The Creator Platform Playbook

Before getting into tactics, I want to explain the core system. Everything we built at Tiki orbited a single loop. If you understand the loop, you understand everything else. If the loop breaks at any point, the whole platform dies.

We called it the Creator Flywheel. Every creator platform lives or dies by this loop:

Real Creators produce Original Content. Original Content retains Users. Retained Users become Creators. And the cycle repeats.

Every successful platform has a flywheel. Amazon's is: lower prices attract more buyers, which attracts more sellers, which enables lower prices. Uber's is: more drivers reduce wait times, which attracts more riders, which attracts more drivers. Ours was: verified real creators produce exclusive original content, which keeps users on the app, and engaged users eventually start creating themselves, which feeds more original content back into the loop.

The critical difference between our flywheel and our competitors' is what starts the loop. Most short-video platforms start with content. They fill the feed first, worry about who made it later. Content-first means you optimize for volume and virality. The result is a feed full of recycled clips (non-original content) because recycled clips are cheap and easy to produce at scale.

We started with creators. Creator-first means you optimize for the people making the content, not just the content itself. The result is a platform where the content is original, because the creators are invested in their own work and held to standards.

This seems like a small philosophical difference. In practice, it changes every product decision, every policy, and every outcome. The rest of this essay explains each spoke of the flywheel.

The first spoke of the flywheel is Real Creators. The question is: how do you attract, develop, and retain genuine talent when your competitors are paying creators fixed salaries to exist on their platforms?

Eugene Wei wrote the definitive essay on how social networks function as status marketplaces. His core argument is that people are status-seeking, and the best platforms offer a unique form of social capital earned through distinctive proof of work. His key line: "Status isn't worth much if there's no skill and effort required to mine it. If everyone can achieve it, it's no status at all. It's a participation trophy."

We took that principle and turned it into an operating system.

We called it Fame as a Service.

What creators actually want

When we surveyed our top-tier creators and asked them to rank what mattered most, we expected money to dominate. It didn't. The top demand was traffic: guaranteed visibility in exchange for quality work. Fame ranked significantly higher than cash.

This makes intuitive sense once you think about it. On most platforms, the algorithm is a black box. Creators pour hours into a video with no guarantee anyone will see it. One video gets massive reach; the next gets almost nothing. The difference isn't effort or talent. It's randomness. For a creator in a small Indian city, that randomness is maddening. A viral video could change their life. But the algorithm feels like a lottery, and lotteries make terrible career foundations.

What creators wanted was a meritocracy. Put in the work, get the eyeballs. A clear relationship between effort and outcome.

We could offer that. Traffic was our most abundant resource. We could guarantee it in exchange for quality, and we just needed for that to be a valuable resource.

The three tiers (Role-Playing Game)

We wanted to design a creator progression system that is transparent, inspired by role-playing games. Think of it as a character leveling system, except for content creators. Every person who joined Tiki entered at the base level. From there, they could work their way up through three tiers, each with transparent requirements and specific rewards.

The specific traffic allocations matter. On most platforms, the top 1% of creators get 90% of the algorithmic love, and nobody knows why. We published the rules. Every creator could see exactly where they stood and what they needed to do to level up. The system was transparent, which meant it felt fair, even when the competition was fierce.

Killing the salary and introducing Tiki Stars

In our cold start phase, we initially paid creators fixed monthly salaries, matching what competitors were offering. Then we looked at the data and saw a pattern: salaried creators were coasting. They posted the minimum their contracts required and collected their paychecks. They weren't building audiences or improving their craft. They were employees pretending to be creators.

So we killed fixed salaries entirely and replaced them with Tiki Stars.

Stars were a virtual currency that fans could purchase and gift to creators they liked. This was fundamentally different from a "like" button, which is free and has been inflated into meaninglessness on every major platform. A Star cost real money. Giving one was a deliberate choice, a sincere vote from a fan to a creator they valued. Creators could convert Stars into real income.

The shift changed the entire economic relationship. Before, creators performed for the platform's management to keep their contracts. After, they performed for their fans to earn Stars. The power moved from the platform to the audience, which is where it belongs. Users are the judge.

The downgrade mechanic

The most controversial design decision, and the one where we had many team disagreements, was making Blue Elite Verification tier status temporary. Every month, we ranked creators. Bottom performers got downgraded. They lost the badge, the traffic, everything. Meanwhile, unknown creators from small towns could fight their way up from nothing purely on merit.

This was not popular with the creators who got downgraded. Some called our team in tears. Competitors tried to poach talent every time we ran the monthly rankings. There were weeks where I genuinely didn't know if we'd destroyed our supply chain.

We hadn't. Nearly half of our Blue tier creators were eventually homegrown: people who started as anonymous newcomers and earned their way to the top. The system self-selected for the creators who actually believed in their talent and were willing to work for their audience rather than coast on a guaranteed paycheck.

For new creators, we had the trainee challenge to break into verified status. They had to submit a talent video and our moderation team would check if it passes for talent. The pass rate was 0.15%. Harder to get into than Harvard.

When something is hard to earn, people feel it's worth having.

The second spoke of the flywheel is Original Content. This is where we made the bet most people in our industry thought was crazy.

Our initial content that generated the most raw views on Tiki was garbage. Random selfies with beauty filters. Screen-grabbed movie clips. Low-effort lip-syncs. This content spread fast because it was easy to consume. But when we analyzed who was actually staying on the platform long-term versus who was churning after a few days, junk content was the problem. Users who discovered us through viral junk had the highest first-day engagement and the highest churn by the end of the first week. They came for the hit, snacked on a few more videos, and never came back.

Worse, junk content was crowding out the good stuff. The algorithm, optimized for engagement, was learning that cheap clips got clicks. It was promoting low-effort content over high-effort content. The creators who were actually investing in their craft couldn't break through the noise.

I thought of this as a restaurant problem. Our competitors were running all-you-can-eat buffets: serve everything, let users sort through the pile. We decided to run a curated dining restaurant. Fewer dishes, higher quality, curated by people who actually cared about the food. Think of Netflix vs Mass Television with ads.

The move most people thought would kill us

We created a content category called "Meaningless." Videos with no story, no talent, no effort. Even if they were going viral, we suppressed their reach. We refused to let the algorithm optimize for empty calories.

We set four explicit content guidelines and communicated them to every creator directly:

If your video didn't do at least one of these things, it wouldn't get distribution. Our moderation team, originally hired only for safety review, was retrained to evaluate content quality.

Then we imposed what we called the HD Mandate. Top-tier content had to be high definition. We taught creators about lighting, framing, and production basics. We provided tutorials and one-on-one guidance.

The result was that our platform looked visibly different from every competitor. Users noticed. Their most common first impression of Tiki was the content quality. We hit 100% original content, compared to 2% on Moj and 8% on Josh.

The quality was the brand.

Democratic quality control: the Judge Panel

India is practically a continent. A comedy format that works in one state may be meaningless in another. Languages, cultural references, humor styles vary wildly between regions. Our moderation team, sitting in an office, didn't have the cultural knowledge to judge content quality across dozens of regions and languages.

So we built a Judge Panel of core community members. Not employees. Our most dedicated users, people who intimately understood their local creative scenes. We gave them actual authority to decide what should be promoted and what shouldn't.

The Panel served two purposes. First, it was a quality filter that was culturally calibrated in a way no centralized team could ever be. Second, it generated a constant stream of local intelligence that we used to improve our automated systems over time. The Panel didn't replace the algorithm. It trained the algorithm.

What it cost, and what it returned

Total view counts dropped after the purge. Internal dashboards went red. Some people on the team thought we were killing the company.

But the metrics that actually predicted long-term success moved in the right direction. Users watched content longer. They shared it more, meaning they were actively recommending what they found rather than passively scrolling. New-user retention improved because the first experience of the app was no longer a wall of noise. We had a ~70% New User Day 1 Retention Rate.

The supply side responded too. When creators saw that the path to fame ran through quality rather than volume, the middle tier started producing better work. The aspiration gradient steepened across the whole platform without us having to coach anyone individually.

The third spoke of the flywheel is Retained Users. This is where we made our most counterintuitive discovery.

Every social app founder I know is obsessed with the recommendation algorithm. Show users the perfect video and they'll keep coming back forever. We spent months optimizing our algorithm. We got fairly good at predicting what someone wanted to watch next. And then we looked at the retention data and realized the algorithm was a distant second to something much simpler: whether the user had friends on the platform.

Users who had joined a "Family Guild," our in-app tribe system, were almost impossible to lose. Their retention was so much higher than solo users that our data team ran the analysis twice because it looked like an error. They also generated the majority of our live-streaming revenue, produced more content, and spent more money on the platform.

I remember the meeting where our analytics lead presented the Guild retention numbers. He put the chart up on the projector and the room went quiet. Someone said, "That can't be right." He said, "I ran it three times." We had spent months optimizing our recommendation algorithm, testing different content mixes, A/B testing onboarding flows. It turned out that none of that mattered as much as whether a user had people on the platform who would notice if they stopped showing up.

The three emotions

When we designed the community layer, we worked from a simple question: when does a creator feel like they've found their home? We identified three emotions that had to be present simultaneously.

These became the design principles for Family Guilds. A Family had a leader, usually a popular creator, anywhere from 50 to 500 members, and its own chat room. The structure was borrowed from multiplayer game guilds. In games like World of Warcraft, guilds don't retain players because the game mechanics are perfect. They retain players because leaving means abandoning your friends. The social obligation becomes stronger than the entertainment value. We applied the same principle to a content platform.

Conflict as a retention tool

To bond the tribes together, we added competition. Family Champions League and Family Battles pitted groups against each other for leaderboard dominance. To win, members had to coordinate: pool their Stars, time their gifts during live streams for maximum impact, strategize in group chats about which battles to enter.

This shifted the user's relationship with the platform. Instead of "I'm watching this stream because the content is good," it became "I'm watching because my Family needs me to show up." The majority of our live-streaming revenue came from these battles. Users weren't spending money on entertainment. They were spending money to defend their tribe's honor. Us vs Them.

500 meetups a month in 200 cities

We wanted to strengthen the bonds of Families and we found that organizing offline experiences enhances the online experiences.

Families started organizing offline meetups on their own. At first, a handful per month. By 2022, they had scaled to over 500 meetups a month across more than 200 cities. We supported this with small budgets: enough for tea, snacks, and branded t-shirts. The Family Leaders organized everything else.

When 50 people in a small Indian town show up wearing matching shirts, eat together, and post photos back to the app, you create something no algorithm can replicate. A user might delete an app to save phone storage. They won't delete the app where their actual friends are.

At our second Tiki Star Awards ceremony, a creator from Bihar who had previously had almost no online presence approached me and said: "Tiki is my world." It felt less like a product launch party and more like a family reunion. That moment crystallized what we'd been building. Not an app. A community with a shared identity.

Retention at scale is not a technology problem. It's a sociology problem. Algorithms serve content. They don't create belonging.

Human connections is what brings us closer.

The fourth spoke of the flywheel is Users Become Creators, which is really about where the next wave of creators comes from. For most platforms, the answer is: from the same pool of urban influencers everyone else is fighting over. For us, the answer was the heartland of India.

Trust is built face-to-face

In smaller Indian cities, digital ads don't build trust. App store rankings mean nothing. What matters is whether someone in your social circle uses the product and vouches for it. We learned this after burning money on digital campaigns with terrible conversion rates in exactly these markets.

The offline meetups solved this. Word of mouth spread like wildfire. When your neighbors see a group gathering with branded t-shirts and having a good time, the social proof is more powerful than any marketing campaign. Creators from competitor apps started showing up to our meetups out of curiosity, which generated word-of-mouth we couldn't have purchased.

Purpose beyond entertainment

We positioned Tiki as a "Make in India" platform, aligning with the national self-reliance movement. During the pandemic, we mobilized local creators to spread safety messages in regional languages that mainstream media couldn't reach effectively. The campaign reached hundreds of millions of views. Through a partnership with a major Indian charity, we helped distribute tens of thousands of meals to underprivileged children. Many of our Tiki meetups were self-initiated charity events to give back to their community.

Creators in small towns started telling their audiences, unprompted, that their work on Tiki was helping feed children. They treated it as part of their identity. That emotional connection to the platform was something no competitor could buy away with a bigger paycheck.

If you've read this far, you've absorbed the full operating system. Let me compress it into what I think are the transferable principles, then be honest about what I got wrong.

Why competitors with $900 million couldn't copy this

A reader might reasonably wonder: if these principles are so effective, why didn't the well-funded competitors just do the same thing? The answer is that every one of these moves required sacrificing something that well-funded companies are structurally unwilling to sacrifice.

Creator-first means suppressing viral content that would boost your short-term metrics. Quality mandates mean your dashboards go red for weeks while retention slowly improves underneath. Tribes require sustained operational intensity, funding offline meetups, managing community leaders, running tournaments, none of which shows up in a performance marketing spreadsheet. And going after Tier 2 and 3 cities means building a ground game in places that don't look impressive on an investor slide deck.

Having scaled many startups before Tiki, I'd seen what happens when growth outruns fundamentals. Most teams optimize for speed because their investors reward speed. Very few are willing to sit on their hands for six months studying the market while competitors burn cash. The playbook we built wasn't just insight. It was discipline that well-funded, growth-addicted companies were structurally incapable of executing.

What everyone gets wrong about creator platforms

The real innovation of Tiki was not short video. It was creator economy infrastructure. Most platforms fail because they misunderstand three things.

They think creator supply is an acquisition problem. It's not. It's a status design problem. Paying creators creates mercenaries. A transparent meritocracy with published rules and specific traffic guarantees creates citizens who police their own quality.

They think retention is an algorithm problem. It's not. It's a community design problem. No recommendation engine can replicate the retention power of a guild system where people have social obligations, shared enemies, and offline friendships.

They think content quality is a moderation problem. It's not. It's a cultural problem. You can't enforce quality from a corporate office across dozens of regions and languages. You have to give that power to the community itself and let local users train your systems.

These three solutions form the Creator Platform Playbook. They're not specific to short video. Any marketplace, community, or platform that depends on a supply side of skilled contributors can apply the same architecture.

We didn't invent any of these ideas individually. Status games are as old as civilization. Tribal belonging is anthropology 101. Quality curation is what every good editor has done for a century. What we did was combine them into a single operating system and prove it works at scale, in one of the most competitive markets on earth, against opponents with a hundred times our budget.

When you can't outspend the competition, you have to out-think them. And the single most important thing we figured out was this: platforms win when they stop treating creators as supply to be purchased and start treating them as citizens to be invested in. Once your creators become citizens, the content improves, the users stay, and the flywheel turns on its own. Everything else follows.