Influencer tiers are a convention, not a law

Every agency chart says a micro-influencer has 10,000 to 100,000 followers. If you run marketing for a herbal tea brand in Can Tho, the question to ask before choosing anyone is: who drew those two lines, and did they measure anything?

How are influencer tiers defined?

Campbell and Farrell (2020) describe an influencer as someone who posts on social media in exchange for compensation, in money or in kind, such as products, services or trips (Campbell and Farrell, 2020). They sort influencers into five groups. The table below gathers what the paper says about each. The price column is what the paper reports, and most of it comes with no source of its own.

Why call it a convention?

Four signs: three in Campbell and Farrell's paper and one in Lou and Yuan's.

First, the thresholds are a synthesis. Campbell and Farrell write that they drew on several sources to build the five groups, including industry material from the Association of National Advertisers (2018) and Launchmetrics (2018) (Campbell and Farrell, 2020). The paper contains no measurement showing that anything changes direction at 10,000 or at 100,000.

Second, the top two groups share the 1 million mark. A celebrity often has over 1 million followers and a mega-influencer 1 million or more. The real divider is whether the person was famous outside social media first (Campbell and Farrell, 2020). The authors also write that the groups rest not on follower numbers alone but on authenticity, accessibility, expertise and cultural capital (Campbell and Farrell, 2020). Once the criteria are more than one number, two people marking the same influencer can put them in different tiers.

Third, price and performance claims are mostly second-hand. The $250,000 per post for celebrities is attributed to a 2017 Digiday article, which the paper describes as saying such people are "infamously known" to charge it (Campbell and Farrell, 2020). The $50,000 and $5,000 figures come with no source. The statement that nano influencers often have the highest engagement carries neither source nor data; the point that micro and macro influencers give better value through higher engagement is cited to a 2018 Forbes article (Campbell and Farrell, 2020). The paper is a review with no data of its own.

Fourth, what actually predicts performance seems to lie elsewhere. Lou and Yuan (2019) surveyed social media users who follow at least one influencer. Their abstract concludes that informative value, trustworthiness, attractiveness and similarity to followers increase trust in branded posts, and that trust affects brand awareness and purchase intention (Lou and Yuan, 2019). Follower count is not among the predictors. I read only the abstract, so I do not know whether the paper tested follower count as a control.

None of this makes tiers useless. The paper ties celebrities to cultural capital and flashy luxury products, and says macro audiences aspire to be like them; micro and nano influencers offer accessibility and authenticity, and many followers find a micro-influencer's recommendations more genuine than a famous person's (Campbell and Farrell, 2020). A tier is shorthand for which function a person leans towards. The harm comes when 10,000 or 100,000 is read as a breakpoint: someone with 9,800 followers and someone with 10,200 do not differ in the way the paper describes.

If not by tier, then by what?

By the function you need. Campbell and Farrell argue that every influencer, large or small, carries three functions: an audience (organic reach, targeting, attention), an endorser (celebrity status, expertise, or being a fellow consumer like the viewer) and a social media manager (producing content, planning, tending the community) (Campbell and Farrell, 2020). They treat follower count as a fair starting proxy for reach, but stress that a full valuation should count all these sources of value (Campbell and Farrell, 2020).

Campbell and Farrell also advise that a brand can use people from different tiers in one campaign to draw on each tier's strengths (Campbell and Farrell, 2020). They note that the brand is responsible for making sure influencer posts carry proper advertising disclosure, and point to United States guidance; this post does not discuss Vietnamese rules (Campbell and Farrell, 2020). The "Look at" column comes from Campbell and Farrell; the "Warning sign" column is my own suggestion.

One check belongs before any signing. Follower count is the headline number, and it is also where the paper locates the risk of bots. Campbell and Farrell, citing Daniels (2018), note that many influencers buy bots even when firms look at comments and shares (Campbell and Farrell, 2020). So the paper's advice, to read the full spectrum of engagement and the history of follower counts, is a reason to treat any tier label with suspicion until the numbers behind it have been looked at.

How should a 30 million dong budget be split?

The example below is entirely hypothetical. The brand "Suong Som Tea", the figures and both engagement rates are my invention, and the rates are not typical of any industry.

A hypothetical example. Suong Som Tea launches a honey-ginger tea in Can Tho with a budget of at most 30 million dong. Option A: one person with 300,000 followers (macro by the table above), quoting 25 million dong a post, with an assumed engagement rate of 0.6 per cent. Option B: ten people under 10,000 followers, averaging 5,000, each receiving product worth 300,000 dong plus a 2.5 million dong content fee, with an assumed engagement rate of 6 per cent.

The last two rows reverse. Per follower, A is several times cheaper. Per engagement, B is cheaper. Neither table says which sells more tea, because that number is not in the example. If the goal is "everyone in Can Tho knows the brand in launch week", I lean to A. If the goal is "strangers dare to try a new flavour", the function needed is trust from people like them, and I lean to B, then check each person's real comments before sending product. Both leanings are my inference from Campbell and Farrell and from Lou and Yuan (Campbell and Farrell, 2020; Lou and Yuan, 2019), not measured results.

This split suits a brand that can count orders per person, for instance through individual discount codes. It does not suit one that cannot tie orders to a source: the table then compares only reach and engagement, and the line between nano and macro adds nothing.

Going deeper: what each source tells us, and what I could not check

I read Campbell and Farrell's paper, "More than meets the eye: The functional components underlying influencer marketing", in full (Business Horizons, 2020, volume 63, issue 4, pages 469 to 479) from a public PDF hosted by a college, not the publisher's page. It is an automatic text conversion; Figure 1 of the paper is an illustration and I rebuilt the table from the text, not from the figure. I have not read the sources the paper cites: Chen's 2017 Digiday article on celebrity prices, Wissman's 2018 Forbes article on micro and macro value, and the industry material used to define the categories (Association of National Advertisers and Launchmetrics, both 2018). For Lou and Yuan I read only the abstract; I do not know how the paper measured followers, who the sample was or how strong the effects were.

Three limits follow. First, the library holds only one source that states thresholds, so I could not set several sources side by side. The claim that tiers are a "convention" rests on the authors describing how they built the table, not on a comparison between sources. Second, the claims about prices and engagement in Campbell and Farrell are mostly unsourced, or cite industry sources of about 2017 to 2018, and are dated. Third, Campbell and Farrell is a review paper with no Vietnamese market data; the thresholds here should not be taken as a rule for Vietnam.