of such
infrastructure, the adoption of online sales channels becomes challenging and less effective.

Third, 2.3.3 Sociocultural Factors

Ttechnology adoption is also affected by sociocultural factors, such as perceived community
norms and trust in electronic transactions (Chittipaka et al., 2023). In some instances, extended
rural families have struggled to adapt quickly to technological changes, mainly due to the fear of
privacy loss and unfamiliarity with social networking. These sociocultural barriers signal the
need for tailored approaches to foster acceptance and trust in digital technologies within rural
communities.
3. Theoretical Background and Hypotheses
Online sales platforms and rural livelihoods are becoming increasingly interconnected, especially
in developing countries like Pakistan where traditional sales methods often limit market access
and income generation. This section explores the key theoretical paradigms that form the basis
for the currentcurrent research, and provideprovides a framework for understanding the
relationship between online sales channels and rural economic development.
3.1 Technology Acceptance Model (TAM)Impact Mechanisms of Online Sales Channels on
Rural Household Income
Combine relevant theories (perhaps the ones you mentioned) to explain the impact mechanism
of online sales channels on rural household income, rather than listing existing theories.
This analysis proposes the following hypothesis:
H1: Households adopting online sales channels will have significantly higher monthly income
compared to non-adopters.
 7

3.2 Mediating Effect of Digital Literacy

explaining the mediating effect of digital literacy……
This analysis proposes the following hypothesis:
H2: The relationship between online sales channels adoption and household income is
positively moderated by digital literacy.
3.3 Mediating Effect of Access to Digital Infrastructure
explaining the mediating effect of access to digital infrastructure……
This analysis proposes the following hypothesis:
H3: The relationship between online sales channels adoption and household income is
positively moderated is positively moderated by access to digital infrastructure.

The Technology Acceptance Model (TAM), introduced by Davis (1989) and further
extended by later scholars like Hjorth (2016), is vital in understanding the factors that
influence the acceptance of new technologies. TAM proposes that perceived ease of use and
perceived usefulness are key factors in determining the individual’s intention to adopt new
technology. In the case of rural Pakistan, such factors are very useful in determining how
rural households perceive online sales channels. For instance, the ease of use of a platform
and its ability to increase profitability can substantially impact its adoption. This
theoretical tool offers a plausible explanation for the differences in technology adoption by
rural communities and emphasizes the necessity for user-centered design and educational
programs aimed at the promotion of online sales platforms.
3.2 Diffusion of Innovations Theory
Rogers’ (1962) Diffusion of Innovations Theory, further refined by Dearing (2018),
analyzes the process through which innovations are transformed into new technologies in
the society. Relative advantage, compatibility, complexity, and observability are the key
factors that significantly influence the adoption rate. In the case of rural Pakistan, the
adoption of online sales channels is determined by their perceived benefits over
conventional sales methods, sociocultural alignment, and ease of integration into daily
tasks. This theory provides a basis for understanding the barriers and facilitators for
 8

adoption, such as digital literacy levels and availability of infrastructures, thus giving an
insight into the more complex issues faced by the rural households.
3.3 Digital Livelihoods Framework
Digital Livelihoods Framework, as outlined by Oberlack (2016), highlights the
transformative potential of digital technologies in improving rural income. This framework
stresses the importance of online sales channels in income diversification, increased market
penetration, and livelihood protection. In the case of Pakistan’s rural households, these
platforms help access larger markets, reduce dependency on middlemen and provide
opportunities for sustainable income growth. By integrating this framework, the study
provides a comprehensive approach on the socio-cultural advantages of online sales
channels, while also addressing contextual factors such as digital literacy and
infrastructure that moderate these advantages.
3.4 Integrative Perspective
These theoretical models collectively offer an integrative framework for examining the
impact and adoption of online sales channels on rural livelihoods. TAM focuses on
adoption behavior at individual levels; the Diffusions of Innovation Theory emphasizes the
societal and systemic factors influencing adoption, and the Digital Livelihoods Framework
links technology adoption to broader economic outcomes. Together, these paradigms
provide a comprehensive basis for exploring how online sales platforms can drive rural
economic development, identify barriers to adoption, and suggest actionable policy
interventions.
4. Hypothesis
This study proposes the following hypotheseis based on the theoretical framework discussed
above:
H1: Households adopting online sales channels will have significantly higher monthly
income compared to non-adopters.
H2: The relationship between online sales channels adoption and household income is
positively moderated by digital literacy.
H3: The relationship between online sales channels adoption and household income is
positively moderated is positively moderated by access to digital infrastructure.
 9

54. Research Design

54.1 Study AreaData Source

This research focuses on the livelihood outcomes of selected rural households that have adopted
online sales channel in three districts of Punjab, Pakistan: Lahore, Kasur and Sheikhupura. These
districts were selected purposely considering the variations in the current status of digital
development and adoption of digitization tools. While Lahore represents a better connected peri-
urban consumers’ example, Kasur and Sheikhupura sites depict relatively typical agri-oriented,
less digitally connected rural consumers.
5.2 Sampling Methodology

We employed a stratified random sampling method to select farms based on criteria such as
farm size, type of agriculture (crops or livestock), and level of digital [Link] stratified
random sampling approach selected farms based on criteria such as farm size, type of agriculture
(crops or livestock), and level of digital experience. This method was designed to ensure
findings could be generalized to both adopters and non-adopters of online sales channels. A
total of 900 households were sampled, including 10% adopters (90 households) and 90%
non-adopters (810 households). While random sampling was conducted within clusters, the
clusters themselves were purposively selected based on geographic and administrative
boundaries. This approach combines purposive and random sampling methods to ensure
representation and address the study'’s specific objectives.

5.3 Data Collection

5.3.1 Quantitative Data Collection
Our data includes both quantitative and qualitative data. On the one hand, Qquantitative data
was obtained through structured questionnaires covering household composition, monthly
income (current and prior to adoption, where applicable), education, access to infrastructure, and
digital literacy. According to the survey findings, incomes in this sample area were higher than
assumed, with adopters having average incomes ranging between PKR 37,500 and PKR 47,500
and non-adopters between PKR 30,000 and PKR 40,000. Additionally, 18% of adopters reported
prior exposure to digital tools, compared to 10% of non-adopters. Adopters also had, on average,
 10

15% larger landholdings and reported a 12% higher rate of mobile internet usage compared to
non-adopters. These values provide a realistic reference point for evaluating the gains from
online sales channel adoption.
On the other hand, 5.3.2 Qualitative Data Collection
Iin-depth interviews were conducted to explore barriers to adoption, perceived benefits, and
sociocultural influences. The qualitative data were analyzed to identify themes such as digital
skills gaps and resistance to technology, providing valuable context to complement the
quantitative findings. Qualitative data analysis included coding responses from interview
transcripts, with specific attention given to identifying themes such as limited digital skills and
sociocultural resistance to technology. For example, 15% of non-adopters explicitly cited
insufficient digital proficiency, while sociocultural resistance—reported in 12% of responses—
was assessed through recurring themes in discussions about community norms and perceived
risks of technology adoption. These insights add contextual richness to the quantitative findings,
highlighting key non-economic barriers to adoption.

45.52 Defining Variables for Analysis

The core variables in this study include online sales channels and household income; each
defined and measured to provide precise and meaningful insights. Online sales channels are
digital platforms that facilitate transactions and market engagement. Examples include e-
commerce websites like [Link], social media marketplaces such as Facebook Marketplace,
mobile apps tailored for agricultural products, and digital payment gateways. This variable is
measured using survey responses that identify households actively utilizing these platforms. It
has a binary form to indicate whether or not the firm adopted online sales channels, with 1
representing adoption and 0 otherwise. Such a categorization avoids confusion and helps in
assessing the adoption trend.
Household income means the total monthly income from sources within and outside
agriculture, of a given household. The measurement entails the use of structured questionnaires
in estimating the amount of EMS exposure from the respondents. Where there is adoption of
online sales channels, both pre and post adoption income is captured in order to analyze the
economic effects. The data is represented as a continuous variable in Pakistani Rupees (PKR) to
provide better statistical analysis.
 11

Digital infrastructure is the other determinant and is represented by access to resources, such
as the internet and quality mobile networks. It is a crucial factor for enabling the rural
households to engage in digital markets (McCluney, 2019). In this study, digital infrastructure is
measured in terms of a digital index varying from 1 (Low) to 10 (High). In addition, this index
includes parameters such as internet reliability, availability of mobile network coverage, and
access to digital devices, thereby providing a thorough assessment of the environment necessary
for effective online selling. The Digital Infrastructure Index was developed as a composite
measure to assess the accessibility of digital resources, ranging from 0 to 10, with higher scores
indicating better infrastructure. It comprised three key components: internet access (40%
weight), mobile network coverage (30% weight), and digital devices (30% weight). Internet
access was assigned the highest weight because of its importance in facilitating digital inclusion
with a scale of 0 for no internet access to 10 for reliable high-speed internet access. Mobile
network coverage estimated signal quality on a scale between 0 and 10 with no and strong
constant signals respectively and was assigned a total of 30% due to its roles in enabling digital
communication. Digital devices section, that also composed 30% of the overall assessment,
assessed the accessibility and quality of devices, which ranged from 0 for no devices to 10 for
multiple and updated devices. The final index was obtained by summing up the weighted score
for each index components, which gives a comprehensive and replicative means of estimating
digital infrastructure and its effectiveness with regards to the adoption of online sales.
Digital Literacy refers to the digital skills and proficiency of a household, measured on a
scale of 0–10. It reflects the individual’s readiness to engage with technology and directly
influences their perceived ease of use and usefulness, as outlined in the Technology Acceptance
Model (TAM). Digital literacy determines a user's ability to effectively utilize online sales
platforms, playing a crucial role in the successful adoption and the benefits derived from online
engagement.
Digital Technologies refers to the use and integration of applications, such as e-shops, mobile
financial services, and digital marketing tools. According to Oberlack (2016), this variable
should be measured through a technology use index that considers both the number and richness
of technology applications. This approach offers a more comprehensive assessment of the extent
to which households utilize digital instruments to enhance their economic activities. The scoring
 12

methodology of “Digital Technologies” takes into account the perceived importance and the
frequency of use of each of the examined technologies in enabling online sales. Those
technologies such as payment gateways, and e-commerce platforms, which are most important
for transaction processes, were weighted more, 50% of the total score. Marketing and analytics
tools, which are supportive technologies including social media, made up 30% while
supplementary technologies including communication apps made up 20%. Representative scores
were obtained by multiplying the proportion of each technology by its relative importance and
then normalizing the overall score to a 0–10 range to equally measure adoption and usage.
All key variables in Table 1 are measured using context-specific indicators to ensure accuracy
and relevancy. For instance, monthly household income is evaluated on an ordinal scale, with
adopters earning PKR 42,500 compared to PKR 35,000 for non-adopters. Other variables, such
as age, education level, family size, and digital literacy are captured as interval/ratio variables.
Digital literacy is measured on the 0-10 scale; highlighting participants’ competence in using the
digital platforms. Adopters reported an average score of 7.0, compared to 4.5 for non-adopters.
These measurements establish a solid foundation for assessing the managerial and sociocultural
factors driving the adoption of online sales channel.
5.6 Explanation of Key Calculations
5.6.1 The Digital Infrastructure Index was developed as a composite measure to assess the
accessibility of digital resources, ranging from 0 to 10, with higher scores indicating better
infrastructure. It comprised three key components: internet access (40% weight), mobile network
coverage (30% weight), and digital devices (30% weight). Internet access was assigned the
highest weight because of its importance in facilitating digital inclusion with a scale of 0 for no
internet access to 10 for reliable high-speed internet access. Mobile network coverage estimated
signal quality on a scale between 0 and 10 with no and strong constant signals respectively and
was assigned a total of 30% due to its roles in enabling digital communication. Digital devices
section, that also composed 30% of the overall assessment, assessed the accessibility and quality
of devices, which ranged from 0 for no devices to 10 for multiple and updated devices. The final
index was obtained by summing up the weighted score for each index components, which gives a
comprehensive and replicative means of estimating digital infrastructure and its effectiveness
with regards to the adoption of online sales.
 13

5.6.2 Digital Technologies: The scoring methodology of “Digital Technologies” takes into
account the perceived importance and the frequency of use of each of the examined technologies
in enabling online sales. Those technologies such as payment gateways, and e-commerce
platforms, which are most important for transaction processes, were weighted more, 50% of the
total score. Marketing and analytics tools, which are supportive technologies including social
media, made up 30% while supplementary technologies including communication apps made up
20%. Representative scores were obtained by multiplying the proportion of each technology by
its relative importance and then normalizing the overall score to a 0–10 range to equally measure
adoption and usage.
5.6.3 Income Difference (PKR): To address potential endogeneity, the outcome measure is
based on the ATT, which estimates the causal impact of adoption on income. This approach uses
the Average Treatment Effect on the Treated (ATT) to focus on the impact of adoption among
those who have actually adopted the technology, while controlling for covariates such as
education, age, and digital literacy. By matching adopters with non-adopters through PSM, this
method reduces bias and provides robust estimation of μ, representing the income difference
attributed to adoption.

Table 1: Variables and Calculations

Variable Description Type Mean ± SD Mean ± SD Calculation Notes

(Adopters) (Nonadopters)
Monthly Income Household Continuous 42,500 ± 35,000 ± 2,500 Reported in survey
(PKR) monthly income 3,000 responses
Age (years) Age of Continuous 36 ± 7 42 ± 6 Direct survey
household head measurement
Education Level Years of formal Continuous 10 ± 2 8±3 Total years of formal
(years) education education
Family Size Number of Continuous 5±2 5 ± 1.5 Total individuals in
individuals in the the household
household
Digital Accessibility to Composite 6.5 ± 2.5 4.2 ± 1.8 Weighted scoring of
Infrastructure the Internet and Index responses
devices
Digital Number of Count 3±1 Not applicable Usage of tools like
Technologies technologies apps, platforms
used for online
 14

sales
Digital Literacy Proficiency in Continuous 7.0 ± 2.0 4.5 ± 1.5 Survey-based scoring
Rate using digital (0–10 scale) reflecting respondents'
tools ability to use digital
platforms
Adoption Status Whether a Binary (1 = 1 0 Derived from survey
household adopts Yes, 0 = No) responses
online sales
channels
Income Difference in Continuous 7500 Derived from
Difference monthly income regression and ATT
(PKR) between adopters calculations

Note: Continuous variables are reported as mean ± standard deviation. Binary variables
indicate proportions or yes/no responses. Composite indices are based on weighted scoring
systems for infrastructure and digital accessibility. Observations vary based on availability and
response completeness.

5.74.3 Propensity Score Matching (PSM) method

To control for selection bias between the adopters and non-adopters of online sales channels in
this study, PSM test was used. Applying the logistic regression model, it was possible to estimate
the probability of household adopting online sales with reference to several covariates.
Cramming down coefficients, logistic regression was used to predict the likelihood of each
household utilizing online sale channels based on predictors like education level, size of
household, infrastructure access, and digital literacy levels. The logistic regression model for
propensity scores (P(X) is specified as follows:

Where;
 P(X) is the probability of adopting online sales channels (treatment),
 X 1, X 2 ,…, X n represent the covariates used for matching (e.g., education level, household
size),
 β 0 is the intercept, and β 1,…, β n are the coefficients for each covariate.
 15

A caliper value of 0.05 was selected for nearest-neighbor matching, following both practical
considerations and established recommendations in the literature. This value is commonly used
as a standard to ensure close matches while minimizing selection bias (Rosenbaum & Rubin,
1985). The caliper restricts the maximum absolute difference in propensity scores between
matched adopters and non-adopters, ensuring that the matched pairs are similar in measurable
characteristics. Restricted to this range, the caliper minimizes the possibility of low quality
matching, which could negatively impact the accuracy of treatment effect estimation.
To evaluate matching efficiency and ensure covariate balance between adopters and non-
adopters, the standardized mean differences (SMDs) of each covariate were assessed before and
after matching. The results showed that matching significantly enhanced the balance. Before
matching, the mean of mean absolute difference (MMD) for education level was 0.35, household
size was 0.28, digital infrastructure was 0.40, and digital literacy was 0.32, which were above the
recommended MMD of 0.1. However, after matching, SMDs for all covariates reached below
the threshold – education level (SMD = 0.04), household size (SMD = 0.06), digital
infrastructure (SMD = 0.08) and digital literacy (SMD = 0.05). These results demonstrate
improved balance, supporting the suitability of the 0.05 caliper in generating comparable
matches and ensuring robust estimation of treatment effects.
The Average Treatment Effect on the Treated (ATT) was estimated as the mean income
difference between matched adopters and non-adopters:

ATT =E[Y i (1)−Y i (0)∣T i=1]

Where:
 Y i (1) is the outcome (household income) for the treated household (adopters).
 Y i (0) is the counterfactual outcome (what household income would have been
without adoption).
 T i=1 indicates households that adopted online sales channels.
This calculation isolated the effect of Adoption on income, controlling for confounding factors.
5.84.4 Moderation Analysis
 16

To assess how variables such as digital literacy and digital infrastructure moderate the
relationship between online sales adoption and income, interaction terms were included in the
regression model. The key variables used in the Moderation Effect Model were adoption of
online sales channels, digital literacy, digital infrastructure, age and education. The equation for
the model is as follows:
Income = α + β 1 (Adoption) + β 2 (Moderator) + β 3 (Adoption × Moderator) + ϵ
Where:
 Income is the dependent variable representing household income,
 Adoption is a binary variable indicating online sales channel adoption (1 for
adopters, 0 for non-adopters),
 The moderator represents the moderating variables (digital literacy or
infrastructure access),
 Adoption × Moderator is the interaction term capturing the moderating effect,
 α is the intercept, β 1, β 2, β 3 are coefficients and ϵ is the error term.

Interaction plots and regression coefficients for β 3 were used to establish the significance of the
moderation effects. The direction of the moderation would show whether higher digital literacy
or better Infrastructure enhances or dampens the influence of online sales adoption on income.
Interaction plots were created by plotting predicted values of household income across different
levels of the moderator variables (digital literacy and digital infrastructure). Predicted values
were calculated using the regression model, holding all other covariates constant at their mean
values. This approach illustrates how the relationship between online sales adoption and income
changes depending on varying levels of the moderators.
54.4 Analytical Techniques

The analytical techniques employed in this study include regression analysis and interaction
modeling to examine the relationships between key variables and their impact on household
income. Interaction terms are incorporated into the regression model to analyze how digital
infrastructure and digital technologies moderate the effects of adopting online sales channels on
household income. For instance, the term Adoption × Digital Infrastructure assesses whether
better infrastructure amplifies the income benefits of online sales. Similarly, Adoption × Digital
 17

Technologies evaluates the impact of digital tools on economic outcomes. These interaction
terms are important for aligning the analysis with the study's objectives, ensuring a precise
understanding of how digital infrastructure and literacy contribute to household income through
online sales adoption.
Furthermore, the inclusion of regression models with interaction terms enables the study to
explore variations in results. For example, households with access to better digital infrastructure
experience significantly greater income gains from adopting online sales channels. This
observation aligns with De Schutter’s (2017) emphasis regarding the role of enabling
environments in promoting rural development.

65. Analysis and Results

6.1 Empirical Results

6.1.1 Income Differences between Adopters and Non-Adopters
The study reveals that households adopting online sales channels earn significantly higher
monthly incomes than those relying on traditional methods. Adopters report an average monthly
income of PKR 42,500, while non-adopters earn PKR 35,000. According to Aker and Cariolle
(2023), digital tools like online sales platforms enhance market access and price negotiation
capabilities, which are critical for increasing rural household incomes. This significant disparity
(p < 0.001, Cohen’s d = 1.04) underscores the economic advantages of embracing e-commerce in
rural settings.
6.1.2 Role of Digital Infrastructure
Adopters score significantly higher on the digital infrastructure index (6.5) compared to non-
adopters (4.2), with a large effect size (d = 0.98). Deichmann et al. (2016) emphasize that access
to reliable internet and mobile networks is essential for facilitating the adoption of online sales
channels. These findings suggest that strengthening digital infrastructure in rural areas can
amplify the benefits of e-commerce, a result also supported by Duncombe (2016).
6.1.3 Role of Digital Literacy
Digital literacy emerges as a critical factor in technology adoption. Adopters demonstrate higher
proficiency (mean = 7.0) compared to non-adopters (mean = 4.5), reflecting a significant positive
association (p < 0.001, Cohen’s d = 1.02). According to Estriegana et al. (2019), digital literacy
 18

not only improves technology acceptance but also enhances user confidence in navigating digital
platforms. This result highlights the importance of targeted interventions to boost digital skills in
rural populations.
6.1.4 Socio-demographic Influences
The study indicates that younger individuals (mean age = 36 years) and those with higher
education levels (mean = 10 years) are more likely to adopt online sales channels. These findings
are consistent with Chittipaka et al. (2023), who argue that younger, educated individuals are
more open to adopting digital innovations. However, family size does not appear to significantly
influence adoption decisions, reinforcing the idea that individual and infrastructural factors play
a more prominent role.
6.1.5 Advanced Statistical Tests
The study employed advanced statistical techniques to compare adopters and non-adopters of
online sales channels, enhancing its conclusions. Independent sample t-tests assessed differences
in predictor variables like monthly income, education, and digital infrastructure. When normality
assumptions were not met, the Mann-Whitney U test was used as a non-parametric alternative,
providing statistical significance at the 5% level. This test compares medians and is less sensitive
to outliers, ensuring valid comparisons. Cohen's d indicated large practical differences in income
(1.04) and digital infrastructure (0.98). Multiple regression analysis considered interactions like
Adoption × Education, while PSM matched groups on age and education. These methods
strengthened the study’s conclusions with robust statistical analysis.

Table 2: Descriptive Statistics for Adopters and Non-adopters

Variable 95% CI 95% CI (Non- p-value Effect Size

(Adopters) adopters) (Cohen’s d)
Monthly Income (PKR) [41,880, 43,120] [34,470, 35,530] < 0.001 1.04
Age (years) [34.9, 37.1] [40.9, 43.1] < 0.001 0.91
Education Level (years) [9.7, 10.3] [7.6, 8.4] < 0.001 0.80
Family Size [4.7, 5.3] [4.6, 5.1] 0.540 0.00
Digital Infrastructure [6.0, 7.0] [3.8, 4.6] < 0.001 0.98
Digital Literacy Rate [6.7, 7.3] [4.2, 4.8] < 0.001 1.02

Note: CI = Confidence Interval; p-values indicate statistical significance levels, where p < 0.05
denotes significance and p < 0.001 indicates highly significant results. Cohen’s d represents the
 19

magnitude of effect sizes: 0.2 = small, 0.5 = medium and 0.8 = large. For example, Monthly
Income (Cohen’s d = 1.04) shows a large and statistically significant difference between
adopters and non-adopters, suggesting income disparity as a strong factor influencing adoption.

Figure X

6.2 Correlation Matrix

The correlation matrix helps identify relationships among variables, mainly focusing on income
relationship with age, education, family size, digital infrastructure and digital literacy. The
correlation matrix reveals several significant relationships that align with the study's objectives.

Interpretation:
Monthly income in Table 3 shows a strong positive correlation with digital literacy (r=0.6r =
0.6r=0.6) and digital infrastructure (r=0.5r = 0.5r=0.5), emphasizing the critical role of digital
capabilities and infrastructure in boosting income levels. The positive correlation between
education level and income (r=0.4r = 0.4r=0.4) supports the notion that higher educational
attainment is associated with better earning potential. Income has a weak negative correlation
with age (r=−0.3r = -0.3r=−0.3), suggesting that younger individuals, particularly adopters of
online sales channels, may have higher incomes. This aligns with the study's focus on the digital
economy, where younger, tech-savvy individuals are likely to thrive. Family size shows a
slightly inverse relationship with income (r=−0.2r = -0.2r=−0.2), indicating minimal impact on
 20

income levels. These findings collectively highlight the importance of digital literacy,
infrastructure, and education in shaping economic outcomes, reinforcing the study's emphasis on
the role of online sales adoption in income generation.

Table 3: Correlation Matrix

Variable Monthly Age Education Family Digital Digital

Income Level Size Infrastructure Literacy
Monthly Income 1.0 -0.3 0.4 -0.2 0.5 0.6
Age -0.3 1.0 -0.1 0.1 -0.2 -0.1
Education Level 0.4 -0.1 1.0 -0.1 0.3 0.4
Family Size -0.2 0.1 -0.1 1.0 -0.1 -0.1
Digital Infrastructure 0.5 -0.2 0.3 -0.1 1.0 0.6
Digital Literacy 0.6 -0.1 0.4 -0.1 0.6 1.0

Note: Correlation coefficients (r) range from -1 to +1, indicating the strength and direction of
relationships. Values closer to ±1 imply stronger associations, while 0 suggests no correlation.
Statistically significant relationships include Monthly Income with Education Level (r = 0.4),
Digital Infrastructure (r = 0.5), and Digital Literacy (r = 0.6), highlighting their positive
influence on income levels. Negative correlations, such as Age with Monthly Income (r = -0.3),
suggest older respondents tend to have lower income levels. Relationships with r ≥ 0.3 are
considered moderate, and r ≥ 0.5 are strong.

6.3 Propensity Score Matching (PSM) Results

Propensity Score Matching (PSM) was applied to create a balanced comparison between
adopters and non-adopters of online sales channels based on key covariates, including age,
education level, digital infrastructure, digital literacy, and family size. By controlling for these
factors, the analysis isolates the effect of adoption on monthly income, enabling a more accurate
estimation of income variation attributable to online sales channel adoption.
6.3.1 Matching Procedure
The matching employed the logistic regression of the propensity score, which gives the
probability of a given household carrying out its business using an online sales channel
depending on covariates such as age of the household, education level, digital infrastructure
access, digital literacy level and number of family members.
6.3.2 PSM Balance Check
 21

Standardized Mean Difference (SMD) measures the extent of covariate balance between
treated and control group. SMD value smaller than 0.1 post-matching is considered ideal,
depicting excellent balance. In this study, Propensity Score Matching (PSM) effectively
enhanced the comparability between adopters and non-adopters of online sales channels by
controlling for main demographic characteristics such as age, education level, family size, digital
environment, and IT literacy. Pre-matching SMD values in Table 4 ranged from 0.10 to 0.55,
depicting moderate level of imbalance for certain covariates. However, post-matching, all SMD
values were below the 0.1 threshold, showing excellent balance across the groups. This
improvement confirms the robustness of PSM approach, ensuring that the estimated effects of
adoption of online sales channels on income are not confounded by pre-existing differences.
Variance Ratio (VR) evaluates similarity in variability between groups. VR nearing 1 indicates
balanced variability across covariates in both matched groups. Post-matching VRs in Table 4
approached 1.0, depicting improved homogeneity between adopters and non-adopters. This
balance reduces the probability of interaction effects and strengthens the credibility of observed
results. Attaining VRs near to 1 depicts that the matching process successfully controlled for the
confounding factors, increasing the reliability and validity of the study’s outcomes.
P-values measure the statistical significance of differences between groups. Post-matching,
p-values > 0.05 for all covariates indicate no statistically significant differences remain,
thus validating the effectiveness of the matching process.
Key Interpretation
After matching, all covariates in Table 4 achieved SMDs below the 0.1 threshold, with p-values
exceeding 0.05. Furthermore, VRs approached 1, confirming that the matching process
successfully balanced the groups. These results ensure a fair comparison for estimating the
treatment effect of online sales adoption on household income.
6.3.3 Propensity Score Distribution Plot
The propensity score distribution plot shows the overlap between adopters and non-adopters
before and after matching. Before matching, the distributions for the two groups are often
distinct, reflecting baseline differences. After matching, an increased overlap in distributions
indicates improved covariate balance and reduced selection bias.
 22

Interpretation: The overlapping distributions observed post-matching confirm that the matched
sample is balanced, making it suitable for comparison. This ensures the credibility of conclusions
regarding the treatment effect of online sales adoption on household income.

Table 4: PSM Balance Check

Variable Pre- Post- Pre- Post- p-value p-value

Matching Matching Matching Matching (Pre) (Post)
SMD SMD VR VR
Age 0.30 0.05 1.15 1.02 < 0.001 0.780
Education Level 0.45 0.08 1.25 1.05 < 0.001 0.560
Family Size 0.10 0.03 1.05 1.00 0.080 0.920
Digital Infrastructure 0.50 0.07 1.35 1.10 < 0.001 0.610
Digital Literacy 0.55 0.06 1.40 1.08 < 0.001 0.590

Note: SMD = Standardized Mean Difference; values < 0.1 indicate excellent covariate balance.
Variance ratios near 1 suggest comparable variability between groups. Post-matching p-values
> 0.05 confirm no statistically significant differences, enhancing result validity and
comparability between groups.

6.3.44 Average Treatment Effect on the Treated (ATT)

 23

Objective: The ATT measures the average monthly income difference between adopters and
non-adopters of online sales channels after matching, isolating the specific effect of adoption.
Interpretation: The results in Table 5 indicate that households adopting the online sales
channels, on average, experience an increase of PKR 10,000 in their monthly income compared
to non-adopters. This result confirms the substantial positive effect of online sales channel
adoption on rural household income. These findings align with previous literature, which shows
that digital sales platforms improve income levels by allowing access to broader markets and
diversifying income sources. These findings are consistent with the hypothesis that adopting
online sales channels provides significant economic benefits to the rural households, thus,
allowing them to improve their financial situation.

Table 5: Average Treatment Effect on the Treated (ATT)

Group Monthly Income (PKR)

Adopters 45,000
Non-adopters 35,000
ATT 10,000

Note: The ATT measures the average monthly income difference between adopters and non-
adopters after matching. At PKR 10,000, adopters demonstrate significantly higher income,
confirming the positive impact of online sales adoption on household income and well-being.
 24

6.5 4 Moderation Analysis Results

Objective: Moderation analysis evaluates the extent to which specific variables influence the
relationship between online sales channel adoption and monthly income. In this study, the
moderating effects of digital Infrastructure and digital literacy are examined, as they are
critical factors enabling digital adoption in rural settings. Additionally, education level and age
are analyzed as secondary moderators to offer additional insights.

6.45.1 Digital Infrastructure as a Moderator

The regression model included an interaction term (Adoption × Digital Infrastructure) to analyze
the moderating effect of digital Infrastructure.
Explanation of Results
Significance: All p-values are highly significant (p<0.001), indicating strong effects.
Interaction Effect: The interaction term depicts that household with better digital infrastructure
experience substantially higher income benefits from adoption. Improved internet connectivity
and access to mobile devices amplifies the economic impact of online sales channels.

Interpretation of Digital Infrastructure as a Moderator

The study demonstrates that digital Infrastructure significantly enhanced the impact of online
sales channel adoption on household income. In particular, the results of the interaction term in
Table 6 indicate that for each unit improvement in the digital environment, the net income
advantage increases by PKR 1,500 for households adopting online sales platforms, including
better internet connection and more access to mobile devices. Since digital or Internet
connectivity is an important driver for leveraging online sales as a marketing strategy, customers
with access to superior digital networks can derive higher economic benefits through online sales
channels. Dorward et al. (2005) emphasize that in regions with infrastructural constraints, the
expansion of the internet and mobile networks can potentially elevate income levels. Therefore,
robust digital infrastructure is critical for supporting online sales platforms and boosting income-
generating opportunities.

Table 6: Moderation Analysis Results for Digital Infrastructure

 25

Variable Coefficient p-value Interpretation

(β)
Adoption Status 7,500 < 0.001 Adopting online sales channels increases
income by PKR 7,500.
Digital 1,000 < 0.001 Each unit increase in infrastructure score
Infrastructure increases income by PKR 1,000.
Adoption × Digital 1,500 < 0.001 The income benefit of adoption increases by
Infrastructure PKR 1,500 for each unit increase in
Infrastructure.

Note: This moderation analysis investigates the role of digital infrastructure as a moderator in
the relationship between online sales channel adoption and monthly income. Interaction effects
highlight that improved digital infrastructure amplifies the economic impact of online sales
channels, particularly in rural contexts. All p-values being highly significant (p < 0.001)
indicate robust effects.

8000
Digital Infratructure as a Moder-
7000
ator
6000
5000
4000
3000
2000
1000
0

6.54.2 Digital Literacy as a Moderator

This analysis examines how digital literacy moderates the relationship between Adoption and
income. An interaction term (Adoption × Digital Literacy) was included.
Explanation of Results
Significance: The interaction term is statistically significant (p<0.01), confirming the
moderating role of digital literacy.
Interaction Effect: Higher digital literacy enables adopters to utilize online platforms more
effectively, thereby enhancing the income benefits associated with adoption.

Interpretation of Digital Literacy as a Moderator

 26

The impact of online sales adoption on household income is also found to be moderated by
digital literacy. The findings in Table 7 indicate that for every one-unit increase in digital
literacy, revenue from adoption increases by PKR 1,200. This highlights the critical role that
digital skills play in maximizing earnings through online sales channels. Digital savvy
consumers are better equipped to navigate e-commerce platforms, utilize social media for
marketing, and employ mobile money effectively. Therefore, improving rural population's digital
literacy level can enhance the efficiency and effectiveness of online selling channels, leading to
higher economic returns. The statistically significant positive relationship between adoption and
digital literacy underscores the importance of education and training in equipping individuals
with digital skills for driving broader economic liberation.

Table 7: Moderation Analysis Results for Digital Literacy

Variable Coefficient p-value Interpretation

(β)
Adoption Status 7,500 < 0.001 Adopting online sales channels increases
income by PKR 7,500.
Digital Literacy 800 < 0.01 Higher digital literacy increases income by
PKR 800 for each literacy level.
Adoption × 1,200 < 0.01 The income benefit of adoption increases by
Digital Literacy PKR 1,200 for each literacy level.

Note: Digital literacy serves as a crucial moderator in enhancing the economic impact of online
sales channel adoption. Households with higher literacy levels experience more substantial
income gains, underscoring its importance in rural settings.
 27

Digital Literacy as a Moderator

Adoption Status

Digital Literacy

Adoption × Digital Literacy

0 1000 2000 3000 4000 5000 6000 7000 8000

6.55.3 Secondary Moderators: Age and Education Heterogeneity Analysis

To provide additional context, age and eEducation were analyzed as secondary moderators.
6.5.1 Education LevelHeterogeneity of Education Level
Net sales also vary significantly with the education level, shaping the impact of online sales
channels. The findings in Table 8 reveal that with each additional year of schooling, the income
advantage of adoption rises by PKR 800. This depicts that individuals with higher levels of
education are better equipped to effectively market and sell their products online. This advantage
can be attributed to a better understanding and knowledge of Internet based marketing.
Policymakers should realize that educated persons are more likely to be conversant with
technology, hence, better placed generate higher income from E-commerce. Thus, education
plays a moderating role in enhancing the impact of online sales channel adoption on household
income.
6.5.2 AgeHeterogeneity of Age
Age significantly influences the extent of income benefits households realize through adopting
internet-based sales channels. The analysis in Table 8 reveals that the interaction between age
and adoption reduces income benefits by PKR 150 for each unit increase in age. This implies
that young people are more likely to experience higher income gains or enjoy enhanced income-
generating opportunities from online sales channels. This outcome can be attributed to younger
generations being more familiar to digital tools and systems than their older counterparts. Their
 28

adaptability and proficiency in using internet-based platforms enable them to better leverage e-
commerce opportunities. Conversely, older individuals may face challenges in fully utilizing the
benefits of adoption due to factors such as limited digital literacy or minimal exposure to modern
technologies. These barriers particularly affect the elderly, underscoring the need for targeted
interventions, such as training programs, to enhance their ability to benefit from digital sales
channels.

Table 8: Moderation Analysis Results for Education Level and Age

Variable Coefficient p-value Interpretation

(β)
Adoption Status 7,500 < 0.001 Adopting online sales channels increases
income by PKR 7,500.
Education Level 500 < 0.05 Each additional year of education increases
income by PKR 500.
Adoption × Education 800 < 0.01 The income benefit of adoption increases by
PKR 800 per year of education.
Age -200 0.10 Older individuals experience slightly lower
income benefits from adoption.
Adoption × Age -150 0.05 Younger individuals gain more from
adoption, with benefits decreasing slightly as
age increases.

Note: This table highlights the moderating effects of education and age on the relationship
between online sales channel adoption and household income. Higher education levels amplify
income benefits, while younger individuals tend to experience greater economic advantages.
These findings underline the importance of targeted education, and support for older
populations to optimize the benefits of digital adoption.
 29

Secondary Moderators: Age and Education

8000

7000

6000

5000

4000

3000

2000

1000

0
Adoption × Age Age Adoption × Education Education Level Adoption Status
-1000

6.6 Summary of Findings

The findings of the current studies confirm that online sales platforms significantly improve
household income with digital infrastructure and digital literacy being key moderators. Better
access to the internet and mobile devices amplifies income benefits, while higher digital literacy
enables households to utilize online tools effectively, boosting their economic gains. Age and
education also play important roles. Younger individuals, being more adept at using technology,
gain higher income advantages, whereas older users face challenges like lower digital literacy.
Similarly, education enhances the benefits of adoption, as educated individuals are better
equipped to leverage online platforms. These findings highlight the importance of digital
infrastructure, literacy, and education in maximizing the income potential of online sales
channels for rural households.
7. Discussion
7.1 Implications of Findings
The study demonstrates that online sales channels are effective in enhancing rural household
income in Pakistan. The positive ATT indicates that the income level of the treated group (online
sales channel adapters) was higher than the untreated group (traditional sales channels relying
group). This implies that online platforms allow rural households to access broader markets;
 30

moving their products beyond local neighbors at better prices and higher sales volumes (Aker
and Cariolle, 2023). A key advantage is reduced reliance on intermediaries, enabling rural
producers to retain higher profit margins. These findings affirm that digital commerce is a viable
model for economic development, particularly for agriculture and artisanal goods that previously
lacked market access.
Furthermore, the study implies that online sales channels can help address the problem of income
disparities in rural areas. By providing market access to households regardless of their location,
these platforms enable small or localized producers to participate in economic activities that may
gradually narrow the rural-urban income gap (Duncombe, 2016). The positive ATT results
reinforce this, indicating that online sales platforms are especially effective in boosting incomes
where IT tools facilitate their adoption and usage.
7.2 Barriers and Moderating Factors
This study implies that access to digital literacy and ICT Infrastructure significantly moderates
the degree of technology-driven sales channels in rural settings. Moderation analysis confirms
that households with higher digital literacy achieve greater economic gains from adopting online
sales channels. Digital literacy serves as both an enabler of technology use and a determinant of
economic success in digital markets (Hjorth & Khoo, 2016). The study also reveals that
enhanced digital skills enable rural households to efficiently operate online platforms, analyze
market trends, and manage e-commerce, such as listing product specifications, handling
payments, and responding to customer questions. In contrast, households with limited digital
skills are often excluded from these economic opportunities, underscoring the critical need for
improving digital literacy in rural areas.
This study also highlights that infrastructure access mitigates the effect of online sales channels
on rural incomes. Households in areas with robust network infrastructure, such as reliable
internet connectivity and transportation, experience improved income from online sales channels.
Stable internet is crucial for rural producers to engage in e-commerce effectively, while efficient
transportation ensures timely product delivery, preventing delays that could reduce customer
satisfaction (Tegegne & Alemu, 2019). Consequently, inadequate infrastructure constrains online
sales adoption and limits rural households' ability to maximize income through digital markets.
 31

These findings underscore the need for targeted efforts to enhance both digital literacy and
supporting infrastructure. Investments in digital skills training for rural households and
improvements in internet and transportation networks can substantially expand the economic
benefits of online sales channels for rural communities (Endris & Kassegn, 2022).
8. Conclusion
8.1 Summary of Key Findings
The findings confirm that rural households in Pakistan benefit significantly from adopting online
sales channels; as evidenced by the statistically significant ATT derived through PSM analysis.
Households selling through e-platforms register higher revenues compared to households relying
on traditional sales channels (Brynjolfsson & McAfee, 2014). Furthermore, moderation analysis
confirms the experimental hypothesis, highlighting the critical roles of digital literacy and
infrastructure access. Households with higher digital literacy and better access to essential
infrastructure, such as reliable internet and transportation, reap more substantial income benefits.
These findings underscore the importance of online sales platforms for increasing rural income
while emphasizing the importance of contextual factors that influence the effectiveness and reach
of these platforms.
8.2 Policy Implications
Several policy recommendations are proposed to facilitate the development and efficiency of
online sales channels in rural areas.
8.2.1 Enhancing Digital Literacy Programs
Digital literacy is a significant determinant of income gains from online sales channels, with a
correlation of 0.6 between digital literacy and income. Policymakers should prioritize focused
interventions to improve digital competence among rural populations. These programs should
target skills that are critical for carrying out business activities online, such as creating online
store, handling electronic payments, and utilizing digital advertising. Effective implementation
strategies include utilizing community based digital training centers, involving school going
children as intermediaries to educate family members and deploying mobile training units to
reach remote rural areas. Improving digital literacy in rural households will help counter barriers
to adoption, enabling them to fully reap the economic benefits of e-commerce.
8.2.2 Strengthening Digital Infrastructure and Connectivity
 32

Digital infrastructure, including internet and mobile networks, is critical for connecting rural
producers to online market sales. The analysis highlights a significant correlation of 0.5 between
infrastructure and income benefits, with each unit enhancement increasing income by PKR
1,500. To strengthen digital infrastructure, policies should focus on expanding broadband and
mobile network access in rural areas, providing subsidies or tax incentives to telecommunication
companies to encourage investments in rural digital infrastructure and establishing free internet
zones in rural centers to promote e-commerce activities and enable producers to access broader
markets efficiently.
8.2.3 Improving Access to Digital Tools and Technologies
The study reveals that consumers using online sales channels experience an average higher
income, primarily due to availability of multiple digital tools. To enable wider access, local
authorities should design programs that would allow the start-ups to acquire smartphones,
computers, and payment processors at affordable rates. This can be done through micro-credit
schemes, grants or direct subsidies to support credit-constrained, low-income consumers in
acquiring essential digital tools and ensuring these tools feature user-friendly interfaces and
support for local languages to increase usability and adoption rates.
8.2.4 Developing Rural-Centric E-Commerce Platforms
Focusing on localized e-business solutions is essential to address the specific needs of rural
producers. These platforms should incorporate minimalistic designs for easy navigation,
translation options, and logistics support for order management. These platforms can be created
by private technology firms with the support from policymakers. Additional features like mobile
payment gateways and real-time tracking systems could further enhance their usability. These
platforms can serve as direct marketplaces for rural producers, reducing their reliance on
intermediaries and increasing profitability. Improved access means that rural households can
expand their operations and diversify their income streams.
8.2.5 Enhancing Transportation and Logistics Infrastructure
A robust transport system is necessary for product quality and timely delivery especially for
agricultural and the artisan products. The lack of formal logistics in rural e-commerce can hinder
transaction efficiency, leading to reduced customer satisfaction and economic gains. Priority
should be given to modernizing rural roads to improve accessibility, developing logistics
 33

platforms to streamline distribution and promoting eco-friendly transportation solutions to align

with sustainable practices. Collaboration between policymakers and private sector stakeholders
can enhance delivery services, improving market access for rural producers and boosting their
ability to penetrate urban and regional markets efficiently.
8.3 Future Research Directions
As with any research, this study presents several paths for future research on adopting
technology in rural environments. Future studies can investigate the dynamic effect of OSC
adoption on rural household income, focusing on income volatility and broader life chances,
including education and health (Savastano, 2017). Subsequently, other studies can extend their
analysis to identify the exact product categories that may be most suitable for online sales and
analyze seasonal trends and periods that maximize economic benefits within agriculturally
oriented rural economies.
An important research avenue can be examining gender as a factor in determining online sales
channel adoption. Understanding how gender influences digital literacy and technology uptake
can highlight challenges faced by female entrepreneurs in rural areas. This insight could guide
the development of gender-sensitive policies to empower women in rural enterprises (Zahra &
Wright, 2016). Finally, cross-regional or cross-country comparisons could help identify how
varying socioeconomic and cultural norms affect the adoption and impact of online sales
channels on rural economies. Such studies can offer valuable insights into designing policies and
interventions to maximize benefits for different demographic and geographical settings.
 34

References

 Abebe, G. K., Bijman, J., & Royer, A. (2016). Are intermediaries facilitators or barriers
to improving smallholders' welfare in rural economies? Empirical evidence from
Ethiopia. Journal of Rural Studies, 43, 203-213.
 Abu-Bader, S. H. (2021). Using statistical methods in social science research: With a
complete SPSS guide. Oxford University Press, USA.
 Aker, J. C., & Cariolle, J. (2023). Mobile Phones and Development in Africa. Springer
International Publishing AG.
 Aker, J. C., Ghosh, I., & Burrell, J. (2016). The promise (and pitfalls) of ICT for
agriculture initiatives. Agricultural Economics, 47(S1), 35–48.
[Link]
 Akinola, A. A. (2017). Influence of socioeconomic factors on farmers' use of mobile
phones for agricultural information in Nigeria. Library Philosophy and Practice, 1688, 1-
14.
 Ali, S., Jabeen, U. A., Nikhitha, M., & India, Z. A. M. B. I. A. (2016). Impact of ICTs on
agricultural productivity. European Journal of Business, Economics, and Accountancy,
4(5), 82-92.
 Brynjolfsson, E., & McAfee, A. (2014). The second Machine Age: Work, progress, and
Prosperity in a time of brilliant technologies. WW Norton & company.
 Chen, J. W., Maldonado, D. R., Kowalski, B. L., Miecznikowski, K. B., Kyin, C.,
Gornbein, J. A., & Domb, B. G. (2022). Best practice guidelines for propensity score
methods in medical research: consideration on Theory, implementation, and reporting. A
review. Arthroscopy: The Journal of Arthroscopic & Related Surgery, 38(2), 632-642.
 Chittipaka, V., Kumar, S., Sivarajah, U., Bowden, J. L. H., & Baral, M. M. (2023).
Blockchain Technology for Supply Chains operating in emerging markets: an empirical
examination of technology-organization-environment (TOE) framework. Annals of
Operations Research, 327(1), 465-492.
 Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of
information technology. MIS Quarterly, 13(3), 319–340. [Link]
 35

 De Schutter, O. (2017). The political economy of food systems reform. European Review
of Agricultural Economics, 44(4), 705-731.
 Dearing, J. W., & Cox, J. G. (2018). Diffusion of innovations theory, principles, and
practice. Health affairs, 37(2), 183-190.
 Deichmann, U., Goyal, A., & Mishra, D. (2016). Will digital technologies transform
agriculture in developing countries? Agricultural Economics, 47(S1), 21–33.
 Duncombe, R. (2016). Mobile phones for agricultural and rural development: A literature
review and suggestions for future research. The European Journal of Development
Research, 28, 213-235.
 Dwivedi, Y. K., Shareef, M. A., Simintiras, A. C., Lal, B., & Weerakkody, V. (2016). A
generalized adoption model for services: A cross-country comparison of mobile health
(m-health). Government Information Quarterly, 33(1), 174-187.
 Endris, E., & Kassegn, A. (2022). The role of micro, small and medium enterprises
(MSMEs) to the sustainable development of sub-Saharan Africa and its challenges: a
systematic review of evidence from Ethiopia. Journal of Innovation and
Entrepreneurship, 11(1), 20.
 Estriegana, R., Medina-Merodio, J. A., & Barchino, R. (2019). Student acceptance of
virtual laboratory and practical work: An extension of the technology acceptance model.
Computers & Education, 135, 1-14.
 Hahs-Vaughn, D. L. (2016). Applied multivariate statistical concepts. Routledge.
 Hayes, A. F. (2017). Introduction to mediation, moderation, and conditional process
analysis: A regression-based approach. Guilford publications.
 Hendricks, S., & Mwapwele, S. D. (2024). A systematic literature review on the factors
influencing e-commerce adoption in developing countries. Data and Information
Management, 8(1), 100045. [Link]
 Hjorth, L., & Khoo, O. (Eds.). (2016). Routledge handbook of new media in Asia.
London: Routledge.
 Imtiaz, S., Ali, S. H., & Kim, D. J. (2020). E-commerce growth in Pakistan: Privacy,
security, and trust as potential issues. Cyber Security and Research, 26(2), 10–18.
 36

 Israr, M., Yaseen, A., & Ahmad, S. (2017). Sustainable rural development and livelihood
sources of the rural households in mountainous Pakistan. American Journal of Rural
Development, 5(4), 97-105.
 Karine, H. A. J. I. (2021). E-commerce development in rural and remote areas of BRICS
countries. Journal of Integrative Agriculture, 20(4), 979-997.
 Keith, T. Z. (2019). Multiple regression and beyond An introduction to multiple
regression and structural equation modelling. Routledge.
 Kshetri, N. (2018). Rural e-commerce in developing countries. IT Professional, 20(2),
91–95.
 Lee, S. W., & Acharya, K. P. (2022). Propensity score matching for causal inference and
reducing the confounding effects: statistical standard and guideline of Life Cycle
Committee. Life Cycle, 2.
 Li, X., Li, Y., & Chen, Z. (2024). Impact of rural e-commerce participation on farmers’
household development resilience: Evidence from 1229 farmers in China. Agriculture,
14(5), 692. [Link]
 McCluney, J. K., Anderson, C. M., & Anderson, J. L. (2019). The fishery performance
indicators for global tuna fisheries. Nature communications, 10(1), 1641.
 Mitzner, T. L., Savla, J., Boot, W. R., Sharit, J., Charness, N., Czaja, S. J., & Rogers, W.
A. (2019). Technology adoption by older adults: findings from the PRISM trial. The
Gerontologist, 59(1), 34-44.
 Nazir, M., & Roomi, M. (2021). Barriers to adopting electronic commerce for small and
medium-sized enterprises in emerging economies. EMAJ: Emerging Markets Journal,
10(1), 43–55. [Link]
 Oberlack, C., Tejada, L., Messerli, P., Rist, S., & Giger, M. (2016). Sustainable
livelihoods in the global land rush? Archetypes of livelihood vulnerability and
sustainability potentials. Global environmental change, 41, 153-171.
 Pan, W., & Bai, H. (2018). Propensity score methods for causal inference: an overview.
Behaviormetrika, 45(2), 317-334.
 Rogers, E. M. (1962). Diffusion of Innovations. Free Press.
 37

 Rosenbaum, P. R., & Rubin, D. B. (2023). Propensity scores in the design of

observational studies for causal effects. Biometrika, 110(1), 1-13.
 Savastano, S. (2017). Agricultural intensification: The status in six African countries:
Agricultural intensification: The status in six African countries.
 Shah, H. J., & Attiq, S. (2016). Impact of technology quality, perceived ease of use and
perceived usefulness in the formation of consumer’s satisfaction in the context of e-
learning. Abasyn J. Soc. Sci, 9(1), 124-140.
 Siegel, D., Acharya, P., & Sivo, S. (2017). Extending the technology acceptance model to
improve usage & decrease resistance toward a new technology by faculty in higher
Education. Journal of Technology Studies, 43(2), 58-69.
 Stromquist, N. P. (2019). World Development Report 2019: The changing nature of
work: By the World Bank. Washington, DC, World Bank, 2019, 151 pp. ISBN 978-1-
4648-1342-9 (hbk). ISBN 978-1-4648-1328-3 (pbk), ISBN 978-1-4648-1356-6 (eBook),
DOI: 10.1596/978-1-4648-1328-3.
 Tegegne, A. K., & Alemu, T. A. (2019). SMS-Based Agricultural Information System for
Rural Farmers in Ethiopia. Journal of Usability Studies, 15(1).
 Tim, Y., Cui, L., & Sheng, Z. (2021). Digital resilience: How rural communities
leapfrogged into sustainable development. Information Systems Journal, 31(2), 323-345.
 Wattoo, M. U., & Iqbal, S. M. J. J. S. O. (2022). Unhiding nexus between service quality,
customer satisfaction, complaints, and loyalty in online shopping environment in
Pakistan. SAGE Open, 12(2), 21582440221097920.
[Link]
 World Bank. (2016). World Development Report 2016: Digital Dividends. Washington,
DC: World Bank. [Link]
 Yadav, V. S., Singh, A. R., Raut, R. D., & Cheikhrouhou, N. (2021). Design of multi-
objective sustainable food distribution network in the Indian context with multiple
delivery channels. Computers & industrial engineering, 160, 107549.
 Zahra, S. A., & Wright, M. (2016). Understanding the social role of entrepreneurship.
Journal of management studies, 53(4), 610-629.
 38

Appendices