Make the best YouTube video
Not the best-produced, funniest or best-looking video. The best YouTube video. Every decision is judged against that.
The handbook opens by rejecting how film and TV work. Most movies and shows would flop on YouTube, with long lead times that can't react to trends. The team aims to make the equivalent of a small movie every week, so it has to stay nimble and work its own way.
Three numbers decide virality
Everyone owns the metrics. Producers, camera and editors are all expected to know the title, the thumbnail and which minute of the video they're working on.
Earn the click
If 100 million people see a thumbnail and 10 million click, that's a 10% click-through rate. The idea itself is chosen for that number.
Pick the idea for the thumbnail
Fifty hours in your front yard and fifty hours in a bathtub of ketchup take about the same effort. Only one makes an image people have to click.
- 01
I Spent 50 Hours In My
Front Yard - 02
I
Spent50 Hours In Ketchup - 03
I Survived 50 Hours In Ketchup
The more extreme, the better
Mild
“I Don't Like Bananas”
Extreme
“Bananas Are The Worst Food On Earth”
The package is a promise. If the title says “World's Largest Bouncy Castle” and the thumbnail shows a yellow one beside skyscrapers, the video must open on exactly that. A red castle in a field feels like a lie, and viewers leave.
Map every minute
On almost every video, the first minute loses the most viewers. On one video, 60 million people clicked and 21 million were gone by 1:00.
0:00–1:00
The first minute
Match the thumbnail right away, light it well, script every word and front-load the most interesting material.
1:00–3:00
Hype to execution
Stop telling people what they'll see and show them. Use crazy progression: cover days of a challenge in minutes so viewers get invested fast.
3:00–6:00
Fall in love
Open with a re-engagement, a spectacle only this channel could pull off. Then the most exciting, simplest content, with quick scene changes.
6:00–end
The back half
A second re-engagement pushes the story on. Longer explanations live here. Never signal the end except to build hype for the payoff.
Get them through half the video and they'll very likely finish. That's why the team always asks which minute mark you're working on.
Watch time wins
Two videos of nearly the same length. People watched one of them 1 minute 38 seconds longer on average. It got almost three times the views.
YouTube shows the exact second every viewer leaves, so the retention graph is the honest review of the content. Hook at the start, move into a story people care about, cut every dull moment, then pay it off and end abruptly.
Formats that hold attention
A good format does part of the retention work for you. The best ones hide the answer until the end.
Last to leave
Who stays in the circle longest wins $100,000. The winner isn't known until the end, so people stay.
Stair-stepping
Start small and escalate: $1, $10, $50 and up to the record. You can wander as long as each step gets cooler.
The chase
Bounty hunters, the military or the FBI hunt Jimmy. Will they catch him? You have to watch to find out.
Retire formats before viewers ask you to. People said they wanted more donating-to-streamers videos, but every channel that rehashes formats for years dies. Space out videos in the same format.
The wow factor
You can't track it, but you can define it: anything no other creator can do.
The easy way
The contestant's house is already standing in the circle when the video starts.
The wow way
Thirty seconds in, the house arrives on a crane.
The crane added nothing to the title or thumbnail, and on paper it was a waste of money. But it hooked viewers and changed how they saw the channel: one more thing nobody else on YouTube could do.
- Spend over $10k? Show it on camera
- Spectacles only you can make
- Fold mishaps into the story
- Never fake intensity
Brand deals are content
A read-from-the-script ad leaves a crater in the retention graph. An ad built into the content barely dents it.
Creativity saves money
Money is rarely the answer. A funnier idea usually costs less.
The same rule applies to cranes, locations and props. Every dollar saved buys more people and more stability for the team.
How the videos get made
Most of the handbook is about not leaving room for error. These are the habits that carry over to any team.
- Protect critical components. Anything the video can't exist without (the island, the thumbnail's promise, the two challenges the boss loved). Check on it constantly, build a backup and never leave it to chance.
- Use consultants first. Need the world's largest slice of cake? Call whoever made the last one. They've already run the tests.
- Video everything. Film every set scout and every order so the whole team shares one picture, not just the person who was there.
- Say the negatives. “I found the castle” is useless without “but it's booked all year and over budget.”
- Name the bottleneck and check in daily. Set a date with whoever blocks you, then check in every day, not just on the deadline.
- Bring options, not questions. Give the decision-maker context, the budget and five pre-approved choices.
- Push through no. One “no” from one person is a starting point. Stop only when every avenue is exhausted.
- Book a backup day. Make sure expensive sets and rentals can be extended.
- Feed your information diet. You can't be inspired by things you don't know exist. Watch YouTube, follow memes and keep a mental library.
The short-form version
The handbook is written for 13-minute videos. Here's how we compress it for TikTok, Shorts and Reels.
| MrBeast long-form | Short-form translation |
|---|---|
| Click-through rate | The first frame and on-screen title are the thumbnail. Make that frame the extreme image. |
| The first minute | The first 1–3 seconds: open on the exact promise, already in motion. |
| Crazy progression (1:00–3:00) | Skip the setup. Show three escalations in the first third. |
| Re-engagement at 3:00 and 6:00 | A twist or a bigger stake at about 40% and 70% of the runtime. |
| Stair-stepping | $1 vs $100 vs $10,000, or level 1 to 10, with the biggest last. |
| Payoff, then abrupt ending | Cut on the reveal. Ending without an outro lets the video loop. |
| Wow factor | One shot only your setup could get: your game, your code, your access. |
Pair this with Playbook 02 to pick the format, and Playbook 01 for platform rules.
Before you film
- 01Package first
Write the title and sketch the thumbnail before anything else. Is it extreme and impossible to scroll past?
- 02Match the promise
Does the opening show exactly what the package promised?
- 03Front-load
Is the best material in the first minute, or the first three seconds?
- 04Plan re-engagements
What's the spectacle at each turning point?
- 05Hold the payoff
Is the answer saved for the end, with an abrupt cut after it?
- 06Find the wow
What's the one thing no other creator could do here?