The Originality Premium

A research note on why creators selected for voice outperform follower-count selection on the metrics that compound — and why the crypto creator-agency model is structurally late to converge with what Web2 marketing settled five years ago.

§ 01 · The QuestionWhat does a crypto creator agency actually select for?

The crypto creator-agency category — KOL networks, distribution shops, attention-as-a-service desks — has operated for roughly five years on a single load-bearing selection criterion: follower count. The largest accounts get the campaigns. The campaigns inflate the largest accounts. The cycle compounds, and a small number of established voices end up doing a disproportionate share of all paid crypto coverage. This is well-documented and operates more or less as the industry's open secret.

What is less well-examined is whether this model is correct. Whether selecting on follower count actually produces the outcomes — for projects buying coverage, for creators building careers, for the markets ostensibly being moved — that the participants assume it does.

This memo argues that the follower-count model is structurally inferior to a voice-based selection model on every metric that compounds, and that the crypto creator economy is roughly five years behind Web2 in arriving at this conclusion. The data Web2 marketing settled in 2020–2022 — micro-influencer engagement premiums, parasocial trust dynamics, the failure of reach-based attribution — has yet to be priced into crypto creator selection. This memo lays out the evidence and proposes what the converged model looks like.

We are writing this because Triad operates by the converged model, and the question of why is the question this memo is built to answer.

§ 02 · The Web2 BenchmarkFive years of data on what follower count actually predicts.

The Web2 influencer-marketing industry — projected at $32.55 billion in 2025 global spend[1] — has gone through the same migration we expect crypto to undergo. The shift began around 2022 as brands started running rigorous post-campaign attribution and finding that follower count was a poor predictor of campaign outcomes.

The headline finding, reproduced across multiple independent datasets, is that engagement rate scales inversely with audience size. A creator with 30,000 followers generates a higher engagement rate per post than a creator with 1,000,000 followers — not by a small margin, but by 3x to 7x depending on the platform and methodology.

The conventional response to this data is that engagement rate understates the value of macro-creators because their raw audience size compensates: a 2% rate on 1M followers is 20,000 engagements, materially more than a 7% rate on 30K followers (2,100 engagements). This is mathematically correct and substantively misleading. The conversion economics work differently.

Conversion economics

When campaigns are measured on action — sign-ups, transactions, retained users — not impressions, the picture inverts. Brands report 3–5x higher ROI from micro-influencer campaigns versus macro spend[5]. 61% of brands explicitly state higher ROI from micros than macros, and micro-led campaigns produce 28% higher repeat-customer rates[2]. The cost gap is equally pronounced:

The implication is straightforward: follower count is a vanity proxy. It correlates loosely with reach but inversely with the qualities that produce campaign outcomes — depth of audience attention, trust per follower, and the willingness of an audience to act on a recommendation. The Web2 industry has been pricing this in for three years.

The parasocial mechanism

Why does the inversion exist? The academic literature on parasocial relationships in influencer marketing is now substantial and consistent. The 2019 study published in the Journal of Marketing Management by Reinikainen et al.[7] established that perceived authenticity moderates credibility more strongly than reach in predicting purchase intent. Subsequent work (van Driel & Dumitrica, Convergence, 2018[8]; Schwemmer & Ziewiecki, Social Media + Society, 2018[9]) extended this: as creators professionalise and audiences perceive content as paid, conversion drops sharply.

The intuition behind the data is simple. Audiences form one-sided emotional bonds with creators whose voice they trust. Trust accumulates with consistency, originality, and visible authenticity over time. It does not accumulate with reach. A creator with 30,000 followers who has built voice over five years has more trust per follower than a creator with 1,000,000 who has built reach through campaign volume. When a brand wants action from an audience, trust is what converts.

"It's no longer about the old-school influencer marketing considerations of how many followers someone may have, but instead taking a media-based approach — retention rates, view-through rates, and individual audiences' trust in their creator of choice." Digiday, January 2026 — on the maturation of the creator economy[10]

§ 03 · The Crypto LagWhy the same data has not yet priced in.

If the Web2 conclusion is so well-established, the question becomes: why does the crypto creator-agency category still operate predominantly on follower count? Three explanations, in descending order of importance.

First: closed attribution loops

The Web2 migration happened because brands ran post-campaign attribution and found macro-spend uneconomic. Crypto campaigns, by contrast, rarely measure against true conversion. The dominant metric — impressions on paid posts — is one of the few metrics where macro accounts genuinely lead. If you optimise for impressions, you select for the accounts that generate them. The loop closes before any quality signal can break it.

This is reinforced by the fact that crypto projects buying coverage are often optimising for narrative ignition rather than retained users. A token launch needs a 72-hour spike of attention to coincide with TGE. After that, the campaign is judged on price action, not user retention. Macro accounts produce spikes. Micro accounts produce retention. The category structurally rewards what it measures.

Second: incentive misalignment in agencies

Agencies whose business model is take-rate on paid creator placements have direct economic incentive to maintain the macro-centric model. Higher placement values mean higher take. A campaign placing $50,000 across 5 macro accounts produces meaningfully more agency revenue than a campaign placing $50,000 across 50 micro accounts, even if the latter outperforms on attribution. This is not a critique of any specific firm — it is a structural feature of how take-rate models operate.

Third: creator concentration is self-reinforcing

Once a small set of macro voices dominates campaign volume, agencies build infrastructure around managing those relationships. Pipeline, contracts, rate cards, payment rails, content review — all calibrated for a relatively small roster of high-volume accounts. Expanding to 100+ micro accounts requires fundamentally different operational infrastructure. The switching cost is real, and the incumbents who would need to bear it are the ones the model is currently serving best.

The result is a category where the data Web2 settled five years ago is structurally suppressed. Every participant has a local incentive to maintain the system; no participant has standing to challenge it. The challenge has to come from outside.

§ 04 · Crypto Case EvidenceWhat happens when reach-based campaigns meet reality.

The follower-count model is most clearly stressed in the cases where it most visibly fails. Two recent examples — both well-documented in public reporting — illustrate the gap between paid amplification and durable outcomes.

Neither case is being cited as an indictment of any specific creator or agency. Many of the accounts that covered Movement and Mantra are excellent creators producing genuine work; the responsibility for what happened lies with the projects' internal governance and the market structure around them, not with the people paid to amplify the launches. The point is narrower: paid amplification, no matter how aggressively deployed, does not produce the audience qualities that determine post-launch outcomes. Coverage is upstream of price; resilience is downstream of trust. These are different mechanisms.

This is the gap the Web2 industry priced in five years ago, and the gap crypto is now beginning to price in — most visibly through tools like Kaito.

§ 05 · The Kaito SignalWhat the rise of mindshare scoring tells us.

Kaito's emergence is the most legible signal that crypto is beginning its own version of the Web2 migration. The platform's "Yapper Leaderboard" — which uses AI scoring to rank creators by mindshare rather than follower count or paid placement[15] — surfaces a parallel selection mechanism. As of early 2025, the program had 250,000+ active users, and the broader InfoFi category reached a $335M market cap with sustained $50M+ daily trading volume[15]. Major projects (Polygon, Berachain, Monad, Eclipse, MegaETH) run campaigns through the leaderboard with $30,000/month reward pools targeting top voices[16].

The mechanics matter. Kaito's scoring system explicitly downweights:

• →Repetitive content lacking analysis
• →Posts below qualified-engagement thresholds (reputation-weighted, not raw impressions)
• →Engagement from low-quality accounts

And explicitly rewards:

• +Depth of analysis and topical originality
• +Engagement from other high-mindshare voices (peer signal)
• +Consistency over time rather than spike behaviour

One Chinese-language analysis[17] made the implication explicit:

"Instead of spending money to have a KOL post an ad, it is more effective to allocate a portion of the budget as a community reward to incentivise engagement on Kaito. KOLs must demonstrate their genuine insights into the project to earn community approval. The era of presumptive trust has concluded." BlockBeats analysis on the Kaito Yap model, 2025[17]

The market is reaching the conclusion through infrastructure rather than discourse: the metric crypto needed to start measuring already exists, the platform that measures it has reached scale, and the projects deploying budget against it are running the experiment in real time. The question is no longer whether the migration happens — Kaito's adoption proves it is. The question is which agencies finish on the right side of it.

§ 06 · What the Converged Model Looks LikeSelection on voice, contract on output, scale through curation.

If the Web2 evidence is correct and the Kaito signal indicates crypto is converging in the same direction, the operational question becomes: what does a creator agency selecting on voice rather than follower count actually do differently? Four structural differences, all of which Triad has built around.

The voice-selection model has one significant operational disadvantage: discovery is hard. Identifying genuinely original creators with 10K–50K followers requires reading work rather than pulling a leaderboard. It does not scale through automation. This is the part of the model that takes patience.

It also has one significant advantage that compounds over time: the roster appreciates. A creator brought into an agency at 12K followers, given access to substantive briefs and partner introductions, who grows to 45K over 18 months on the strength of their own voice, becomes more valuable to the agency every quarter. A macro-creator brought in at 500K, who continues to take paid campaigns indefinitely, depreciates as their audience treats their feed as ad inventory. The math on which model produces a more valuable roster over a five-year horizon is not close.

§ 07 · ImplicationsFor projects, for creators, for the category.

For projects buying coverage

The 2025 collapses make clear that there is no longer a credible argument that aggressive paid amplification produces durable outcomes. Projects launching tokens or campaigns in 2026 should be running attribution against retained holders / users at T+90 and T+180, not impression counts at T+7. Where paid coverage is used, it should be structured as a smaller component of a broader distribution strategy that includes mindshare-scored organic engagement and curated micro-influencer placement. Budgets allocated entirely to macro-account inventory are misallocated against the available evidence.

For creators

The premium on voice is real and growing. Creators currently in the 10K–80K follower range who continue producing original work — analysis, framing, conviction — are the demographic that the converging market is most likely to reward over a three-to-five-year horizon. The temptation to optimise for short-term paid placements should be measured against the long-term cost of audience erosion. A feed that reads as ad inventory has a ceiling. A feed that reads as original analysis does not.

For the category

Crypto creator agencies that have built infrastructure around managing a small roster of macro-accounts are not structurally positioned to migrate. Their pipeline, pricing, and operational expertise are all calibrated for the model that the data suggests is in decline. The agencies that will succeed in the converged market are either new entrants building on the voice-selection model from day one, or incumbents capable of running both models in parallel during a multi-year transition. The middle position — half-committed macro shops trying to add micro pipelines — is the most exposed.

§ 08 · Where Triad OperatesThe thesis applied.

Triad's creator practice is built on the voice-selection model — explicitly, structurally, from the founding cohort onward. The cohort is capped at 100 seats. Selection is by application and partner interview, not follower count. Roster economics are calibrated for creators in the 10K–100K range producing original work, with retainers structured to give creators stability to focus on long-form output rather than chasing short-term placement income.

This is not a critique of how the established crypto agencies operate. Many of them are excellent firms run by serious people, and a paid-amplification campaign run against the right project at the right moment is genuinely valuable. The argument is narrower: the converged market will require a meaningfully different selection model than the one that built the current category, and Triad is built for the converged market.

If you are a creator producing original work, the application is open. If you are a project evaluating distribution strategy for 2026, the question we would invite you to ask is what your campaign produces at T+180 days — and whether the answer to that question matches the campaign budget you've allocated.

The full apparatus — application, roster, retainer structure, campaign workflow — is at triadnetwork.xyz/creators.