Most developers who start a channel stop because of one question asked every day: what should I make today? Without a method, the answer depends on mood, and mood produces videos about whatever you happened to read last night.
A topic decision should take ten minutes and leave a paper trail: the candidates you considered, why one won, and the sources that prove it's true. This lesson is the method I use for my own channel, HemantInsights on YouTube, where I explain real tech incidents in about two minutes.
1. Know what you've already madeLink to this section
Before looking for new ideas, list what's already published, grouped by topic, with views. You're looking for two things: what works and what's missing.
Here is a snapshot from my channel's Shorts in the same week (views at the time of writing):
| Short | Pillar | Views |
|---|---|---|
| true + true + true = 3?! | JavaScript | 1.3K |
| This Bug Scored 10/10 | Security | 1.2K |
| The Night One Command Deleted Production | Stories | 1.2K |
| 1.5 Million Keys. One Missing Setting. | Security | 1.1K |
| The Chatbot Lied. Who Pays? | AI | 6 |
| WordPress "Critical Error"? Find the Cause in 4 Steps | WordPress | 41 |
The pattern is clear: a concrete story with a number or a surprising result in the title wins. "The Chatbot Lied" is a good story, but its title has no number and no searchable topic word. The WordPress Short is useful, but it's a how-to dressed as a Short, and it had the weakest hook.
The gap is just as clear: WordPress had one video, while security had four. That makes a WordPress security story worth extra points this week.
If your channel syncs to your website (lesson 3 shows how), this list is one command away. Otherwise, YouTube Studio's content tab sorted by date works fine.
2. Collect fresh candidatesLink to this section
Look for stories from the last one to two weeks that fit your pillars. Good hunting grounds for developer topics:
- Security advisories and exploited-vulnerability lists: vendor security pages, CISA's Known Exploited Vulnerabilities catalog, GitHub security advisories for the frameworks you use.
- Official release notes: WordPress, Next.js, React, Node.js, PHP. A security release with a twist is a ready-made story.
- Incident write-ups: post-mortems published by companies after outages.
- Questions people keep asking: forum threads and comments on your own videos. One comment on my 0.1 + 0.2 Short became a whole new video.
Write down five to eight candidates. For each, note at least two sources you could cite.
Evergreen still counts. A ten-year-old incident told with a new angle can beat this week's news. Mix both: fresh stories bring timeliness, evergreen ones keep earning search views.
3. Score them, in writingLink to this section
Give every candidate a 1–5 score on seven questions:
| Criterion | The question |
|---|---|
| Hook | Can the first two seconds be a surprising statement that is true? |
| Fit | Does it match one of your pillars and the people you want to reach? |
| Freshness | Did it happen in the last two weeks, or is it evergreen with a new angle? |
| Search | Will someone type it into YouTube or Google: an error message, a product name, a "how to fix"? |
| Lesson | Is there one concrete thing the viewer can do afterwards? |
| Gap | Haven't you covered it, and is its pillar due? |
| Site link | Is there an article, tool or service on your website it naturally leads to? |
Add the scores and pick the highest. Here is a real table from a recent day:
| Candidate | Hook | Fit | Fresh | Search | Lesson | Gap | Site | Total |
|---|---|---|---|---|---|---|---|---|
| WordPress core update flaw, actively exploited | 5 | 5 | 5 | 5 | 5 | 5 | 5 | 35 |
| Node.js flaw with a federal patch deadline | 4 | 4 | 5 | 4 | 4 | 4 | 3 | 28 |
| npm malware hiding its server on a blockchain | 5 | 4 | 5 | 3 | 4 | 4 | 3 | 28 |
| Next.js monthly security release | 3 | 4 | 5 | 4 | 4 | 2 | 4 | 26 |
The WordPress story won on every row: a paradox for a hook ("the update was the way in"), a live threat, a huge searchable audience, a clear action, a pillar that was due, and a direct link to WordPress work I do for clients.
The scores aren't scientific. Their job is to force you to look at every angle, and to make tomorrow's decision as quick as today's.
4. Verify before you writeLink to this section
This is the step that separates a channel people trust from one they scroll past. For the chosen topic, go to the primary sources:
- the vendor's own advisory or release notes;
- the original researcher's write-up;
- a government catalog entry or an established outlet quoting them.
Then build a small table: every fact you plan to say, next to the link that proves it. If a fact has no source, it doesn't go in the script.
In the WordPress story, several aggregator sites said the bug affected versions "4.7 through 7.1.3". The researcher's own write-up said 2.7.0 through 7.1.3; 4.7 is where the fix was backported down to. That's an easy mistake to repeat, and exactly the kind a knowledgeable viewer will correct in your comments.
Never round up a scary number. If a security company says it blocked 13,095 attacks in a day, say exactly that. Don't turn it into "13,000 sites hacked". Accurate numbers are still dramatic, and they keep your credibility.
A few rules worth writing at the top of your sources file:
- Say "reportedly" when the source does.
- Don't name or blame private individuals; use companies, products and public reports.
- Keep a short "avoid saying" list for claims that sound right but aren't supported.
5. Plan the pair: one Short, one long videoLink to this section
Each topic becomes two pieces:
- a Short (30–50 seconds, vertical) that tells the story and ends by pointing to the full version;
- a long video (two to four minutes while your channel is small) that explains how it happened and what to do.
Post the Short first and the long video the next day, then set the long video as the Short's related video. Shorts bring most of the views on a small channel; the long video is where viewers who want more end up, and where a potential client sees how you think.
6. Make it a habit, not a huntLink to this section
The method only works if you run it every day. What helps:
- A fixed slot. Research at the same time daily; post at the same time daily. Several videos posted within an hour compete with each other for the same small test audience.
- One topic per day. Batch the research if you like, but publish one.
- A folder per topic.
content/videos/2026-11-09-topic-slug/withshort.md,long.md,publish.mdandsources.md. Next week you can see exactly why you chose what you chose. - A weekly look at the numbers. Make more of what's working, drop what isn't, and update your pillars.
In lesson 3 I'll show how I let Claude do the collecting, scoring and source-checking each morning, with me approving the pick, so the whole step takes a few minutes.
Frequently asked questionsLink to this section
How many pillars should a small channel have?Link to this section
Three to five. Enough variety to post daily, but few enough that a new viewer can tell what the channel is about from three videos.
Should I chase trending topics that don't fit my channel?Link to this section
Rarely. A trending topic with a fit score of 1 brings viewers who won't watch your next video. The scoring table makes this trade-off visible instead of tempting.
What if I can't find two sources for a story?Link to this section
Skip it or wait. A day without a video costs less than a correction pinned under one.
Test yourself
3 quick questions on this lesson.
Finished this lesson?
Mark it complete to track your progress through the learning paths.