Companies now spend billions of dollars paying social media personalities to promote their products. Spending on influencer marketing has ballooned from $1.7 billion in 2016 to an estimated $24 billion by the end of 2024. But a basic question has lingered behind all that spending: Do influencers actually outperform other marketing options, and if so, under what conditions?
A new meta-analysis, which is a study that statistically combines the results of many previous studies, examined 135 experiments involving more than 44,000 consumers. The research, published in the Journal of the Academy of Marketing Science, found that social media influencers do have a significant positive effect on consumer engagement and purchase intention. But the results also reveal that the size of an influencer’s following matters in ways that many marketers may not expect.
A question of conflicting evidence
The study was led by Mojtaba (Moji) Barari of the University of Newcastle in Australia, along with Martin Eisend of the University of Vienna and Shailendra Pratap Jain of the University of Washington. The team set out to address four gaps in the existing research on influencer marketing.
First, prior studies had produced conflicting findings about whether influencers with large followings or small followings are more effective. Some research suggested that influencers with more followers generate more engagement, while other work found the opposite. Second, there was limited generalizable evidence about how social media influencers compare to alternatives like celebrities, computer-generated “virtual influencers,” or brands’ own social media posts. Third, researchers had proposed multiple theories for why influencers work, but no one had tested an integrated model combining these theories. Fourth, factors like the type of message, the kind of product being promoted, and the age of the influencer had not been studied in a systematic way.
To understand the team’s approach, it helps to know some background. Social media influencers are individuals who have built followings on platforms like Instagram, TikTok, or YouTube by creating content in specific areas of interest. Companies pay them to showcase products in a way that feels more personal and authentic than traditional advertising. Celebrities, by contrast, leverage fame from fields like sports or entertainment. Virtual influencers are entirely computer-generated characters. Each type of endorser has different strengths and weaknesses, and brands must decide which to use.
How the researchers combined 135 experiments
The team searched nine major academic databases using terms like “social media influencer,” “influencer marketing,” and “influencer endorsement.” Their search, completed in July 2024, initially returned 2,075 papers. They then narrowed the pool by including only experimental studies, which are studies where researchers control conditions to test cause-and-effect relationships. Each included study had to compare a social media influencer endorsement condition against a control condition, such as a celebrity endorsement, a virtual influencer, a brand’s own social media post, or no endorsement at all.
After applying these filters, the team was left with 71 papers containing 135 independent experiments and 571 effect sizes. An effect size is a statistical measure of how large or small the difference is between two groups being compared. The researchers converted all effect sizes into correlations and adjusted them for measurement errors. They also checked for publication bias, which is the tendency for studies with positive or significant results to be published more often than those without. Their tests indicated that publication bias was not a major concern in the dataset.
To examine the process by which influencers shape consumer behavior, the team used a technique called meta-analytic structural equation modeling. This allowed them to test how influencer effects flow through a chain of consumer responses, from initial perceptions of trustworthiness and appeal, through attitudes toward the ad and the brand, and eventually to engagement and purchase intention. To identify which factors strengthen or weaken influencer effectiveness, they used a statistical method called hierarchical linear modeling, which accounts for the fact that many studies reported multiple outcomes.
What the numbers revealed
The analysis showed that social media influencers have a positive and statistically significant effect on consumer perceptions of credibility, attitudes toward ads, attitudes toward brands, engagement, and purchase intention. The one exception was attractiveness: influencers were perceived as no more attractive than other types of endorsers, including celebrities and virtual influencers. This suggests that attractiveness is not a unique advantage of social media influencers, while credibility is.
When compared directly to other endorsement types, social media influencers outperformed brand-generated social media posts and no-endorsement strategies on both engagement and purchase intention. They also outperformed celebrities in driving engagement, though they were roughly equal to celebrities in driving purchase intention. Against virtual influencers, social media influencers were more effective at driving purchase intention but performed similarly on engagement.
The structural modeling revealed that social media influencers do not directly drive engagement or purchase intention. Instead, their effect flows through a chain of responses. Influencers first shape how consumers perceive their credibility and attractiveness. These perceptions then shape attitudes toward the advertisement and the brand. Those attitudes, in turn, influence engagement and purchase intention. In other words, if an influencer is not seen as credible, the downstream effects on brand attitudes and buying interest weaken.
The size question: Small versus large influencers
One of the study’s most notable findings involves influencer size, defined by the number of followers. The researchers split influencers into two categories: “large” (more than one million followers) and “small and medium” (fewer than one million followers). Previous research had been divided on whether bigger is better, and this analysis offers a resolution.
Smaller influencers were found to be more effective at driving engagement, which includes activities like liking, commenting, sharing, and word-of-mouth about a product. Larger influencers, on the other hand, were more effective at driving purchase intention. The researchers suggest this pattern may exist because smaller influencers maintain closer relationships with their audiences, making their endorsements feel more personal and authentic. Larger influencers, meanwhile, signal greater popularity and status, which may make their recommendations more persuasive when it comes to actual buying decisions.
Influencer size also interacted with other factors. Larger influencers were better matched with concrete, evidence-based messages and with familiar, well-known brands. They also performed better when endorsing “search products,” a term for goods that consumers can evaluate before purchasing, like electronics or appliances. Smaller influencers, by contrast, were better suited to abstract, recommendation-style messages and to “experience products,” which are goods consumers can only fully evaluate after using them, like restaurants or skincare.
Other factors that shape influencer effectiveness
The age of the influencer was linked to a decrease in effectiveness, suggesting that younger influencers may resonate more with audiences on social media. Regarding the type of message, more abstract messaging and recommendation-based endorsements were generally linked to stronger influencer effects. New products and experience goods also showed higher influencer impact compared to familiar products and search goods.
Interestingly, whether the message used an emotional or rational appeal did not significantly affect outcomes overall. Neither did the influencer’s gender, the specific social media platform used, or whether the influencer and product were perceived as a good fit. However, these null findings should be interpreted with caution, as the number of studies testing some of these factors was limited.
What this means for marketing decisions
The findings point to several practical considerations for companies investing in influencer marketing. The first is that the choice between a social media influencer and another type of endorser should depend on the campaign’s goal. If a brand wants to boost engagement, social media influencers or even virtual influencers can be effective. If the goal is to drive purchases, social media influencers and celebrities are stronger options.
Second, influencer size should align with the campaign objective. Brands looking to spark conversation and community interaction around a product may benefit from working with smaller influencers who have tighter relationships with their audiences. Brands aiming for broad reach and direct sales may be better served by larger influencers. The message strategy should follow suit: larger influencers pair well with factual, demonstration-heavy content, while smaller influencers are better suited to storytelling and general endorsements.
Third, product type matters. For products that consumers can research and compare before buying, larger influencers can provide the detailed, credibility-building content that aids decision-making. For products that require personal experience to evaluate, smaller influencers can use personal narratives to make the case more convincingly.
There are important caveats. The vast majority of the experiments analyzed relied on purchase intention rather than actual purchase behavior. There is a well-documented gap between what consumers say they intend to buy and what they actually buy. The researchers note this as a significant limitation. Most studies also assessed outcomes immediately after exposure, meaning the long-term effects of influencer marketing remain largely unknown. Additionally, a disproportionate number of the original studies focused on Instagram, which may limit how well these findings apply to platforms like TikTok or YouTube, where content formats and audience behaviors differ.
The study also does not account for the cost differences between large and small influencers or between influencers and celebrities. A campaign’s return on investment depends not just on effectiveness but on price, and the analysis does not address that dimension. Still, for marketers trying to navigate a rapidly growing and often confusing field, the research offers a data-driven framework for matching the right influencer type to the right objective.
