Pay-Per-Post Turns Communities Into Content Mills

Opening

Reader, if you’ve ever run a community or a product, this meeting scene will feel familiar. Right around when engagement starts to sag, someone floats the idea: “What if we put points on every post? Wouldn’t that get things moving again?” It sounds intuitively right. People respond to incentives, after all.

But an analysis of exactly this idea, tested at a scale of 570,000 people with real money on the line, was just presented at an international conference. And by coincidence, at almost the same moment, one of the world’s largest “pay-per-post” experiments was forcibly shut down in the market.

Today, I want to read these two events together. Here’s the conclusion up front: paying for posts doesn’t revitalize a community. It turns the community into a subcontracted content market. And the fork between success and failure isn’t the size of the reward — it’s cashability.

---

We Paid 570,000 People to Post

Farcaster is a blockchain-based social network. What makes this platform precious to researchers is something specific: several distinct kinds of rewards coexist within a single service. There’s the official reward the platform pays out in the stablecoin USDC, third-party volatile-token rewards like DEGEN or MOXIE, and tips that users send each other directly. In effect, it’s a natural laboratory for reward design.

An international team of researchers analyzed this laboratory’s records in full and presented the findings at ACM SIGMETRICS 2026, the leading conference in computer systems measurement. The sample covered 574,829 users who had linked wallets — 64.25% of all users. As someone who teaches data analysis, if I had to name this study’s single virtue, it’s that it went after causation, not correlation. Rather than simply noting “people who got paid wrote more,” the researchers estimated the effect using difference-in-differences¹, comparing changes before and after rewards were introduced against a control group.

The results boil down to three lines. First, most token rewards increased the volume of posts and replies. Second, quality didn’t improve — under some conditions it actually dropped. Third, users who repeatedly received algorithmic rewards showed a cumulative pattern of learning not how to write well, but behavior that games the reward system. There’s also the finding that competition to gain followers intensified, while the act of following others didn’t increase at all. In other words, people started broadcasting instead of conversing.

But there’s one part of the paper I lingered on longest. Farcaster’s official USDC reward — a predictable, cash-like payout — had no significant effect on the volume of posts and replies. What actually moved behavior was the volatile token, the one whose price swings wildly. What moved people wasn’t a wage; it was a lottery ticket. Fittingly, the Gini coefficient of reward distribution ranged from 0.72 to 0.94 — a lottery-like structure where a tiny few captured most of the payout.

To be fair, the paper isn’t all bleak. Peer-to-peer tips flowed 1.3 to 2 times more often to people outside a user’s follow network, which eased the tendency toward cliquish clustering. Today’s focus, though, is the side that community operators actually design — the pay-per-post rewards an algorithm dispenses.

The Same Experiment Ran Four Times

Is this result just a quirk of crypto culture? Trace the lineage back, and the same experiment has run at least four times.

The first was Steemit. Launched in 2016, this blockchain social network paid writers in STEEM coin, which could be cashed out on exchanges when a post got upvoted. It has deep ties to Korea, too — by the community’s own 2017 analysis, Korean was the second-largest language on Steemit after English. Many readers will remember what happened next: controversies over self-voting (upvoting your own posts) and vote-trading rings dragged on, and when the coin’s price crashed, the community sank with it. In 2020, the founder sold the platform and the community splintered, and the experiment’s momentum effectively died.

The second was Quora. In 2018 it created the Partner Program, splitting ad revenue on “questions” — and questions were mass-produced in exactly that unit. A question factory ground away, chasing pageviews. The English-language program shut down in September 2022, and the remaining language versions followed in March 2023.

The third is X. Starting in 2023, it began sharing revenue tied to engagement metrics, and what kind of posts proliferated on the timeline afterward hardly needs explaining.

The fourth is the experiment that just ended: Kaito’s Yaps. Launched in December 2024, Yaps was a system that measured influence and awarded points when users posted about crypto projects on X. According to reports citing figures aggregated by the data platform Dune, monthly active participants topped 200,000. What’s interesting is that Kaito fought farming² fairly seriously. It touted an algorithm that supposedly evaluated content quality and authenticity rather than raw activity, and it kept tightening eligibility and ranking criteria. Even so, both Korean and international reporting consistently noted that posts climbing the rankings with sensational images and repetitive comments kept appearing.

The ending came on January 15 this year. Nikita Bier, X’s head of product, announced the platform would cut off API access for apps that pay rewards for posts, citing a surge in low-quality AI-churned posts and comment spam originating from InfoFi³. Kaito surrendered the same day. After 13 months, it shut down Yaps and its leaderboard and pivoted to Kaito Studio, a curated marketplace where brands directly select and contract with vetted creators. Reports said the token price dropped more than 20% right after the announcement. Cookie DAO, another InfoFi service, announced the end of its own reward program within 10 minutes of X’s announcement, and word spread that a Yaps participant community of roughly 157,000 people had been banned from X. Notably, international reporting pointed out that Korean users made up a particularly large share of these participants.

Two points are worth flagging here. One: AI was the final nail in this model’s coffin. In a world where the marginal cost of producing a single post is effectively zero, “pay-per-unit” simply doesn’t add up. One Silicon Valley investor summed it up this way: “The median crypto-Twitter post has become Kaito slop.” The other is the counter-argument: since X itself runs its own pay-for-engagement revenue share while it was InfoFi specifically that got cut off, some critics argue the real issue was never the reward model itself but who gets to keep the fee. That criticism isn’t wrong in principle — after all, the quality problems plaguing X’s own timeline stem from the same structure.

But right here, the real question remains: if every money-based reward ends this way, what was different about the reward systems that are still working just fine?

The Fork Isn’t the Amount — It’s Cashability

Think about it, and plenty of communities already run on rewards. Naver Knowledge iN, Korea’s Q&A platform, gives out “Naegong” points; Stack Overflow gives reputation scores; Danggeun, Korea’s neighborhood marketplace app, gives “manner temperature” ratings. Gamification — designing tiers and badges — has long been a standard tool of community management. Same idea, reward-wise. So why does this kind survive while pay-per-post dies?

The difference lies in whether the reward can be cashed out beyond the community. Naegong, reputation, and manner temperature are all status currencies that only work inside that community. No matter how much you farm them, you can’t turn them into an hourly wage, so the profit-and-loss math of farming never even gets off the ground. More important is which way the incentives point. Status only holds value as long as the community stays healthy. If I wreck the place, my reward disappears with it. Cash and tokens, by contrast, leave the building. Even if the community collapses, the money I’ve collected stays in my account. At that point, treating the community as disposable becomes the individually rational strategy. This is where the grammar of the subcontracted market begins.

Money changing hands doesn’t just change the math — it changes how people feel. Behavioral economics calls this motivation crowding-out⁴. The famous stage for this experiment was a group of daycares in Haifa, Israel. When parents who picked up their kids late were fined, lateness didn’t fall — it rose. The social norm of feeling guilty had been replaced by a price tag: “just a service you can pay for.” Even after the fine was removed, lateness never returned to its original level. Research on blood donation points the same way. In 1970, sociologist Richard Titmuss argued that paying for blood donations could actually reduce giving, and a 2008 Swedish field experiment partially confirmed this. But the real twist in that experiment comes next: when donors were offered payment with the option to redirect it to charity, the crowding-out effect disappeared. Strip the cashability out of the money, and the norm was restored. Rarely does evidence show so cleanly that cashability is the switch.

Korea carries deep traces of this same experiment too. Take reviews. As photo-review reward points and “experience group” sponsorships (product-seeding programs) became commonplace, formulaic reviews were mass-produced, and distrust toward sponsored reviews built up until it erupted in the 2020 “hidden advertising” scandal. That year, the Fair Trade Commission tightened its ad-disclosure guidelines. What I find especially telling is a phrase born out of this process: naedon-naesan (“I bought it with my own money”). It means trust had gotten to the point where you had to actively prove you weren’t paid to be believed. It reads like a gravestone marking the spot where reviews — a community asset — were handed over to advertising subcontracting because of pay-per-post money.

So what if we just shrink the amount? Unfortunately, a middling sum is the worst option of all. The title of another paper by Gneezy and Rustichini, the economists behind the Haifa experiment, says it best: “Pay Enough or Don’t Pay at All.” A small reward crowds out intrinsic motivation without being large enough to fill the vacuum it creates.

Oswald’s Lens

This subject isn’t theoretical for me. I say this as someone who personally built Notion’s early Korean community. Back then, the community grew explosively without any money changing hands. What made people build templates and write how-to posts all night wasn’t cash. It was the chance to present on stage at meetups, the status of being an “ambassador,” early access to new features and the team behind them, and the identity of being “someone who’s good at Notion.” Every one of those is a currency that can’t be cashed out beyond the community. What if I’d put ₩5,000 (~$3.6) on every post back then? Template sharing would have become a delivery, and I would have become a quality-control clerk instead of a community lead.

So when I advise on reward design, my principle comes down to two lines. First: reward the majority of members with currency that can’t be cashed out — status, a stage, access, and mechanisms that make contribution visible. Second: don’t scatter money throughout the community — pay it through contracts with a small, vetted set of professionals. As you may have noticed, this is essentially the answer Kaito arrived at after paying two years of tuition. The Studio model is exactly this structure. One line from a go-to-market perspective: the moment you burn money per unit inside a community, it stops being a community and becomes a low-cost-per-unit advertising channel. And the fate of any such channel always converges on a race to the bottom on price.

Closing

Let me sum up. Pay-per-post increases the number of posts, but it can’t buy quality or trust. Data from 570,000 people and four separate market experiments all point to the same ending. The fork in the road is cashability. Rewards that convert outside the community make farming a rational strategy; status currency that only works inside aligns everyone’s interests.

Try this exercise this week: split the list of rewards you currently run into two columns — “cashable outside” and “only works inside.” If the first column is tied to content volume or view counts, farming is a future you’ve already booked.

I’d love to hear about your experience on the ground, Reader. Whether you offered points or fees and watched farming take over, or moved people with non-monetary rewards like tiers and badges, tell me in the comments. If enough cases come in, I’ll turn them into a “Korean reward-design casebook” in a future issue.

---

💬 Share your reward-design successes and failures in the comments — I’ll fold them into a future issue. 📨 If you know a colleague wrestling with community or reward policy, pass this along.

---

References & Further Reading

Primary sources

• Yang, W. et al., “Beyond Single-Tokenomics: How Farcaster’s Pluralistic Incentives Reshape Social Networking”, Proceedings of the ACM on Measurement and Analysis of Computing Systems (ACM SIGMETRICS 2026), 2025. Link ··· This is the backbone of today’s piece. The causal-analysis table mapping which reward moved which behavior is the highlight.
• The Block, “X users celebrate crackdown on ‘plague’ of AI-led reply spam as InfoFi platforms seek alternatives”, 2026.1. Link ··· Lays out the industry’s for-and-against arguments at the moment InfoFi got cut off. You can also find the counter-argument targeting X’s own revenue share here.
• Gneezy, U. & Rustichini, A., “A Fine Is a Price”, The Journal of Legal Studies, 2000. Link ··· The original source of the Haifa daycare experiment. It’s short, so it’s a good one to read in full.
• Mellström, C. & Johannesson, M., “Crowding Out in Blood Donation: Was Titmuss Right?”, Journal of the European Economic Association, 2008. Link ··· The source of the twist that “adding a charity-donation option eliminates the crowding-out effect.”

Background

• Titmuss, R., The Gift Relationship: From Human Blood to Social Policy, Allen & Unwin, 1970. ··· The classic that started the debate over rewards versus norms.
• CoinGecko, “What Is Kaito? 2026 Guide to Studio, Markets & KAITO Token”, 2026. Link ··· Shows Kaito’s whole trajectory at a glance, from the end of Yaps to the pivot to Studio.
• TechCrunch, “Quora shutting down English version of Partner Program”, 2022.8. Link ··· The record of how the experiment of paying for questions came to an end.
• Blockmedia, “Google Cuts Apps, X Cuts Yapping⋯Big Tech’s Blade Shakes the Digital Asset Industry,” 2026.1. Link ··· Korean-language reporting summarizing the situation before and after the yapping ban.

Past issues worth reading alongside this one

• A Pay Stub for 30,000 Followers ··· Covers the opposite route money takes to reach creators. Read it alongside this piece to see why commissions tied to sales performance and pay-per-post fees tied to volume meet such different fates.
• The AI That Aced the Math Olympiad — With No Proctor Watching ··· Covers how measurement breaks down when optimization pressure gets applied to a metric. It’s the assessment-side version of today’s story.

---

📝 Glossary

The author, Kwangseob Ahn, is a professor of business administration at Sejong University and lead consultant at OBF (Oswarld Boutique Consulting Firm). He teaches statistics and data analysis, including business data management and business analytics, while leading GTM and AI strategy consulting in the field, designing the seam between technology and business. He has published academic research on a memory architecture for AI dialogue systems (HEMA) and runs Daily Arxiv, a daily curation of global AI papers. He holds a master's from Korea University's Graduate School of Technology Management and a KMBA. He is the author of Homo Brainless: The People Who Outsource Their Thinking.

Footnotes

1. Difference-in-Differences: A statistical technique that estimates a treatment’s pure effect by subtracting the before-and-after change in a treated group from the before-and-after change in an untreated comparison group. ↩

2. Farming: Exploiting loopholes in a reward system by repeatedly satisfying the reward’s conditions — regardless of the actual value of the content — to harvest gains. The term comes from repeatedly hunting for items in video games. ↩

3. InfoFi (Information Finance): A crypto-industry model that measures information and attention like financial assets and rewards them with tokens. Services that pay points or coins for writing posts fall into this category. ↩

4. Motivation crowding-out: A phenomenon where an external reward displaces intrinsic motivation (enjoyment, sense of duty, norms), so that behavior decreases or quality drops even though a reward was introduced. ↩