Fiche 9

Posted on by

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.