Are influencers something new under the sun or just celebrities by another name?
Perhaps you think the most appropriate response to this question is, ‘Who gives a shit?’ After all, this is marketing, not philosophy, and debating taxonomies doesn’t move metrics.
I’ve got a lot of sympathy for that attitude, but influencers emerged from the intersection of youth culture and technology, and this industry has repeatedly demonstrated that it cannot be trusted to assimilate new phenomena within these spheres in level-headed fashion.
Marketers as a group over-index on neophilia and status anxiety, and influencers are the perfect trigger for those tendencies. Without intervention, the urge to romanticise influencers and the relationship they have with audiences will triumph to the detriment of best practice.
So there’s more at stake here than semantics. A robust classification grounded in empirical evidence could curb that instinct to put influencers on a pedestal, and break the inevitable cycle of inflated expectations and monomania.
Apples and influencers
The IPA research published in October did nothing to temper the idea that influencers are special and should be treated as a genus unto themselves. A sample of 18 UK case studies showed that influencer marketing had a long-term ROI 50% higher than the all-channel average, but also that there was an abnormally large amount of variation between the performance of the campaigns.
Scott Guthrie, the director general of the Influencer Marketing Trade Body, says that he interpreted the results to mean that while a bad TV or radio ad will still work if it’s on air enough, you can’t use a large media budget to spend your way to effectiveness with influencers.
Other people came to the same conclusion, and rightly so. The data does indeed suggest what Guthrie says it does. My only issue is that comparing human beings who endorse products with advertising formats, like TV and online display, is bound to produce results that make influencers seem unique.
Use traditional celebrity endorsers as the basis for comparison instead and you’ll get the same insights, but without the kind of contrasts that imply influencers are re-writing the rules of marketing.
A meta-analysis paper published in 2016, for example, quantified the impact of celebrity endorsements across 46 independent studies and found an average effect of zero, but with ‘strong attitudinal and behavioural effects when including theoretically relevant moderator variables.’
Essentially, the authors are saying that everything depends on using the right celebrity endorser in the right way because large media budgets can’t make up for bad content — sound familiar?
A study published in 2019, which compared influencers with celebrities, put it more bluntly. The researchers showed ads for the same beauty product, with either a famous model or an influencer, to hundreds of participants and found ‘no evidence that fit [between the brand and the endorser] is more important for influencers than celebrities.’
And if celebrities and influencers do work the same way, then it’s useful to know that the meta-analysis also found endorsements by famous people were generally less persuasive than quality seals, industry awards, or even endorsements by another brand in the same category.
Leave authenticity behind
This is not meant to downplay the importance of influencers in the media ecosystem, or imply that brands should not use them to promote their products. Despite the caveats mentioned above, the authors of the meta-analysis still concluded that celebrity endorsements were ‘undoubtedly an effective [form] of marketing communication’, and I have no doubt that the same is true of influencers.
If this article is intended to undermine anything, it’s the way that marketers constantly talk about influencers in terms of authenticity. The term is too vague to have any real descriptive or predictive ability. Invoking it as the source of influencers’ potency only adds another unhelpful layer of mystique because it can’t be measured or tested. But even as a vague descriptor, authenticity no longer serves the same function that it once did.
Strategist Graeme Douglas-Kilgannon wrote earlier this month that content now succeeds on social media when ‘it matches, reinforces, or slightly disrupts learned patterns inside a low-attention, high-velocity cognitive environment’, and not when it aligns with our values.
Essentially, he argues that when the platforms switched from friend graphs to interest graphs, it skewed the game in favour of content that interrupts our mindless scrolling by appealing to the lizard parts of our brains. In these environments, large followings count for less, and so does producing content that reflects the real you.
Identity solution
Of course, it is possible to be too cynical. We’re not ruled by our lizard brains all the time, and people do still seek out online personalities because they feel a connection with them. So the way that influencers cultivate audiences is absolutely something that brands should think about — just not in terms of authenticity.
A study published in the Journal of Marketing Communications in June, which looked at the effects of influencer credibility and influencer-consumer similarity, offers a better way.
According to the authors, influencer credibility is made up of two components. Expertise (how much do they know?) and trustworthiness (how likely are they to tell you what they know?)
Influencer-consumer similarity is more subjective, but it’s still easier to define than authenticity. It’s just how much people think this influencer is like them.
The study found that an influencer’s credibility has a direct effect on how their audience responds to brand collaborations. Influencer-consumer similarity has an indirect effect because it encourages influencer-consumer identification — a deeper psychological bond than similarity — which, in turn, improves brand image.
It’s true that almost all the research agrees that people are more likely to feel similar to, or identify with, or trust influencers than traditional celebrities, but this seems to correlate with audience size, not any personality traits. If people identified with influencers simply because they are more authentic than celebrities, they would not also consistently identify more with smaller influencers than bigger ones — but they do.
I also looked at virtual influencers, to see if they were something new under the sun in terms of how people respond to them. There’s not as much evidence to go on yet, but in 2025, researchers from the Netherlands conducted an experiment with women and saw that they formed para-social relationships with virtual and human influencers in exactly the same way.
Fame is a spectrum
You shouldn’t draw firm conclusions from a lone (single-sex) experiment, but the evidence keeps piling up that fame is fame, whatever form it takes. The platforms have created environments that can support lower levels of fame, and it turns out that people are more likely to identify with personalities at that end of the scale, but this isn’t something that authenticity, or possibly any behaviour, can reliably mitigate.
And as media continues to fragment and social media platforms continue to behave more like broadcasters, the distinctions between old and new kinds of celebrity are likely to become even more confusing. Where will the stars of micro-dramas sit on the fame spectrum, for example?
In this environment, time spent thinking about the ontology of different kinds of influencers isn’t time wasted — it’s a way to keep perspective when everything appears to be changing.
And there’s usually some practical tips to be got from even the most abstract-sounding studies. Take the one above, about how women form para-social relationships with virtual influencers. It suggests that brands should treat virtual influencers the same as real ones, not like fictional characters. If that was more widely known, we might have avoided the situation where any marketer — even one with the very best intentions — would think it was a good idea to give Lil Miquela leukemia…
This article was first published in MediaCat’s free-to-download report on Influence, which is available here.