A creator platform is only as strong as the behavior it rewards when people start believing in it.
From the outside, it looked like a short-video app. Internally, the product was status, coaching, quality, community, and trust.
Tiki Short Video was one of the most intense products I have built. We launched in India in February 2021. Within a month, we crossed 1 million DAU. By May, we crossed 5 million DAU. By December, we crossed 10 million DAU.
Tiki internal BP materials put the platform at more than 32 million monthly active users and a Tiki community of more than 100 million. Those numbers still sound unreal when I read them back.
At the same time, Tiki eventually shut down in 2023. I am proud of the run, but the useful lesson is in the parts I would rebuild.
We were trying to build a place where unknown local creators could become somebody. A creator from a tier 2 or tier 3 city could earn attention, receive support, belong to a group, and eventually make money from talent.
The line we kept coming back to was simple: a place for real talents.
I still believe that instinct was right. The creator economy has changed, though. AI has made content production cheap. Distribution is more fragmented. Creators now live across short video, live, commerce, messaging groups, offline events, private communities, and brand deals.
The platform that wins now will be the one that creates the most trusted creators. That is the part I would rebuild around.
The 2026 version should stop disguising it. It would still have a feed, live rooms, creators, fans, guilds, rankings, and rewards. The real product would be the full stack that helps unknown creators in underserved markets become trusted local media entrepreneurs.
Recruit for hunger, not only follower count. The best supply often grows with the platform.
Protect taste before the feed trains everyone into the same format.
Make creators understand where they stand and what they need to improve.
Turn creators, fans, guilds, and cities into a world people want to defend.
Monetization works best when it feels like support, not extraction.
This essay is the Tiki chapter inside the broader Distribution Is Part of Product framework.
oBike taught me access. OYO taught me supply. Tiki taught me status: the product has to decide what kind of contributor the platform wants to create.
Gold, Blue, Grey, White, UGC. The labels changed over time, but the lesson stayed: traffic carried salary, status, proof, and hope in one object.
Top creators got more public traffic and higher status. That worked when the platform could point to them as proof of taste, craft, and aspiration.
Better version: make top status visible, but never make it feel permanent.
When we gave a creator traffic, we were telling them the platform believed they could represent us. When we removed traffic, we were telling them their status could be lost. That made Tiki feel alive, but also dangerous. Every ranking rule shaped behavior.
Our creator survey made this clearer than I remembered. In one survey of 289 Blue and Gold V creators, "I like to create" ranked ahead of money as a reason they joined. When asked what they wanted from Tiki, traffic came first, then money, then happiness and friendship.
Creators care about money. Many of them first want proof that the platform can make them matter.
High-profile L1 influencers averaged around 36 million views with a 3.12% like rate in one internal comparison. They helped cold start, but they did not automatically create culture.
The strongest Tiki-native creators averaged around 232 million views and a 5.16% like rate. Users could feel the difference between fame and effort.
The early metrics were strong because the feed had taste. The algorithm mattered, but the operating discipline mattered more.
Tiki's active retention was listed at D1 70%, D7 57%, and D30 40% in old decks. Another India update showed DAU growing from 3 million to 7.5 million while D1 retention stayed around 70.2%.
Video views per user increased. Playtime per user increased. Organic new users moved from around 60,000 a day to 100,000 a day.
We also saw that Tiki's average video views were 40% higher, average playtime was 20% longer, and share rate was 2x compared with Likee Global in the materials.
The easy explanation is algorithm. I think the better explanation is discipline.
We cared whether the video had a reason to exist. Internally, we kept repeating content values: spread happiness, share knowledge, tell a story, showcase rhythm. Those words gave the team a taste language.
Short-video teams get bullied by dashboards. A cheap clip can spike views and still train the platform in the wrong direction.
The more I reread the old files, the more obvious this becomes. Tiki was growing through identity.
The Family Guild MVP is the best example. Creators formed teams, earned points, received stars, completed missions, and fought for rank. Nearly everyone used their team avatar pendant. That last detail matters. People change their avatar when they want others to know who they are with.
Offline made the same point. We held creator meetups across states and cities. Some were official. Some became creator-led. I remember Mumbai, Indore, Assam, and other cities showing that the product had entered real life.
One internal note counted 32 offline meetups across 8 states and 14 cities with 350 creators. Internal community materials described thousands of creator and charity meetups across 100+ cities with hundreds of organizers.
Competitors can copy coins, badges, creator funds, and leaderboards. Copying a feeling is harder.
India taught me scale. MENA taught me that "local" is smaller than country, smaller than language, and much harder to fake.
The India playbook broke in MENA. That is the part I would bring forward, because it is the part most platform teams underestimate.
In MENA, we also saw users who wanted to become stars. In one survey, most users said they came to Tiki because they wanted to be a star. The supply side was more complicated. We recruited many creators from Egypt because Egypt had a deeper entertainment market. Then we saw that GCC users often wanted different creator supply.
Same language, different market.
Saudi Arabia, Egypt, Morocco, UAE, and Iraq can break if treated as one creator culture because Arabic overlaps. Humor, gender norms, aspiration, spending power, music, family expectations, religious constraints, and what counts as high quality all move by market.
The old MENA update captured an early operating phase, not the final ceiling: DAU around 30K+, retention around 35%+, and the team still working through cold local supply, leaderboard competition, Future Stars, and Blue/Grey/White creator performance. India gave the scale receipt. MENA gave the sharper lesson: you cannot average a region into one creator market.
Chinese consumer platforms are strong at feed iteration, creator incentives, live monetization, traffic allocation, internal tooling, and operating cadence. Likee, Kuaishou, ByteDance, and BIGO all understand that creator platforms are systems, not pages.
The next layer is local translation at the level of supply, norms, trust holders, and incentives. A product can be technically strong and still need market-specific proof before creators and agencies risk their reputation on it.
For a Chinese consumer or AI platform entering MENA, I would start with one operating question: which local trust map makes this product believable enough for creators, agencies, fans, merchants, and regulators to risk their own reputation on it?
That is where product, distribution, local operators, and creator economics become one system.
Status attracts ambition. It also attracts gaming. The more valuable the ladder becomes, the more carefully it has to be governed.
Once stars, traffic, badges, and ranking had economic value, some creators optimized for the reward instead of the craft. Internal notes mentioned cringe, sympathy, and shocking content being used to get views and stars.
Other notes raised concerns about semi-explicit content, low-quality uploads, and creators chasing views without improving. That is where the platform has to be honest with itself.
If you pay for views, you manufacture views. If you reward originality, the team needs a way to define and defend originality.
Opening paths for Grey and White creators was necessary. It also made the feed uneven when direction and review tools lagged behind scale.
Hashtags gave creators direction, then sometimes made everyone produce the same dance, joke, or format.
Delayed or unclear creator payments damage trust fast. Creator platforms need strong rails earlier than teams want to build them.
Offline campaigns, merchandise, creator spotlights, and social pushes needed cleaner tracking from day one.
Country and language are too broad. A creator market has its own trust holders, taste map, and social limits.
The warning signs usually appeared before the dashboard moved: a creator asking about payment, a leaderboard that felt political, a campaign with no attribution, a market treated too broadly. The rebuild lesson is to give those warnings an owner before they turn into culture.
I would still build around real local creators. The rebuild would start from different first principles.
Every creator should know where they stand, what they need to improve, what they can earn next, and why someone else is ahead.
Promotion and demotion should feel fair before they feel exciting. Creator trust is built when the rules are clear, the feedback is specific, and the path is hard but believable.
In 2021, training creators manually was expensive. In 2026, the coaching layer should sit inside the app.
A creator uploads a draft. The system gives feedback: the first three seconds are unclear, the story has no payoff, the audio is weak, the content is drifting from the creator's category, or the format is too close to a recycled trend.
Traffic is the most powerful incentive in a creator platform. I would treat it like a currency with rules, audits, and explanations.
Creators can disagree with a decision and still trust the system if the rules feel real.
Every creator should have a place to belong. Every fan should be able to support a group. Every market should have local leaders.
Trust often travels through people before it travels through software. A meetup, a local leader, a shared shirt, a city cleanup, or a small award can create more belief than another feed tweak.
Creator CRM. Payment rails. Referral attribution. Campaign tracking. Music rights workflows. Creator manager dashboards. Content quality review tools. Leaderboard audit logs. Creator support tickets. Local event verification. Safety escalation.
These sound like back office. In a creator platform, they are trust infrastructure.
I used to think the hardest part of Tiki was growth. Growth was hard. The bigger challenge came after growth started working.
The hardest part was designing a system that protected the reason people joined after growth started working.
If I rebuilt Tiki today, I would spend less time asking how to get more creators and more time asking what the system teaches creators to become.
If you are a Chinese consumer or AI platform expanding into MENA, this is the operating call I run: map the local trust holders, creator supply, status incentives, and distribution loops before the launch plan hardens.
Teach creators to chase shortcuts and the platform becomes cheap.
Teach creators to improve, support each other, build local trust, and earn status through real contribution, and the platform can become something much bigger than entertainment.
Tiki proved that unknown creators in hard markets want a stage. It also proved that a stage needs rules, taste, governance, community, and trust.
The 2026 version would still try to create stars. I would be much more careful about what the product rewards on the way there.