Choosing a topic, and a niche

§1Why the topic carries the Short

On a faceless channel there is no familiar face to come back for and no personality the viewer already trusts. What stops the thumb is the subject itself. YouTube says that Shorts are ranked based on performance and viewer personalisation, and among the signals it names are the share of viewers who chose to view a Short rather than swipe past it, and how much of it they watched². In other words, every Short is judged on how real people respond to it, one viewer at a time.

YouTube also says its system uses fresh performance data for each individual video rather than relying on past results⁴. That cuts both ways. A small channel is not locked out by its history, and a big channel is not carried by it. Each topic gets its own hearing, which is exactly why choosing it well is worth more time than most creators give it.

Good editing can rescue a decent topic, but it cannot rescue a dull one. A beautifully cut Short about "interesting facts about Rome" will still lose to a plainer Short about the day Roman engineers shipped an entire Egyptian obelisk across the sea, because the second one is a story and the first is a heading.

§2What makes a topic watchable

We suggest testing every candidate topic against three questions. They are simple, but most weak topics fail at least one of them.

1. Is there one specific, surprising fact, and can you check it? Not "volcanoes are dangerous" but "a town in Italy was buried so quickly that the bread was still in the ovens". The surprise is what earns the first second; the fact being verifiable is what earns trust and survives the comments.
2. Is there a question the viewer wants answered? A good Short opens a loop: why did a city build its streets on stilts, how did one misprinted map send ships to an island that does not exist. The viewer stays to close the loop. If there is no question, there is no reason to reach the end.
3. Is the subject concrete? A person, a place, an object, a single day. Abstract themes such as "the economy" or "the history of science" are hard to picture and harder to film. A specific bridge, a specific shipwreck, a specific law with a strange consequence gives the footage something to point at.

A useful trick is to write the topic as the first line a narrator would say. If that line already makes you want to hear the second one, the topic is probably alive. If it sounds like a chapter title, keep digging until you find the story inside it.

A topic is a promise. Pick the ones you can keep before the viewer's thumb moves.

YouTube's own beginner guide makes the same point from the other end: capture attention in the first few seconds to prevent viewers from scrolling⁶. Those first seconds are mostly the topic speaking. The hook line is just the topic said well.

§3A consistent niche without repetition

There are two ways to get the niche question wrong. One is to have no niche at all: a Short about ancient Egypt, then one about phone batteries, then one about a football transfer. A viewer who enjoyed the first has no reason to expect the next one will be for them, so subscribing feels pointless. The other is to have a niche so narrow and a format so fixed that every Short is the same video with a new noun.

The second mistake has a policy behind it. In July 2025 YouTube renamed its "repetitious content" policy to "inauthentic content", covering content that is repetitive or mass-produced, and said such content has always been ineligible for monetisation¹. The examples it lists include similar or repetitive content with low educational value, commentary or narrative, or minimal variation across videos; image slideshows, templated storylines or scrolling text with little or no narrative or educational value; and AI-generated content made with generic or unoriginal templates that gives the impression of mass production¹. YouTube's CEO letter for 2026 also describes building on its anti-spam systems to reduce the spread of low quality, repetitive content⁵.

So the sensible shape, in our view, is a clear territory with many doors into it. Keep the subject area recognisable, and vary the angle: one Short is a disaster, the next is an invention, the next is a person, the next is a map, the next is a law. Vary the period and the place. Vary the kind of surprise. A viewer should be able to guess what your channel is about without being able to guess the next Short.

Here is what that looks like for a channel about the history of cities. One week might hold a Short about a street that was raised by several metres while people still lived on it, one about the morning a capital's traffic switched to the other side of the road, one about a lost river buried under a capital, and one about a mayor's strange law that is still on the books. Same territory, four different doors, four different kinds of surprise. None of them could be made by swapping a name into the previous script.

How a video is made is not what decides this. YouTube says videos made with AI can be recommended as broadly as videos made without AI³, and its monetisation policy applies to the content regardless of the tools behind it. What matters is whether each Short brings something of its own: real research, a real source, a narrative that explains rather than just lists.

§4Where to find topics

Topics rarely arrive one at a time when you need them. It works better to keep a few wells and draw from them regularly. Inside ViewMade there are two rooms built for this, and neither of them is a guessing game.

• Discover shows what is breaking out in a niche right now, and who published it. Use it to see which kinds of subjects are finding an audience in your territory this week, then ask what angle nobody has covered yet. A Short that is breaking out is a signal about the subject, not a script to copy. "Film something like this" opens Studio with that title, and the free footage check runs before anything is charged.
• Ideas writes subjects worth filming in batches: press "Generate 5 ideas", steer the batch with a theme if you like, and each card gives you a title and an angle. "Is there footage?" checks for free whether real footage exists, and "Film it" sends the idea to Studio. Use it when you know your niche but the list has run dry.
• Keywords scores demand and competition for a search term before you commit to it, which helps when a topic could be phrased several ways.
• Daily Ideas suggests fresh, channel-fit ideas from real YouTube data once your channel is connected.
• Outliers finds your own breakout videos, the ones beating your channel's median, so you can see which of your past subjects deserve a second angle.

Outside the product, the best wells are the places where curious people already ask questions: the comments under your own Shorts, the "what happened next" replies under a popular post, the footnotes of a good history book, the anniversary of an event in your territory. A museum's collection page can give you ten objects, each with a story. A city's old newspaper archive can give you a year's worth of strange local events.

§5Build a thirty-topic list

Thirty is not a magic number. It is simply enough to see patterns in your own choices, enough to cover several weeks of publishing, and small enough to finish in one sitting. A list also makes variety visible: when the topics sit next to each other, you notice at once if eight of them are the same idea.

1. Write one sentence for your territory

   Name the subject area your channel covers, in a sentence a viewer would understand, such as "strange true stories from the history of cities". This is the fence; everything on the list sits inside it.

2. Fill from three wells

   Take about ten candidates from Discover and Ideas, about ten from public tools such as Keywords and Daily Ideas, and about ten from your own reading and your comments. Different wells give different kinds of topic.

3. Run the three questions

   For each candidate, write the specific fact, the question it opens and the concrete subject. Cut anything that is missing one of the three, or anything whose central fact you cannot verify from a named source.

4. Check the spread

   Tag each surviving topic by angle (person, place, object, event, law, invention) and by period. If one tag dominates, swap some topics out until the list has real variety.

5. Write the first line

   Give every topic the opening line a narrator would say. Put the strongest five at the top; those are your first week.

Once the list exists, the production side is the easy part. In ViewMade each topic on it is one line in the Studio: you type the subject, and the service researches it, writes the script, sources real archival footage, records the narration and returns a finished vertical Short with captions and an upload package. If you see the Automatic chip on the composer, full members can also give it a niche or a topic list and have Shorts made overnight. Your list is still the best guide to what that plan should cover, because the territory and the variety are decisions only you can make.