Originality, Above Everything

Why Triad selects creators on voice instead of follower count. A founder note, with the receipts attached.

§ 01The one number

For five years, every KOL shop in crypto has selected on one number: follower count. The biggest accounts get the deals. Size is the product. We built Triad on the opposite bet, and this note is me explaining why.

The follower-count loop is simple. Large accounts get the campaigns. Campaigns inflate the large accounts. A small circle of established voices ends up doing most of the paid coverage in the industry, and everybody involved knows it. It is the category's open secret.

What almost nobody asks is whether the model is correct. Whether selecting on size actually produces the outcomes that projects pay for and creators build careers on. Web2 asked that question years ago. It got a clear answer. Crypto has mostly not looked.

§ 02What Web2 already settled

Influencer marketing outside crypto is a $32.55 billion industry this year[1], and around 2022 it started doing something crypto still rarely does: measuring campaigns against real conversion instead of impressions. The finding that came back, again and again, across independent datasets, is that engagement rate runs inverse to audience size. Not slightly. By 3x to 7x.

The standard rebuttal is arithmetic: 1.8% of a million followers is still more raw engagement than 7% of thirty thousand. True. It is also beside the point, because impressions are not what anyone is actually buying. When campaigns get measured on action, the picture flips. Brands report 3 to 5x higher ROI from micro-creator campaigns than macro spend[5]. 61% of brands say micros beat macros on ROI outright, and micro-led campaigns produce 28% higher repeat-customer rates[2]. Then there is the price of admission:

Follower count is a vanity proxy. It tracks reach, loosely, and runs inverse to the things that actually convert: attention depth, trust per follower, an audience willing to act on a recommendation.

The academic work explains why. Audiences form one-sided bonds with creators whose voice they trust, and that trust is what converts. Reinikainen and colleagues showed in the Journal of Marketing Management that perceived authenticity predicts purchase intent more strongly than reach does[7]. Follow-on studies found the flip side: the more an audience reads a feed as paid, the less it converts[8][9]. Trust accumulates with consistency and originality over time. It does not accumulate with reach[10].

§ 03Why crypto has not priced it in

If the evidence is that settled, why does crypto still buy on follower count? I see three reasons, and I have watched all of them operate from inside the industry.

First, crypto campaigns rarely measure against conversion. The metric that gets reported is impressions on paid posts, which happens to be one of the few metrics where macro accounts genuinely lead. Optimise for impressions and you will select for the accounts that produce them. The loop closes before any quality signal can break it. It does not help that most token launches are optimising for a 72-hour attention spike around TGE rather than retained users. Macro accounts produce spikes. Micro accounts produce retention. The category rewards what it measures.

Second, follow the agency economics. A shop that earns take-rate on placements makes more money putting $50,000 across five macro accounts than across fifty micro accounts, even when the fifty outperform on every attribution metric. That is not corruption. It is just how take-rate models behave, and it is why the pressure to change will not come from inside.

Third, the incumbents are built for the old model. Pipeline, contracts, rate cards, payment rails, all calibrated for managing thirty relationships, not three hundred. Switching costs are real, and the firms that would have to pay them are the ones the current model serves best.

Everyone has a local incentive to keep the system running. Nobody inside it has standing to challenge it. So the challenge comes from outside.

§ 04What it looks like when it fails

You can see the model stress-tested in public, because 2025 supplied two clean experiments.

To be clear about where the blame sits: not with the creators who took those briefs. Plenty of them are excellent at their jobs, and the failures above were governance failures inside the projects. The point is narrower. Coverage is upstream of price. Resilience is downstream of trust. Paid amplification, at any budget, buys the first and cannot buy the second.

§ 05The Kaito signal

The market is now correcting this with infrastructure rather than discourse. Kaito's Yapper Leaderboard scores creators on mindshare instead of follower count or paid placement[15], and it has reached scale: 250,000+ active users, an InfoFi category at a $335M market cap with $50M+ daily volume[15], and serious projects (Polygon, Berachain, Monad, Eclipse, MegaETH) running $30,000-a-month reward pools through it[16].

Look at what the scoring punishes:

• →Repetitive content with no analysis
• →Engagement below reputation-weighted thresholds
• →Engagement from low-quality accounts

And what it rewards:

• +Depth of analysis and topical originality
• +Engagement from other high-mindshare voices
• +Consistency over time rather than spikes

That is a machine built to price originality. One analysis of the model put it plainly[17]:

"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 migration is not a prediction anymore. It is running, with real budgets. The only open question is which agencies end up on the right side of it.

§ 06The model we run

So what does an agency actually do differently when it selects on voice? Side by side:

I will be honest about the cost of the right-hand column: discovery is brutal. Finding genuinely original creators at 10K to 50K followers means reading their work, not pulling a leaderboard. It does not automate. That is the tax, and we pay it every week.

Here is what you get for the tax: the roster appreciates. A creator we sign at 12K followers, who gets substantive briefs and real introductions and grows to 45K on the strength of their own voice, becomes more valuable to us every quarter. A macro account signed at 500K that keeps taking paid campaigns depreciates, because its audience learns to read the feed as ad inventory. Run that math over five years. It is not close.

§ 07What I would tell you

If you run a project: after 2025, there is no serious argument left that aggressive paid amplification buys durable outcomes. Measure your campaigns on retained holders and users at T+90 and T+180, not impressions at T+7. Use paid coverage as one component, not the strategy. A budget spent entirely on macro inventory is misallocated against the evidence above.

If you create: the premium on voice is real and it is growing. If you are at 10K to 80K followers producing original analysis, you are the demographic this migration rewards over the next three to five years. Every short-term placement you take prices against long-term audience erosion. A feed that reads as ad inventory has a ceiling. A feed that reads as original work does not.

If you run an agency: the shops built around thirty macro relationships are not positioned to migrate, and the half-committed middle position is the most exposed one on the board.

§ 08Where we stand

Triad runs the voice-selection model structurally, from the founding cohort onward. One hundred seats, filled by application and interview, never by follower count. Economics calibrated for creators in the 10K to 100K range producing original work, with retainers built so they can focus on long-form output instead of chasing placements.

None of this is a shot at the established firms. Some of them are run by serious people, and a well-timed amplification campaign has its place. My claim is narrower: the converged market will select for a different model than the one that built this category, and we built Triad for the converged market.

If you create original work, the application is open. If you are planning distribution for 2026, ask one question of every campaign you are quoted: what does it produce at T+180? Then check whether the answer matches the budget.

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