Reference : Wang, L., Huang, N., He, Y., Liu, D., Guo, X., Sun, Y., & Cheng, G. (2025). Artificial intelligence (AI) assistant in online shopping: A randomized field experiment on a livestream selling platform. Information Systems Research, 36(4), 2358–2374. https://doi.org/10.1287/isre.2023.0103

Key words: livestream selling • artificial intelligence (AI) • AI streaming assistant • human-AI interaction • product return • randomized field experiment.

Summary:

This article studies the impact of an AI-powered streaming assistant in livestream shopping platforms. The research investigates whether an AI assistant can help solve the tension between:

• streamers’ limited ability to answer all customer questions,
• and consumers’ need for immediate personalized information

Development: 

The study uses a large-scale randomized field experiment on a major Asian livestream shopping platform to examine how AI assistants affect:

• purchase behavior,
• consumer decision-making,
• and product return rates

Methodology : quantitative study

H1.aThe implementation of an AI streaming assistant during streaming sessions increases the number of purchases in livestream selling.

Hypothesis 1b. The implementation of an AI streaming assistant during streaming sessions decreases the number of purchases in livestream selling.

H2. The implementation of an AI streaming assistant during streaming sessions decreases product return rates in livestream selling

Research Design

The study used:

• a randomized field experiment
• on a leading livestream shopping platform in Asia.

Key findings:

AI assistants increase purchases

AI assistants reduce product return rates

AI reduces uncertainty in decision-making

AI works better for uncertain products

Reference:

Tian, Z., Dew, R., & Iyengar, R. (2024). Mega or Micro? Influencer Selection Using Follower Elasticity. Journal of Marketing Research (JMR), 61(3), 472–495. https://doi-org.devinci.idm.oclc.org/10.1177/00222437231210267

Keywords: Influencer marketing ,Streaming video & television, Causal inference, Deep learning Tags (Metadata) Online social networks

Summary: this article adresses the criteria for selecting influencers to partner .

While some firms collaborate with“mega” influencers with millions of followers, other firms partner with “micro” influencers with only several thousand followers,
but who also cost less to sponsor. To quantify this trade-off between popularity and cost, the authors develop a framework for
estimating the follower elasticity of impressions (FEI), which measures a video’s percentage gain in impressions (i.e., views) corresponding
to a percentage increase in the number of followers of its creator. Computing FEI involves estimating the causal effect
of an influencer’s popularity on the view counts of their videos, which is achieved through a combination of (1) a unique data set
collected from TikTok,

(2) a representation learning model for quantifying video content, and (3) a machine learning–based causal
inference method. The authors find that FEI is always positive, averaging .10, but often nonlinearly related to follower size.

They examine the factors that predict variation in these FEI curves and show how firms can use these results to better determine
influencer partnerships.

Development: 

Concepts :

Social network and influencer Marketing : how the structure of social networks can
impact downstream micro- and market-level outcomes.

• Mega-influencers are useful for awareness.
• Micro-influencers are perceived as more authentic and relatabl

Advertising:  the authors position influencer marketing as a form of social media advertising.

They study  the relationship between advertising investment and performance outcomes like sales or engagement.

The main research objective here is to know How does an influencer’s popularity (number of followers) causally affect the impressions/views of their videos?

in order to answer this  question the authors adobt a   causal inference and machine learning framework :

• Representation Learning using Variational Autoencoder (VAE) to analyze video content.
• Deep Instrumental Variables (Deep IV) to estimate the causal relationship between followers and impressions with content type and appeal type as moderators.

The data was collected from the Discover page of TikTok over six months (October 2020 – April 2021).

• 216 hashtags.
• 30 sponsored hashtags.
• More than 500,000 videos.

Key findings:

FEI is positive but non linear meaning that the following elasticity of impressions is positive on average so more followers generally increase impressions but the relationship is not linear.

Midtier influencers generate the most marginal returns compared to mega or micro influencers.

the effectiveness of the influencer depends on the content type (category) and the engagement goal (informative, entertaining, socializing).

Limitations: 

-Tiktok specific content : the data was only collected through the tiktok platform and no other social media platforms so the results cannot be generalized all across platforms.

-The focus was only on impressions / views and didnt focus on sales or purchase behavior which is the ultimate goal.

Reference:

Sun, Y., Wang, Y., Zhong, Y., Zhang, Z., & Zhu, M. (2024). How technical features of virtual live shopping platforms affect purchase intention: Based on the theory of interactive media effects. Decision Support Systems, 180. https://doi-org.devinci.idm.oclc.org/10.1016/j.dss.2024.114189

Keywords: Virtual live shopping platforms, anthropomorphism,media richness,psychological distance,customer engagement, purchase intention.

Summary:

This study addresses the research gaps by developing a theoretical
model based on the theory of interactive media effects (TIME) to investigate the influence of VLSPs’ technical features on customers’ purchase intentions.

results indicate that psychological distance plays a mediating role between anthropomorphism (full mediation) and customer engagement, as well as between media richness (partial mediation) and customer engagement in a survey of 299
VLSP users.

Development :

Virtual live shopping platforms (VLSPs) are an innovative form of intelligent shopping DSS that offer brands novel opportunities to interact with customers.

VLSPs represent the application system of artificial intelligence, 3D modeling, deep learning, and speech synthesis technologies in the live e-commerce field.

personified virtual streamer makes it easier to establish relationships with users, thereby promoting familiarity with their characteristics.

2 key features of VLSP:

• Media richness: The level to which a medium can supply communication capabilities to media users. /the medium’s ability to produce and
  transmit a diversity of sensory stimuli, as well as the ability of numerous
  cues to decide the channel’s ability to convey rich information (Huang and li su)

elements of Media richness : immediate feedback,
linguistic diversity, personal attention, and multiple cues

• Anthropomorphism :

  making virtual streamers appear more human-like in behavior, appearance, or interaction style.

Theoritical Framework : Theory of Interactive Media Effects (TIME)

Explains how media technology features influence consumers through psychological mechanisms.

Variables :

1-Independant variables : Media Richness, Anthropomorphism

2-Dependant variable: Purchase Intention

3-Pshychological (mediators)

• Pshychological distance: The extent to which consumers perceive something as close, tangible, and present
• Customer engagement: A psychological state of continuous attention while using VLSPs.

H1. Psychological distance mediates the relationship between
anthropomorphism and customer engagement.
H2. Psychological distance mediates the relationship between media
richness and customer engagement.

H4. Psychological distance and customer engagement play a role in
mediating the chain between anthropomorphism and purchase
intention.
H5. Psychological distance and customer engagement play a role in

Methodology : Quantitative research with data collected through a survey with 402 questionnaires ( 299 valid responses).

Key findings:

-Anthropomorphism has no significant direct effect on customer engagement or purchase intention. but has an indirect effect through psychological distance → customer engagement

-Media richness has a significant direct effect on customer engagement. It also has a significant indirect effect via psychological distance.

-Customer engagement is confirmed as a critical mediator leading to purchase intention.

-Pshychological distance has a strong indirect effect via customer engagement.

Limitations :

Data collected only from Taobao VLSP users.

reference :

Shao, Z. (2026). Dual congruence in live-streaming commerce: A mixed-method to examine the role of virtual influencers and live content on consumer purchase behavior. Journal of Retailing and Consumer Services, 88. https://doi-org.devinci.idm.oclc.org/10.1016/j.jretconser.2025.104546

Keywords:  Virtual influencer, Live-streaming commerce ,Perceived value theory, Source credibility theory ,Congruity theory Mixed-method

Summary : This article is about examining how the Dual congruence in live-streaming commerce affects consumer purchase behavior.

The authors study :

• The congruence between live content and brands
• The congruence between virtual influencers and brands

and how it affects :

• percieved hedonic and utalitarian value
• Source credibility

and their impact on the purchase intention of chinese consumers.

Development :

Virtual influencers are AI-generated or computer-created influencers that interact with audiences through live-streaming platforms and promote products or brands in other words They are constructed utilizing AI technologies, computer-generated imagery, and machine learning algorithms to create entities for avatars with digital personalities and realistic looks.

They engage with the audience , showcase emotional expressions through content that mirrors real-life scenarios, they seek to foster emotional bonds with their audience.

Source credibility:

• Credibility
• Expertise
• Attractiveness

Methodology : a mixed method :

1- a quantitative research :

Data collection: Questionnaire distributed through a Sina Weibo “Live Stream” group

H1.Congruence between live content and brands is positively related to the hedonic value of live content in live-streaming commerce.

H2.Congruence between live content and brands is positively related to the utilitarian value of live content in live-streaming commerce.

H3.Congruence between virtual influencers and brands positively related to credibility of virtual influencers in live-streaming commerce.

H4.Congruence between virtual influencers and brands positively related to attractiveness of virtual influencers in live-streaming commerce.

H5.Hedonic value of live content positively relates to purchase behaviors in live-streaming commerce.
H6.Utilitarian value of live content positively relates to purchase behaviors in live-streaming commerce

H7. Credibility of virtual influencers positively relates to purchase behaviors in live-streaming commerce.
H8. Attractiveness of virtual influencers positively relates to purchase behaviors in live-streaming commerce.

H5 and H8 were not supported.

2- Qualitative research: 30 participants from Study 1

Key findings:

• Congruence between live content, virtual influencers, and brands positively influences consumer perceptions
• Credibility and utilitarian value are the strongest drivers of purchase behavior
• Entertainment and attractiveness alone do not significantly increase purchases
• Consumers prioritize authenticity and professionalism over visual appeal

Reference :

Zhang, M., Liu, Y., Wang, Y., & Zhao, L. (2022). How to retain customers: Understanding the role of trust in live streaming commerce with a socio-technical perspective. Computers in Human Behavior, 127. https://doi-org.devinci.idm.oclc.org/10.1016/j.chb.2021.107052

Keywords: Live streaming commerce,Trust,Continuance intention,Social interactivity,IT affordance.

Summary : this article is about what role does trust play in customer’s buying decision in the context of live streaming shopping. The author explains how both social factors (interaction) and technical factors (platform features) can build trust in live streaming commerce.

Development: 

The author distinguishes two types of trust in this article :

trust in the streamer

Trust in the product .

Method : Online questionnaire survey with users of TAOBAO (chinese live streaming commerce platform).

Valid responses : 446

Statistical method: Structural Equation Modeling (SEM)

Conceptual model : based on  the Socio-Technical Systems Theory.

Social enablers :

• Active control
• Two-way communication
• Synchronicity

Technical enablers:

• Visibility affordance
• Personalization affordance

Influence : trust in streamers / product

Moderating effect :of live streaming genre

Key findings: 

Trust is one of the strongest drivers of continuance intention in live streaming commerce.

Trust is multidimensional : a customer can trust both the streamer and the product  and trust in the streamer can transfer to the trust in the product.

Real-time interaction and visibility are especially important in reducing uncertainty.

Reference

Tian, Z., Dew, R., & Iyengar, R. (2024). Mega or Micro? Influencer Selection Using Follower Elasticity. Journal of Marketing Research (JMR), 61(3), 472–495. https://doi-org.devinci.idm.oclc.org/10.1177/00222437231210267

Keywords: influencer marketing, causal inference, deep learning, representation learning, heterogeneous treatment effects, video data
Online supplement.

Summary: This article focuses on one of the main criteria to choose an influencer to partner with based on their effect on the consumer’s buying decision which is the influencer’s popularity .

Mega-influencers have millions of followers and large reach, but they are expensive. Micro-influencers have smaller audiences but are cheaper and often seen as more authentic. The authors wanted to measure the real causal impact of follower size on video performance.

the authors develop a framework estimating the follower’s elasticity of impression (FEI) and the calculates the causal effect between an influencer’s populairty on the view counts of their videos with data collected from Tiktok.

Development: 

1- follower elasticity of impressions (FEI) : measures the percentage increase in video impressions generated by a 1% increase in followers.

2-Representation Learning Framework (SMVAE):

AI model extracts and compresses information from:

• text,
• images,
• audio,
• editing styles/effects

In order to create its representation

3- Deep IV (Deep Instrumental Variables) :

estimates the causal effect of followers on impressions while controlling for:

• nonlinear relationships,
• heterogeneity,
• and unobserved confounders.

Independent Variable

• Influencer follower size

Dependent Variable

• Video impressions/views after 2 weeks

Moderators

The relationship changes depending on:

• content type (food, gaming, beauty, etc.)
• engagement style:
  • entertaining,
  • informative,
  • socializing/emotional.

Key findings: 

• More followers generally increase video impressions, but the relationship is nonlinear.
• The average Follower Elasticity of Impressions (FEI) is about 0.10, meaning a 1% increase in followers leads to a 0.10% increase in views.
• The FEI curve is inverted U-shaped: mid-tier influencers generate the highest marginal gains in impressions compared to micro or mega influencers.
• After controlling for content and hidden confounders, mega-influencers are not always the most effective choice.
• The effectiveness of influencer size depends on the type of content and the campaign objective (informative, entertaining, or socializing).
• Some campaigns benefit more from mega-influencers, while others perform better with smaller or mid-tier creators.
• The study shows that brands should select influencers strategically rather than assuming that bigger influencers always produce better results.

Reference:

Teresa Borges-Tiago, M., Santiago, J., & Tiago, F. (2023). Mega or macro social media influencers: Who endorses brands better? Journal of Business Research, 157. https://doi-org.devinci.idm.oclc.org/10.1016/j.jbusres.2022.113606

Keywords:  Celebrity endorsement, Brand equity, customer-brand management, Brand credibility,customer-endorser envolvement.

Summary: 

This study conducted an online survey to determine the antecedents and consequences of endorsers’ participation in marketing and communication strategies. The results of path analysis showed that brand and endorser credibility played a significant role in determining customer brand engagement and brand equity. Endorser credibility impacted brand equity only in the case of mega-influencers. Smaller influencers exhibited higher prowess than celebrities to engage customers, thus suggesting that “less is more.”

Development: 

The research mode was l built on the Associative-Network Memory Theory

The study employed a two-phase exploratory research design:

Phase 1 : Influencer pool selection

mega-influencers: christiano ronaldo endorsing nike

macro-influencers: couple Ana Guiomar and Diogo Valsassina endorsing Vodafone

Phase 2 : Data collection

online survey with 270 valid responses from portugese consumers  between 18-25 years old

. 61.1% evaluated the mega-influencer scenario and 38.9% evaluated the macro-influencer scenario.

Key findings: 

Brand credibility has a massive direct impact on brand equity (0.697) and customer brand engagement (0.318).

A customer’s emotional/cognitive involvement with an influencer strongly and positively affects both the influencer’s perceived credibility and customer-brand engagement.

the less is more theory was rejected (H12) type of influencer does matter significantly.

Mega-influencers effectively boost brand credibility and brand equity, but requires a re-existing congruence between the brand and the celebrity

Macro-influencers (smaller networks) display a higher prowess for generating direct consumer engagement

Limitation:

The data collection was limited exclusively to Portuguese consumers so it cant be generalizable to other countries or cultural contexts.

Reference :

Conde, R., & Casais, B. (2023). Micro, macro and mega-influencers on instagram: The power of persuasion via the parasocial relationship. Journal of Business Research, 158. https://doi-org.devinci.idm.oclc.org/10.1016/j.jbusres.2023.113708

Keywords : Influencer Marketing , Social Media Influencers, Micro-influencers,Parasocial relationship,Instagram.

Summary:

• This study analyzes the effect of the parasocial relationship on the audience’s intention to adopt the recommendations of micro, macro and mega-influencers, considering the number of followers, perceived popularity and opinion leadership. It used a sample of  140 portuguese influencers and classified them into  micro, macro or mega-influencers.
• A quantitative research method was applied using a questionnaire validated with  577  responses and analysed.

Development: 

this research  focuses onthe effect of the parasocial relationship on the persuasive power of different-sized SMis on their audiences.

They define SMis as an ordinary internet user who has cultivated a substantial audience in an organic way (Ki & Kim, 2019 and categorises them by the number of followers.

• Micro-influencers: 1,000 to 20,000 followers.

• Macro-influencers: 20,000 to 100,000 followers.

• Mega-influencers: More than 100,000 followers.

Analysis: The data was processed using Andrew Hayes’ macro PROCESS for SPSS to execute conditional process and moderated mediation analysis.

• H1a & H1b: Larger SMI size (more followers) directly increases their perceived popularity and attributed opinion leadership.

  H1c: Higher perceived popularity positively drives attributed opinion leadership.

  H2a & H2b: Perceived popularity and opinion leadership directly and positively impact the audience’s intention to adopt recommendations.

  H3a & H3b: Parasocial relationships act as moderators. Notably, the authors hypothesized that a strong parasocial relationship would negatively moderate the influence of opinion leadership, meaning emotional bonds would reduce the importance of an influencer’s perceived status or audience size

Key findings: 

The findings suggest that the indirect effect between the number of followers and the intention to adopt SMIs’ recommendations is mediated by the perceived popularity and opinion leadership, and are moderated by the parasocial relationship.

Significant differences are found between micro, macro and mega-influencers in terms of credibility, attractiveness and established relationship.

Limitation: 

This research was mainly conducted in the context of a small country (portugal) and didnt take in concideration a more global context

It was  conducted strictly on female influencers and specified one sector (fashion)

and the data was collected in  pre-pandemic landscape .

Reference: 

Wang, S., Yin, Y., Yuan, C., & Zhang, Z. (2025). Exploring consumer behavioural inertia in live streaming commerce. Journal of Marketing Management, 41(5/6), 485–507. https://doi-org.devinci.idm.oclc.org/10.1080/0267257X.2025.2495341

Keywords:

Live streaming commerce (LSS); consumer behavioural inertia (CBI); consumer confirmation (CC); value similarity (VS); socio-technical theory (STT); expectation-confirmation theory (ECT); structural equation modelling (SEM).

Summary: 

The article  aims to understand how the advantages of live streaming shopping influence consumer behaviour / consumer confirmation (CC). 

Then how consumer confirmation influences Consumer behavioural inertia (CBI).

And finally, whether value similarity (VS) strengthens these relationships.

Development: 

Live streaming shopping has become an increasingly interior choice of consumers. This article investigates how consumer behavioural inertia (CBI) is formed in live streaming commerce (LSS).

LSS allows real time interaction between streamers and viewers/ consumers increasing engagement and immersion (Chen, Zhang, & Zhao, 2022; Sun et al., 2019). This offers many advantages / dimensions such as : 

Perceived :  synchrony, Proximity , authenticity,Price discount,Convenience.

Which are grounded in Socio-Technical Theory (STT) (Li et al., 2016)

Consumer confirmation (cc): is a concept derived from  Expectation-Confirmation Theory (ECT) (Oliver, 1980).

Which explains that confirmation occurs when consumers’ expectations are either met or exceeded , Confirmation increases consumers satisfaction and this satisfaction increases continued usage and repurchase behaviour (Bhattacherjee, 2001; Hsu et al., 2015).

Consumer Behavioural Inertia (CBI) has 3 dimensions : 

-Cognitive inertia

-Affective inertia

-Behavioural inertia

Methodolgy : 

Data was collected through an online questionnaire of 224 respondants. 

Sample : chinese consumers with LSS experience.

The Analysis method used by the authors is : Structural Equation Modelling (SEM) 

Control variables: 

Age, gender, monthly  frequency of LSS usage.

Hypothesis : 

H1: LSS advantages positively affect CC.

H2: CC positively affects affective, cognitive, and behavioural CBI.

H3: Value similarity strengthens the relationship between LSS advantages and CC.

H4: CC mediates the relationship between LSS advantages and CBI.

Key findings

Perceived proximity, perceived authenticity, and perceived price discount positively affect consumer confirmation.

CC significantly influences all three types of CBI 

CC partially mediates the relationship between LSS advantages and CBI.

perceived synchrony and perceived convenience do not significantly influence CC.