# I've generated 18 billion views. Here's how to go viral every time.
**作者**: Reece | Clipping Agency
**日期**: 2026-07-23T08:47:48.000Z
**来源**: [https://x.com/rhysclipping/status/2080213276399399196](https://x.com/rhysclipping/status/2080213276399399196)
---
I've generated 18 billion views.
Here's how to go viral.
most people in this space think they know how virality works, but the problem is most of them are wrong.
i've generated over 18 billion views across 60+ industries.
my agency puts out up to 750 million views every single month. and i can tell you with certainty virality is not random, it's engineered.
what most people get wrong
most brands think virality is about content quality.
write better scripts, and hire better editors, or post more often, find the right niche.
it's true but not how people make it out to be.
obviously your content isn't going to perform if it's straight up trash so it needs to actually be good.
but for a lot of them it's not the actual problem.
the real problem is the distribution:
for example, a 10,000-view clip relies on the brand's existing audience.
it reaches the people who already follow them and not many beyond that.
a 10,000,000-view clip relies on human psychology and algorithmic triggers and it doesn't care whose account it's posted from.
it engineers a response.
the algorithm rewards watch behaviour.
specifically, it watches three signals:
does the viewer start watching
do they keep watching
do they watch again.
quality has a small effect on those three signals directly.
what drives them is mechanical.
here's what the data actually shows.
## the 3 mechanics that separate 10k views from 10M views
the competence gap in the hook
a 10,000-view hook says: "here are 3 tips to grow your instagram."
a 10,000,000-view hook does something different.
it attacks a specific mistake the viewer is already making, or destroys something they believed was true.
it creates an immediate need in the viewer's brain to know the answer before they can move on.
"you've been posting reels wrong and here's the proof" compared to "here's how to post better reels."
the first one creates a competence gap.
the viewer thinks wait, am i the one doing this wrong?
that question makes scrolling past psychologically difficult.
the hook's only job is to create a gap that can only be closed by watching the rest of the clip. not to summarize. not to explain. to open a gap.
information density and visual pacing
a 10M-view video has zero dead air. not reduced dead air. zero.
every 3 to 5 seconds, there's a visual reset
the other half of this is progressive disclosure. instead of showing a completed diagram, you build it live on screen.
instead of stating a list, you reveal it item by item. the viewer always feels like the next piece of value is 3 seconds away. that feeling is what kills scroll impulse.
the perfect loop
this is the one almost nobody talks about, and it might be the most important of the three.
the algorithm's primary distribution metric is average watch time as a percentage of video length.
a 100% watch time means the viewer watched the whole thing.
above 100% means they watched it more than once.
the algorithm interprets that as "this video is so good people can't stop watching" and pushes it to cold audiences at scale.
the final sentence of the clip has to bleed back into the first sentence of the hook.
the viewer finishes the video and the brain doesn't register a stopping point.
it loops before the person consciously decides to. the watch time counter keeps climbing.
## the four frameworks we actually use
every clip we produce is built on one of four internal frameworks depending on the goal. clippers don't improvise, they work off the brief.
> the authority myth-buster. opens with a contrarian hook, drops an authority anchor immediately, delivers a 3-part solution, ends with an ideological cta. used to build trust fast with new audiences.
> the delta shift. shows what the average person does first, then introduces what the top 1% does instead, then gives one immediately actionable takeaway. drives saves more consistently than almost any other format.
> the progressive debunking (the iceberg). introduces a polarizing topic, dismisses the surface-level tactics everyone obsesses over, and progressively reveals the deeper mechanics underneath. maximises watch time because the viewer keeps expecting the next revelation.
> the HYPE/HEIT framework. promises a named system with the key asset deliberately blurred on screen, rapidly breaks down the parts, and ends with a keyword cta. pure lead generation.
the frameworks are the skeleton.
on top of that, we're constantly analyzing what's exploding in real time across every niche, reverse-engineering the mechanism behind it, and rebuilding it as a new format for our clients.
by the time something looks like a trend, we've already moved past the format that created it.
## the distribution layer
the clip is only half the machine.
we run a single brand's content across 62,900+ vetted clippers and creators at full scale.
tiktok.
instagram reels.
youtube shorts.
x.
thousands of accounts that all feed attention back to the main brand.
adobe: 14 million views from 867 submissions.
okx: 248 million views in 30 days.
stake: 742 million views delivered.
every campaign has its own dashboard, and every view is anti-bot verified before it counts.
the numbers we generate come from combining a high-volume distribution machine with content that's actually built to travel.
these are the basics of scaling organically and if there's one thing i want you to take away it's:
EVERYTHING IS IN THE DATA
## 相关链接
- [Reece | Clipping Agency](https://x.com/rhysclipping)
- [@rhysclipping](https://x.com/rhysclipping)
- [3.8K](https://x.com/rhysclipping/status/2080213276399399196/analytics)
- [Upgrade to Premium](https://x.com/i/premium_sign_up)
- [4:47 PM · Jul 23, 2026](https://x.com/rhysclipping/status/2080213276399399196)
- [3,825 Views](https://x.com/rhysclipping/status/2080213276399399196/analytics)
- [View quotes](https://x.com/rhysclipping/status/2080213276399399196/quotes)
---
*导出时间: 2026/7/24 10:22:46*
---
## 中文翻译
我创造了 180 亿次浏览量。以下是如何每次都实现病毒式传播。
**作者**: Reece | Clipping Agency
**日期**: 2026-07-23T08:47:48.000Z
**来源**: [https://x.com/rhysclipping/status/2080213276399399196](https://x.com/rhysclipping/status/2080213276399399196)
---
我创造了 180 亿次浏览量。
以下是如何实现病毒式传播。
在这个领域，大多数人都认为自己知道病毒式传播是如何运作的，但问题在于他们大多数人都错了。
我在 60 多个行业创造了超过 180 亿次浏览量。
我的代理公司每个月能产生高达 7.5 亿次的浏览量。我可以肯定地告诉你，病毒式传播不是随机的，它是精心设计的。
大多数人搞错的地方
大多数品牌认为病毒式传播取决于内容质量。
写出更好的脚本，聘请更好的剪辑师，或者更频繁地发布，找到正确的细分领域。
这话没错，但并不是人们理解的那样。
显然，如果你的内容完全是垃圾，它就不会有好的表现，所以它确实需要好。
但对很多人来说，这并不是真正的问题。
真正的问题在于分发：
例如，一个 1 万次浏览量的片段依赖于品牌现有的受众。
它触达那些已经关注他们的人，除此之外影响甚微。
一个 1000 万次浏览量的片段则依赖于人类心理学和算法触发机制，它不在乎是从哪个账号发布的。
它诱导了一种反应。
算法奖励观看行为。
具体来说，它关注三个信号：
观众是否开始观看
他们是否持续观看
他们是否再次观看。
质量对这三个信号有轻微的直接影响。
驱动这些信号的是机制。
这就是数据实际显示的内容。
## 区分 1 万和 1000 万浏览量的 3 个机制
钩子中的能力差距
一个 1 万次浏览量的钩子会说：“这里有 3 个涨粉技巧。”
一个 1000 万次浏览量的钩子做法不同。
它攻击观众正在犯的一个具体错误，或者摧毁他们原本认为是正确的事情。
它在观众的大脑中制造了一种立即知道答案的需求，否则无法继续。
“你发 Reels 的方式一直是错的，这是证据”对比“这是如何发更好的 Reels”。
第一个创造了一个能力差距。
观众会想：等等，我就是那个做错的人吗？
这个问题使得划走变得心理上很困难。
钩子的唯一任务是创造一个只能通过观看片段其余部分来填补的缺口。不是为了总结。不是为了解释。而是为了制造缺口。
信息密度和视觉节奏
一个 1000 万浏览量的视频没有“死气时间”。不是减少，而是零。
每 3 到 5 秒，就有一次视觉重置
另一半是渐进式披露。不要展示一张完成的图表，而是在屏幕上实时构建它。
不要陈述一个列表，而是一个接一个地揭示。观众总是感觉下一个价值点就在 3 秒之后。这种感觉扼杀了划走的冲动。
完美循环
这是几乎没人谈论的一点，它可能是这三个中最重要的。
算法的主要分发指标是平均观看时间占视频时长的百分比。
100% 的观看时间意味着观众观看了整个视频。
超过 100% 意味着他们观看了不止一次。
算法将其解释为“这个视频太好了，人们停不下来”，并大规模将其推送给冷流量受众。
片段的最后一句必须无缝衔接到钩子的第一句。
观众看完视频，大脑没有注册到一个停止点。
它在人有意识地决定之前就开始循环。观看时间计数器不断攀升。
## 我们实际使用的四个框架
根据目标，我们制作的每个片段都基于四个内部框架之一。剪辑师不即兴发挥，他们按照简报工作。
> 权威迷思粉碎机。以反直觉的钩子开场，立即放下权威锚点，提供 3 部分解决方案，以意识形态性的 CTA（行动号召）结束。用于与新受众快速建立信任。
> 差距转变。首先展示普通人的做法，然后介绍前 1% 的人的做法，然后给出一个立即可执行的要点。与其他任何格式相比，它能更持续地驱动收藏。
> 渐进式揭穿（冰山）。引入一个两极分化的话题，摒弃人们痴迷的表面策略，并逐步揭示其下的深层机制。最大化观看时间，因为观众总是期待下一个启示。
> HYPE/HEIT 框架。承诺一个命名系统，关键资产在屏幕上故意模糊，快速分解各个部分，以关键词 CTA 结束。纯粹的潜在客户开发。
框架是骨架。
在此基础上，我们要不断分析每个细分领域实时爆火的内容，反向拆解其背后的机制，并将其重建为客户的一种新格式。
当某样东西看起来像一种趋势时，我们已经超越了创造它的那种格式。
## 分发层
片段只是机器的一半。
我们大规模地在 62,900 多个经过审核的剪辑师和创作者账号上运营单一品牌的内容。
TikTok。
Instagram Reels。
YouTube Shorts。
X。
成千上万个账号将注意力回馈给主品牌。
Adobe：867 个投稿带来 1400 万次浏览。
OKX：30 天内 2.48 亿次浏览。
Stake：交付了 7.42 亿次浏览。
每个活动都有自己的仪表板，每次浏览在计入前都经过反机器人验证。
我们产生的数字来自于将大量分发机器与真正适合传播的内容结合起来。
这些是自然增长的基础，如果你只能从中学到一点，那就是：
一切都在数据中
## 相关链接
- [Reece | Clipping Agency](https://x.com/rhysclipping)
- [@rhysclipping](https://x.com/rhysclipping)
- [3.8K](https://x.com/rhysclipping/status/2080213276399399196/analytics)
- [Upgrade to Premium](https://x.com/i/premium_sign_up)
- [4:47 PM · Jul 23, 2026](https://x.com/rhysclipping/status/2080213276399399196)
- [3,825 Views](https://x.com/rhysclipping/status/2080213276399399196/analytics)
- [View quotes](https://x.com/rhysclipping/status/2080213276399399196/quotes)
---
*导出时间: 2026/7/24 10:22:46*