Hypermoefication
Why virtual influencers are incredibly creepy and weird but vtubers are hecking good puppers
It’s surprisingly hard to find solid information on the virtual YouTuber industry. Part of it is just how new all this is -- I completely forgot that Kizuna Ai, generally considered the first vtuber as we now understand the term, only debuted in 2016 -- in light of how fast it’s metastasized into whatever it is now. Another big part is that much of the English-language coverage is fan run and far more invested in passing around insider joke memes and memes of memes while a small minority with proficiency in Japanese disseminate the tiny bits
of real information that ever get out. As of right now one of the best primary sources for a basic primer on the business model and strategy of vtuber agencies is this video series with YAGOO of Cover/Hololive from several years ago.
What I do know about to a decent degree is so-called influencer marketing. Paid endorsements have been around forever and influencers are just the latest digitized incarnation of getting your face printed on the Wheaties box. Basically, if you’ve seen an Instagram celeb artfully eating a very specific brand of Greek yogurt on the TL or a Youtuber asking if you hate going to the post office then you’ve been exposed to influencer marketing.
Influencers are a big enough business that many consumer brands have dedicated serious budgets towards them, running influencer programs alongside other established marketing acquisition channels like TV, print, paid search, and organic search. In this format influencers are part of so-called “performance” marketing, an aspect of marketing that can be tracked with tangible numbers and Key Performance Indicators. Those KPIs can vary, but more often than not they’re directly associated with a revenue-generating “conversion,” referring to the conversion of a user into a customer. Some examples include clicking on a link to a product page with the influencer’s affiliate hash on the URL, or signing up for a service using a discount code you got from the influencer. The whole reason for those unique identifiers is to be able to track the performance of each influencer: how many people do they convert versus the total amount of people who look at the post, how many people click on a link AND purchase the good/service, how long on average does it take for someone to convert, etc. These KPI numbers can then be used by both the influencer/agency and client company to negotiate the rates and terms of the influencer program. If the influencer is being paid per click on the affiliate link, what should that per-click cost be using their average conversion rate across all their sponsored posts and the overall audience for their web presences? On the client side, what is the overall ROI on an influencer program given that we spend $X per conversion they generate, our product costs $Y, and the program generates an average number of monthly conversions B. Therefore the actual monthly amount we’re making for the company on each sale is ($Y-$X)*B=ROI of influencer program. How does that math square up versus our other performance channels and should we turn up/turn down the faucet, or put the squeeze on the influencer(s) to get better value?
The nature of the shoutouts depends on the influencer themselves, but the most subtle (read: insidious) forms simulate so-called earned media, or positive mentions of your brand or products that you didn’t pay for.
The problem with running influencer campaigns, however, is the influencers. The brand has to cede a level of control of its marketing operation to a third party. While this does happen all the time in all the channels, dealing with influencers is different from dealing with the agencies or professional contractors you might make deals with elsewhere. Unless you’re interfacing through the intermediary layer of an agency that keeps a bunch of influencers stabled under it, influencers can be far more, how do you say, quirky. The same aspect that makes them a valuable channel -- the authenticity and social proof they receive from their followers -- also makes them much more likely to do stuff like read out your corny outreach emails on stream with a laughing-with-tears-in-my-eyes emoji in the corner.
So I wasn’t surprised at all when I found out about the virtual influencer thing. It’s exactly the sort of product born out of tech industry-style business thought patterns that reduce every issue to an algebra problem where the solution to the equation is always technology. Retain the acquisition and brand marketing benefits of influencer marketing without all the hassle of dealing with influencers’ personalities and insistence on getting paid by building your own influencer from scratch. While big names like Lil Miquela are ostensibly independents who can shill for the highest bidder (in addition to her own music and other products), the current big story, Seraphine, represents a new iteration on the idea: a virtual influencer born not from a marketing agency, but direct from a consumer brand and ultimately designed to sell said brand’s product exclusively.
At first Seraphine’s association with Riot Games wasn’t explicit and she appeared to be run like any other “run-of-the-mill” virtual influencer of the Lil Miquela bent. As time went on, however, the hints and references to League of Legends became more frequent, culminating in an official statement from Riot that Seraphine is a product owned directly by the company.
This is true vertical integration of the influencer marketing model. If it succeeds (Seraphine already has 326k Twitter followers and 394k on Instagram -- no numbers on what percentage of those are astroturfed) it not only represents a new way for corporations to wrest control of their marketing channel away from the clutches of independent influencers who fund a significant portion of their increasingly precarious existences off these marketing deals, but a slew of new monetization streams. After all, Seraphine isn’t just selling LoL, she’s part of an all-virtual K-pop group that’s dropped a single and has an EP in the works.
This is why it’s slightly annoying when I see the parallel phenomenon of vtubers seemingly conflated with virtual influencers. The business models are, for the time being, completely different.
While virtuals like Miquela and Seraphine are creating their own cultural products at a rapid rate, by and large they still exist to move some other product besides themselves. Hence the whole “influencer” name. Whereas sales is more or less the active coercion of you the consumer to purchase the product, all marketing is in effect an attempt to influence you to perform a specific action. That action could be paying for premium skins in LoL, subscribing to a newsletter, filling out a webform, or adding a product to your cart. Regardless of what it is, that action is the metric by which the success or failure of the marketing campaign is judged.
In contrast vtubers aren’t orchestrated marketing campaigns to persuade you to buy another product. Vtubers seem to exist as a product in and of themselves. The typical CTA, or Call to Action, you’re likely to hear on a vtuber stream is to subscribe to their YouTube or Twitch channel(compare to the average virtual influencer whose standard platform for “engagement” is social media), or give them a superchat to get your question/inane comment to the top of their docket. To my mind this puts them on a completely separate evolutionary path from virtual influencers. The core business of a vtuber agency like Hololive, Nijisanji, or Upd8 is to sell the vtubers themselves and pursue relentless growth of their audiences. For time being at least, they aren’t a lead gen op for anything else.
Another key difference is the nature of the personalities themselves. While it’s undeniable that both virtual influencers and vtubers are manufactured characters -- multiple Nijisanji and Hololive vtubers have made reference to auditioning for certain characters whose personalities were already drafted out in advance by the agency -- the artifice of the vtuber is far less absolute. The average virtual influencer may not express any overt expectation that any of their fans consider them a “real” person, but the consistency of their “canonical” personality is absolute. Most if not all the influencers are the work of an entire team of people managing their various social media presences, writing and producing the music, drawing and rendering the character art, etc. There’s no “actor” behind the proverbial mask to leak through.
Contrast that against vtubers where there’s some kayfabe around each character’s canonical personality, but the boundary between the character and performer is way more porous. While the earliest vtubers like Kizuna Ai, Eilene, Mirai Akari, and, to a lesser degree, Kaguya Luna maintained the pretense of being artificial intelligences speaking to you from direct from the internet, the current generation of agency-managed vtubers have no compunctions about acknowledging that there are real people behind the characters. Look at Hololive vtuber Oozora Subaru’s extended story about an awkward encounter with a foreign moving company worker, or Nijisanji vtuber Honma Himawari’s anecdote about actually auditioning to get a job at Nijisanji.
It’s obvious that while each vtuber’s personality is preconfigured by the agency before a performer is ever even hired to play them, there exists some level of play between the character’s personality and the actor’s personality bleeding through. This simply wouldn’t be possible with a virtual influencer as there’s no one creative-minded actor behind the character. I have a hunch this modicum of authenticity is a factor in the rather insane quarter-on-quarter superchat revenue both Hololive and Nijisanji vtubers have enjoyed as well as the jaw-dropping daily subscriber growth numbers of some of the top Hololive YouTube channels.
The irony is that at the same time vtubers are allowed a level of cartoonish eccentricity and irreverence that simply don’t seem possible within the strictures of a corporate-backed virtual influencer project. As I write this newsletter Hololive EN vtuber Gawr Gura, who debuted last month, has just achieved an unprecedented one million subscribers to her YouTube channel. This already dwarfs the aforementioned social media followings of Seraphine. (Gura’s Twitter following stands at 369k) This is also despite Gura being a gremlinesque shark-human hybrid in a great white shark onesie whose trademark is to say “a” and have millions of people lose their minds. Virtual influencers on the other hand come off much more like they were designed by committee, all potentially rough edges sanded away and the remaining traits tuned to emulate those of a standard real world Zoomer influencer personality. A corporate design agency’s conception of a personality who would inspire 10x growth and loyal parasocial attachment from followers. Even when virtual influencers dip their toes into “rougher” edges, such as Riot’s recent attempts to introduce mental health issues as a part of Seraphine’s character, it comes off as exceedingly hollow and the popular reaction is unsurprisingly negative.
But while vtubers remain a separate evolutionary branch for now there could be changes coming down the line as growth numbers in the time of COVID and quarantine show no signs of slowing. YAGOO has made it clear that he always envisioned Hololive as a new twist on the traditional Japanese idol agency business model. On its surface the idol agency model looks similar to traditional literary or Hollywood acting agencies, however on a deeper level the idol system is less about marketing their talent to get work in a specific trade but finding attractive youths and training/curating them in a variety of different talents and characteristics. The goal is not to create talented actors or pop musicians, but someone who can do all of the above while also projecting a girl/guy-next-door personality that will resonate with a big tent audience. The pop albums and movie appearances may be what the consumers purchase, but the true product they’re buying is the idol themselves and their manufactured personality. This is secret behind why so many idol-centric pop tracks and TV dramas from Japan tend to fall flat for foreign audiences who don’t exist within the current of idol development. It really isn’t about how good the showor music is, it’s about the fact that it’s the latest thing by [name]-chan and in this house we stan [name]-chan.
“The idol and their personality is the true product.” That doesn’t sound all that different from vtubers does it? While that’s true it’s also the case that traditional idol agencies are notorious for the amount of “curation” they do of their talent’s personalities. In pursuing that guy-next-door shine many idols more or less have to adopt 100% fictional personalities and even backstories -- which sounds way more virtual influencer than vtuber. I somehow doubt a heyday-era agency boss like Johnny Kitagawa would’ve been amused with his talent talking about who they ship each other with, mocking their own audiences, or discussing the kind of porn they like to look at before going to bed.
What remains to be seen is whether YAGOO and the bosses of the other agencies recognize a large part of why their product has seen so much growth is precisely because of their decidedly surreal and un-corporate candidness, and that fucking with that might cause things to go horribly wrong. Judging from recent big-name collaborations on the Hololive side though, I wouldn’t be surprised if there might be some changes coming down the pipe.