YouTube Viral Video Methodology: From Shorts Ideas and Script Adaptation to AI Video Workflows
My first YouTube video reached 3 million views. I did not write its script from scratch; I based it on a video that had already gone viral.
My first YouTube video reached 3 million views. I did not write its script from scratch; I based it on a video that had already gone viral.
This experience made me realize how important it is to find worthwhile references before making a video. But I have also been stuck at around 10,000 views. Seeing a viral video and understanding why it went viral are two different things. The format, script, characters, and timing of your follow-up all affect the final result.
So I want to connect these experiences and explain them clearly: how beginners can choose a content format, how to analyze and adapt a viral video after finding it, how to build a script library, and how to increase output through a workflow. Along with Shorts, I will include examples of long-form videos and livestreams so you can see how production differs across directions.
When it comes to finding references, I will introduce how I use ShortsMonkey. It can help us discover Shorts and long-form videos on YouTube with low subscriber counts and high view counts, browse recent viral videos in one place, search keywords, and research channels. The time saved on finding examples can go into evaluating topics and refining scripts.
The figures and screenshots below are the records I saved when collecting these examples. They were not all observed at exactly the same time.
1. Two Types of Viral Shorts: Format Appeal and Storytelling
1. Format-Driven Hits: Let Viewers Grasp the Hook Immediately
I call the first type a “format-driven hit.” Many of the videos I collected in this category are only 5–10 seconds long: the visuals are striking or novel, or a single action makes people want to see the result. They do not necessarily need to tell a complete story.
For example, Haaland kicks a soccer ball toward a stadium from high above, which requires almost no explanation of his background. A blue watermelon ice treat being displayed and bitten focuses attention on its color, texture, and the action. Viewers quickly understand what the video is going to show.
Watch the example: Haaland’s high-altitude shot, about 10 seconds
Watch the example: watermelon ice ASMR, about 10 seconds
I looked at several videos from this channel together. With the same Haaland and a similar long-range shooting premise moved into different settings, the videos reached about 5.94 million, 3.75 million, 15.42 million, and 1.42 million views. Of course, some videos performed much worse. When I study references, I compare both the high- and low-performing videos in a series to judge how much room the format still has for variation.
For the channels and videos below, I suggest breaking them down as “first-glance hook → action process → presentation of the result”:
| Reference | What I focus on |
|---|---|
| Paws-andVibes | Format-driven hits; first test whether viewers can quickly understand the visuals and premise |
| Reference video 1, reference video 2 | I compare these two format examples side by side, looking at the opening, action, and result |
| Staircasesurprises | A reference for format-driven hits |
| IceVibes-i8c | AI ASMR, Rumi, and watermelon-eating themes; observe the sensory experience |
| commanderbed | Soldiers and heartwarming family skits |
| Product-promotion format reference | I mainly look at how it integrates a product demonstration into the content; sales performance needs separate verification |
| Pupshade | The same template was posted dozens of times in a row, so I can study how it becomes a series |
| Split-screen video of a real person recreating AI | Juxtaposition, recreation, and contrast between a real person and AI |
| maddarkkart | These Haaland formats remind me of earlier Messi and Cristiano Ronaldo videos: the people change, while some of the concepts continue |
| Toothpaste-squeezing example | How an everyday action creates an immediate reason to watch |
When researching this kind of content, first summarize its format in one sentence. If you still cannot explain why viewers would watch after replacing the character, you have not yet found the part truly worth reusing.
2. Story-Driven Hits: Give Viewers a Reason to Want the Ending
I call the second type a “story-driven hit.” It first helps viewers understand what a character wants and what difficulty they face, then moves the story toward its ending through actions and turns. Viewers keep watching because they care about what the character will do next.
I selected three clips for you to compare:
- The phone-wish story: centers on a child who wants a phone, linking the wish, the process of obtaining it, and the characters’ reactions.
- The swallowed-ring story: uses a proposal, an accidental swallow, and the recovery of the ring to create an exaggerated comic plot.
- The donuts-and-sharing story: contrasts the situations of people inside and outside a car in relation to food, making viewers care about what happens next.
When I saved the screenshot, the donut example had already reached about 210 million views. What I especially want you to notice is the relationship between the characters: who has the food, who is waiting, and what the characters do next. Break these beats apart and you can see how the emotion builds step by step.
Format and story can also work together: first use powerful visuals to make viewers stop, then use a clear short story to keep them watching.
2. Five Approaches to Creating Viral Videos: From Fast Learning to Evidence-Based Adaptation
Method 1: Learn from Viral Videos, and Follow Up Quickly
My earliest method, to put it bluntly, was to “copy a viral video directly.” I made no changes to the entire script of my first video and recreated it one for one. It ultimately reached 3 million views.
Looking back, I think an important reason it took off was that I followed up quickly. Competition is different when a concept has just started to explode than after it has already been repeated many times.
That raises a practical question: How can I find out sooner which of someone else’s videos has gone viral?
I divide the search into two methods. The first is to scroll through videos and accounts yourself, looking one by one. This takes a lot of time; sometimes, after several hours, I find only two or three examples worth studying. The second is to use a tool such as ShortsMonkey to browse recent viral videos in one place, then open the original videos and channels for closer research.
This is not just my experience. When my views were stuck at around 10,000, I asked JUGO, a fellow community member who had shared during the livestream at our previous offline meetup in Guangzhou. He has made videos with hundreds of millions of views, and he emphasized the same idea. I asked whether I should try to finish an adaptation within a week after the original video takes off, because waiting too long might produce a different result. His answer was direct: timing matters, and you need to move fast.
I do not treat “one week” as a fixed formula, but that conversation made me pay more attention to the interval between discovery and production. Once I find a suitable example, I need to break it down quickly and make my own version so I can get feedback in time.
However, I do not treat my first recreation as the whole method for running a channel over the long term. YouTube requires original value. Template-based content with very little variation and reused content without substantive changes may be ineligible for monetization; permission to repost does not necessarily mean the content meets monetization requirements. YouTube channel monetization policies
So I recommend going one step further: understand why it is engaging, then write your own plot and expression. That brings me to the second method.
Method 2: Micro-Innovation—Capture the Story’s Core, Then Change the Characters, Goal, and Process
For me, the first step in micro-innovation is to identify the most compelling core of the original video. The characters, props, and visual style can change, but I first need to know what I intend to keep and what I intend to change.
For example, around the premise of “entering a body to complete a mission,” I can write three different story chains:
- Recovering a ring: A proposes → B accidentally swallows the ring → A enters the body → A finds the ring → A successfully proposes.
- Breaking in and escaping: A steals something → A is eaten by B → A enters the body → A causes havoc inside → B expels A.
- Helping and healing: A and B enter C’s body → A and B deal with the bacteria → C recovers → A and B come out.
All three stories share the premise of “entering a space that is normally invisible,” but their character motivations, actions, and endings differ. The core worth studying is the change of space and the advancement of a mission. The new script needs to rewrite the causal chain clearly.
How I Choose What to Adapt at Different Stages of Competition
When selecting references, I also consider how far the niche has developed. My priorities for researching examples differ at the early, middle, and late stages:
| Niche stage | Supply and demand | What I look at first | Priority of reference types |
|---|---|---|---|
| Early | Blue ocean; supply is lower than demand | My own viral videos and recent viral videos first | Viral videos with the same IP → viral videos with a different IP → viral live-action videos |
| Middle | Normal market; supply and demand are balanced | My own viral videos and recent viral videos first | Viral videos with a different IP → viral videos with the same IP → viral live-action videos |
| Late | Red ocean; supply exceeds demand | My own viral videos and recent viral videos first | Viral videos with a different IP → viral live-action videos → viral videos with the same IP |
As for the scale of adaptation, my experience is that in a blue-ocean stage I can begin with smaller changes around the core structure. The more recent the viral breakout, the more worthwhile it is to validate quickly. When using a viral live-action video as a reference, I can first preserve the clear story relationships and then change how they are presented.
As similar content becomes more common, I broaden my references and look for opportunities across characters, topics, and live-action-to-animation conversions. “Changing less” here means first concentrating on validating key changes; the final work still needs its own creative content.
Pet Stories: What Can Change Beyond the Characters?
Consider this group of pet stories: an adult walking a turtle, an adult walking a crocodile, an anthropomorphic cat with a dragon, and an adult riding a dinosaur with a pet. Looking at these images, I considered several directions: Could the “treasure egg” story have an adult version? Would replacing the character with an adult work? Does the dragon element attract traffic? Could the adult-and-pet combination avoid plots that put children in danger?
You can view the four examples separately: adult and turtle, adult and crocodile, anthropomorphic cat and dragon, and adult and pet riding a dinosaur.
But I recommend using these examples only to learn adaptation ideas. All of these videos were later taken down, and what I learned at the time was that the reason involved harmful content. So I cannot simply assume that replacing a child with an adult solves the problem; I still need to review the entire plot.
The Same Wish Structure Can Transfer Across Characters and Visual Styles
Here is an easier example. A girl wants something, but a boy cannot afford it, so he enters a competition, wins money, and buys it for her. When I watched this animation, I thought of the earlier story about a child and their mother. The characters, object, and visual style changed, but the structure of desire and effort remained.
Watch the example: an animated adaptation involving a wish, a competition, and a gift
I can record this structure as: A wish appears → resources are insufficient → the protagonist takes action → resources are obtained → the wish is fulfilled. During research, I can also pay attention to the order in which information is revealed. Showing the gift first and then revealing the protagonist’s effort to obtain it can create a different emotional effect.
Method 3: Combine Viral Openings with Viral Elements
The third method is to break a viral video into smaller parts. I look separately at its opening, characters, props, conflict, and ending, then consider which parts can be combined into a new story.
For example, I collected several types of openings:
- Pouring liquid: makes viewers immediately wait for what will change next.
- Eating diamonds: creates suspense through an action that is unusual in real life.
- Shipwreck: quickly establishes the characters’ situation and a survival mission.
- Either-or choice: uses a choice to create anticipation, with the FunnyStoryDis channel as a reference.
- Judgment Day: establishes conflict through a setting that implies judgment.
I also record recurring elements separately: drinking a potion, a slide, shaving hair, soccer figures such as Cristiano Ronaldo, an intestinal perspective, inflating a large belly, smashing glass at the opening, and entering a body. Whenever I record an element, I keep asking: What role does it play in this video? Does it create suspense, advance the plot, or deliver a reversal?
Breaking it down further, I focus on the emotion it gives viewers: joy, anger, anxiety, nostalgia, sadness, fear, and surprise can all be reasons to keep watching.
Another angle is human desire, curiosity, and value judgments. I include motivations often called “human weaknesses,” such as laziness, voyeurism, greed, lust, arrogance, jealousy, and vanity, together with justice, sympathy, gratitude, wishing, and curiosity in my analysis checklist. Which one I use depends on what the story is about.
When applying this checklist to creation, I can ask three specific questions: What emotion does the opening evoke? Who does the middle make viewers care about, and what are they waiting for? How does the ending answer that expectation?
For example, an “either-or choice” can establish a decision, “insufficient resources” can create an obstacle, and “helping family through hard work” can provide the emotional payoff. The story becomes coherent only when every element has a clear function.
Method 4: Build Around Goodness, Helping Others, and Family Bonds
I summarize the fourth method in one sentence: videos about doing good deeds and helping others. Selfless sacrifice between family members, in particular, makes it easy for viewers to understand why a character is willing to give.
These stories usually have an easy-to-understand core: one person needs help, another takes action, and the character’s situation changes as a result. Viewers can quickly understand why the character acts this way and become emotionally invested in the outcome.
I study the following two videos together:
- An AI story about a golden beard and helping others: the visuals progress from a character undergoing a special transformation to scenes of helping others.
- A girl protects puppies: an example clip from “A 10,000-Character Breakdown of the YouTube AI Video Adaptation Workflow—Practical Experience Earning $3,000 in Two Months”: about 15 seconds long, it completes a short story by showing puppies in danger, a girl taking action, and her ultimately staying with the puppies.
I also collected three examples that can be broken down further:
| Channel in the example screenshot | Views shown in the screenshot | Visual cues |
|---|---|---|
| ASAD 3lo | 30.46 million | Interaction between people inside a car |
| Siuuu | 57.88 million | Soccer figures, a simple birthday cake, and a surprise |
| Powerful - Asher | 150 million | A father, a child, and an AI story about helping others |
What I think is most worth learning from these videos is how they turn kindness into a visible action: sharing food, preparing a gift, caring for family, or helping someone in difficulty. Only when emotion has a concrete vehicle do viewers develop feelings for the characters. If a story only repeats the same suffering and rescue, it quickly becomes thin.
Method 5: Research Indian Audiences and Localized Topics
The fifth direction is the Indian market. My approach is to adapt proven stories with characters, language, and settings familiar to the target audience, then observe the feedback.
I have seen videos with rough visuals and relatively crude production still achieve very high view counts. So I consider whether the earlier stories might also have an opportunity if they use character IP familiar to local audiences. Of course, I do not just replace a face; I also examine whether the character relationships, props, and expression make sense to the target audience.
The image above shows an example related to a golden-goat story. The screenshot shows that the Aas AI World video had about 21.5 million views. Another image presents a series format in which “an old man makes wooden statues of celebrities”: the production process stays fixed while the person being carved changes.
The wooden-statue series screenshots feature subjects such as Goldberg, Brock Lesnar, Michael Jackson, Rohit Sharma, and Tamannaah Bhatia. The highlighted samples had about 43.62 million, 3 million, and 5.7 million views, respectively, while some works had lower view counts. The series offers research clues about how the same production format can be adapted to different people.
You can also look at this India-themed Short together with the ShakilDiaryOfficial channel.
But I have one clear warning about this direction: In my experience, RPM is lower in the Indian market. If you have other suitable options, I do not recommend targeting India solely for high view counts. RPM measures creator revenue per thousand views, and the exact figure still depends on the channel and audience. When choosing a market, I consider view potential, production cost, and revenue potential together.
3. Fifteen Long-Form Video Examples: Finding Research Subjects Among High-View Videos from Small Channels
In addition to Shorts, I compiled 15 long-form video examples worth studying. They cover animated stories, life on a boat, outdoor survival, painting, gaming, English learning, travel, and interviews.
When screening this kind of video, I first look at the views on an individual video, the channel’s subscriber count, and the ratio of views to subscribers. When a small channel gets views far beyond its own size, I want to keep investigating: Is the topic strong, is the title strong, or is there genuine audience demand for the subject?
Below is my complete table of examples. Both the accounts and representative videos can be opened directly. “Subscribers” in the table means the channel’s subscriber count.
These videos were published in July and August 2026. The table contains data I saved when collecting the examples, not measurements taken at the same moment. View and subscriber counts displayed in units of 10,000 were rounded, while the recorded views-to-subscribers ratios were retained. As a result, dividing the figures shown in the table may produce small discrepancies.
How I Study This Group of Long-Form Videos
Animation and character stories. I compare Angel Spark TV, Ghibli Night, Nova Saga, Dabin Toons, 𝑴𝒂𝒛𝒚𝒅𝒂𝒂𝒓 𝑫𝒖𝒏𝒚𝒂, The Story Cafe, Goglipop Rhymes, and UM Cartoon TV in terms of characters, conflict, worldbuilding, and narrative length. I first check whether the same channel has multiple videos with similar performance, then judge whether its characters or plots can sustain continued development.
Life processes and outdoor survival. Vietnam Daily’s life on a boat and Tan SoLo Survival’s raft survival in a heavy rainstorm are useful for observing how environment, tasks, and process sustain longer viewing. The living conditions or extreme environments described in the titles need further verification through the original videos.
Demonstrations and gaming. Big Bro Teo’s spin art and MARINARA’s multi-character GTA scenes provide references for process demonstrations and game-based narratives, respectively.
Learning, travel, and conversation. Family Life English’s slow English, Joel Griffin’s cruise experience, and No Magic Pill with Blake Mycoskie’s interview about mental health represent different reasons people watch long-form videos. They require different knowledge, filming, or communication abilities, so I cannot apply the same production plan simply because each has high views.
When looking for similar examples, I recommend starting with the low-subscriber long-form rankings or keyword search on ShortsMonkey. After finding candidates, I return to the channel to compare older and newer videos on the same topic and record my assessment.
4. Eight Livestream Examples: Ongoing Processes, Sensory Experiences, and Repetitive Tasks
When I look at livestream examples, I focus on different questions: Can the visuals keep changing? Can the task continue naturally? Can viewers who join midway immediately understand what is happening?
For the following eight examples, I list the data together with my assessment. My judgments about account networks, reused material, and traffic quality in the table are leads for further verification.
| Example | Account | Representative livestream | Views | Subscribers | Views/subscribers | Date | Duration / dimensions | Assessment |
|---|---|---|---|---|---|---|---|---|
| Example 01 | PRATHNA RC | View livestream | 7.668 million | 20,000 | 375.9× | 2026-08-12 | 4 hr 00 min 720×1280 |
Forest-survival process stream; strong demand, but high risk around cross-account reuse and copyright. |
| Example 02 | Dani4x fitness | View livestream | 6.017 million | 196,000 | 30.7× | 2026-07-30 | 5 hr 13 min 1080×1920 |
Fixed-camera home workout; I can break down its pacing and exercise rotation. |
| Example 03 | Yana Outdoors | View livestream | 3.894 million |
6,250 | 623.1× | 2026-08-07 | 4 hr 01 min 720×1280 |
The same forest-survival formula; high views on a small account, but it appears to be part of a network using the same material. |
| Example 04 | Pigeon Health Lab | View livestream | 3.532 million | 148 | 23864.9× | 2026-08-13 | 9 hr 09 min 610×1280 |
An extreme small-account outlier; first verify traffic quality and reproducibility. |
| Example 05 | MF ASMR Studio | View livestream | 3.522 million | 13,000 | 268.9× | 2026-08-16 | 2 hr 02 min 1080×1920 |
Hands-on pegboard ASMR; low cost and low language dependence, so it is a testing priority. |
| Example 06 | Dani4x Empire | View livestream | 2.641 million |
58,000 | 45.2× | 2026-08-02 | 2 hr 03 min 1080×1920 |
The same fitness formula works on a sibling account, a strong signal of network reuse. |
| Example 07 | Sisco Candy ASMR | View livestream | 2.418 million | 67,000 | 36.1× | 2026-07-25 | 1 hr 22 min 1080×1920 |
Sorting and returning beans to place; strong sense of completion, global appeal, and low cost. |
| Example 08 | BigTee | View livestream | 1.841 million | 44,000 | 41.8× | 2026-08-04 | 1 hr 12 min 1080×1920 |
Fake-insect cleaning ASMR; highly stimulating, with risks involving disgust, deception, and moderation. |
I recorded video view counts here without separating views generated during the livestream from replay views, so these figures cannot be treated as concurrent viewer counts. I retained the views-to-subscribers ratios as recorded when I collected the examples.
From a production perspective, I continue breaking them down in four directions:
- Forest-survival process streams. For PRATHNA RC and Yana Outdoors, I compare their settings, actions, and sources of material. When similar content appears across accounts, I also need to verify reuse and authorization.
- Fixed-camera workouts. For Dani4x fitness and Dani4x Empire, I break down exercise rotation, pacing, and camera position. I examine them as possible “sibling accounts” because similar formulas appear on different accounts; the exact relationship between the accounts needs further confirmation.
- Hands-on operations and sorting tasks. MF ASMR Studio’s pegboard and Sisco Candy ASMR’s bean sorting rely little on language, show a clear process, and provide a sense of completion. If I were to test one low-cost direction first, I would prioritize a format such as the pegboard.
- Highly stimulating content and anomalous samples. For BigTee’s fake-insect cleaning, I consider the viewing experience, authenticity of the presentation, and moderation issues together. Pigeon Health Lab has only 148 subscribers but about 3.532 million views; when I encounter such an extreme value, I first investigate traffic quality and reproducibility.
These eight livestream examples range from 1 hour 12 minutes to 9 hours 09 minutes, so the investment differs greatly from a ten-second Short. Before choosing a direction, I recommend first working out what process you can sustain, then studying how references structure time.
5. Build a Script Library: Make Every Breakdown Leave Something Reusable
If these methods leave you with only one long-term practice, the one I most want to emphasize is this: Keep analyzing viral videos, break their scripts apart, and record what you learn.
If I only bookmark a link, I may need to understand the entire video again next time. Only by recording the story structure, opening, character relationships, and key elements can I gradually build my own creative resources.
Here is a document titled “Viral Video Analysis—Continuously Updated—07-01”, which you can use as a reference for organizing scripts. For me, recording “drinking a potion,” “smashing glass in the opening,” and “entering a body” is only the beginning; I also need to explain what function each serves in the story.
To help you start immediately, I listed the fields I recommend recording in the table below:
| Field | What to record |
|---|---|
| Example and date | Video link, channel link, publication date, and observation date |
| Data | Views, subscribers, and views-to-subscribers ratio; note whether each is from a screenshot or a live query |
| Niche and audience | Content type, language, character or character series, and target audience |
| Opening | What appears at first glance and what expectation it creates |
| Story framework | Character goal, obstacle, action, turn, and result |
| Elements | Props, settings, visual changes, sound, and repeated actions |
| Emotion | Surprise, sympathy, gratitude, joy, and so on, plus the specific beats where the emotion changes |
| Adaptation directions | The character motivations, course of events, settings, and ending I plan to change |
| Production requirements | Live action, animation, or AI; required shots, character assets, and time |
| Post-publication review | My work’s performance, what differed from expectations, and what to adjust next time |
A script library should contain both successful examples and attempts that did not meet expectations. Only by comparing results with the parts I changed can I judge which changes have value.
6. How to Increase Output: From Step-by-Step Production to a Workflow or Agent
After finding a topic, the production pipeline can be divided into scripting, images, video, and editing.
Traditional Approach: Complete Four Types of Tasks Step by Step
Start with the most basic process, which I divide into four steps:
- Break down the reference video.
- Produce text-to-image and image-to-video prompts.
- Generate images in batches.
- Generate videos from images in batches.
These four steps turn an idea into material that can be edited. Actual delivery also requires sequencing, editing, sound, subtitles, and a final-video check.
Workflow or Agent: Connect Inputs, Outputs, and Acceptance Criteria
A workflow is suitable for linking relatively fixed steps, while an Agent can participate in task breakdown, asset association, and progress tracking. Both require a clear definition of what each step delivers and when it is allowed to advance to the next.
| Stage | Input | Core operation | Output | Criteria for moving to the next stage |
|---|---|---|---|---|
| 0. Preparation | Accounts, API keys | Log in to tools, configure model services, and run a small test task | A working workflow environment | Text, image, and video models can all be called successfully |
| 1. Reference video | Reference-content link | Download the video and compress it if needed | Video file no larger than 20 MB | The file plays normally and can be uploaded for analysis |
| 2. Storyboard script | Reference video, adaptation requirements | Configure text nodes, then generate and revise the script | Final storyboard script | The story, shots, duration, and adaptation direction are clear |
| 3. Image assets | Final script, visual-style and character requirements | Generate characters, settings, products, and key storyboard images | A set of key visual assets | Characters, props, and setting styles are basically consistent |
| 4. Video generation | Script, key storyboard images | Split tasks intelligently, associate images, and generate in batches | Multiple video clips | All necessary shots have been generated successfully |
| 5. Editing and acceptance | Video clips, script | Download, sequence, edit, and add sound and subtitles | Final video | The narrative is clear, visuals are continuous, and the ending provides a payoff |
This example workflow keeps videos used for analysis within 20 MB. I am referring to the input requirement of the analysis tool; this figure does not apply when uploading to YouTube. If you change tools, adjust according to the corresponding requirements.
I split the production sequence and the review-and-rework process into two diagrams. First, look at the main workflow and confirm what each step should deliver:
Next, look at review and rework. When there is a problem, I return to the relevant stage to correct it and preserve the parts that have already passed:
In actual operation, I recommend connecting the following review and retry steps as well:
- Preparation. Log in to the tools, configure the model APIs, and use a small task to confirm that the text, image, and video models can all be called successfully.
- Acquire the reference. Select reference content, obtain a video file suitable for research, and check whether it plays. If it exceeds the workflow’s 20 MB requirement, compress it with a local tool first.
- Generate the script. Create a workflow, configure text nodes, upload the video, and state the adaptation requirements clearly. If the script is unsatisfactory, duplicate the task and rerun it, or edit it manually and save it.
- Generate visual assets. Associate the script with the characters, settings, products, and key storyboard images. If the images are inconsistent or unusable, revise the requirements first and create a new image node to retry if necessary.
- Split video tasks. This example workflow handles four storyboard shots per task. Correct the duration of the final group, then associate images in batches, generate the videos, and monitor progress. The group size can be adjusted to the tool and shot requirements.
- Handle failures. Rerun only failed or substandard tasks and preserve usable shots.
- Download and edit. Download everything as a package, arrange it according to the script, edit it, and add sound and subtitles.
- Accept and archive. If the final video fails acceptance, locate whether the problem comes from the script, assets, video generation, or editing, then return to the corresponding stage. Once it passes, deliver the final video and archive the project and assets.
What I want a workflow to save is the time spent on repetitive operations, leaving more energy for judging scripts and key shots. ShortsMonkey can help find examples, while the appropriate production tools can generate storyboards, images, and videos, connecting all the stages.
7. My Recent Observation: Proven Scripts Keep Appearing with New Characters and Formats
While organizing these examples, I had a strong impression: the visual styles and characters in many videos have changed, but the underlying scripts feel familiar. A classic comedy sketch can be recast with realistic animals, a live-action story can become AI-generated, and 2D animation can adopt a different form of expression. You can review the following examples alongside my observations.
| Reference example | My observation | Question I would continue analyzing |
|---|---|---|
| Realistic animals perform a classic comedy sketch | A classic sketch recast with realistic animal characters | Does the original joke still work after the characters change? |
| Animal comedy example | Another animal-comedy reference | Which actions and reactions carry the joke? |
| Viral-element reference video | Its viral elements can be broken down; when comparing montage formats like this, I also consider producing them with AI | Is the contrast clear, and how are the pieces of material connected? |
| catsfamilyemportuguesbrasil | When I see these familiar formats, I sometimes truly wonder what year it is | Is it an old story structure, or does it only look visually familiar? |
| Celebrity IP applied to a story structure | A reference for video generation and character transfer | Beyond the character, what does the new plot contribute? |
| Live action + AI | A viral reference for a hybrid format | How does switching between live-action and AI visuals create a reason to watch? |
| vocavoca | A children’s channel; I observe its shift from 2D animation to AI expression | How do the existing characters and stories connect with the new presentation style? |
| Climbing a castle by grabbing hair | Uses an exaggerated action to create a straightforward plot | How does the first shot establish the task, and how do later shots answer it? |
| LiveActionAnimeLab | A KPOP account with Japanese titles; when I observed it, it had been updating consistently since March | Are the characters, title language, and series content consistent? |
| Moituber23 | A Spanish-language children’s educational reference | How can educational information be combined with language children understand? |
| BhuvneshMaurya-y8k | I saw what looked like a repost of a year-old video gain traction, which really made an impression on me | The content source, permission, and changes still need investigation; high views do not justify direct reposting |
These examples make me value the script itself even more. Classic sketches, family affection, helping others, and completing missions can still be understood and enjoyed after the characters and language change.
So when I look at a new viral video, I ask two questions at once: What new format does it use? Is there a story structure behind it that I have seen before? Recording these separately helps me assess the current opportunity while continuing to build my script library.
8. Reaching YPP with Long-Form Videos: My Experience and Expectations
If I had to sum up my feeling in one sentence: long-form videos are difficult, but once they work, I value their potential to continue bringing views and revenue. I sometimes call this “passive income,” though the channel still needs maintenance and revenue can change.
Starting a long-form channel takes time. In my experience, if I use multiple accounts in the same niche, I consider launching 10–20 accounts at once, then continuing to operate and observe them. Over about 1–3 months, 3–5 accounts may gain traction. This is the range from my experience; your outcome still depends on your production capacity, content quality, and investment.
I also watch for whether a video breaks 10,000 views. To me, that is a signal that the content is beginning to receive feedback. If I want to move toward the YPP goal, I first hope to see such a breakthrough.
If you are working alone or do not yet have a stable production process, I do not recommend spreading your attention thin just to copy this number of accounts. First choose a clear niche, estimate how much content you can produce, leave time for validation, and then decide whether to expand.
If you are interested in long-form videos about space, you can continue with this document: “YouTube Space Niche References”.
9. Put the Method into Practice: Position First, Research Next, Then Build Your Own Series
Finally, I want to emphasize the three things I value most.
First, value recurring people and character series. My experience making YouTube videos differs from watching Douyin in China. In the Shorts I research, many viral videos center on a fixed person or IP. Each video continues that character’s story, viewers gradually get to know them, and new events have a familiar protagonist.
Second, find strong references and choose your niche before production. I do not recommend making one video every time you see a viral hit. First work out who the content is for, which person or topic it will center on, what viewing experience it offers, and whether you can keep producing it. Then screen viral videos that fit the channel.
Third, treat the script library as a long-term asset. This is something I especially want to remind you about. Trends change, characters change, and production tools change, but the motivation, conflict, reversal, and emotional payoff in a good story can continue inspiring new work. I believe scripts are assets that can span cycles and platforms, making them worth accumulating over time.
If you are ready to begin, I recommend proceeding in this order:
- Write a clear channel position and define the niche and range of characters you want to research.
- Find candidates in recent rankings, keyword searches, and existing references.
- Return to the original video and assess its reference value together with the channel’s other work.
- Record the opening, story framework, key elements, emotion, and data date.
- Write your own adaptation plan and define the new plot and expression you will add.
- Use a workflow to complete the storyboard, images, video, and editing, accepting each stage along the way.
- Record the results after publishing so the next creation has concrete evidence to build on.
You can start by finding an example related to your channel positioning on ShortsMonkey, breaking it down completely, and then writing your own version. More than anything, I hope that after reading this, you will truly add the next script to your library and test your judgment in later work.