Two Weeks on the Video, Forty Seconds on the Title
He filmed it over two days and spent a week editing.
Then graded the colour, remixed the audio, and corrected the subtitles word by word.
Then uploaded it.
And while it was uploading, he typed the first title that came to him into the title field. In the description field he wrote a single line. Then pressed publish.
Two weeks of work, and forty seconds for the text that decides whether anyone opens the video at all.
A week later he looked at the numbers: a large number of people had been shown the video, and very few clicked.
The problem wasn't the video. The video was good, and nobody saw it.
Why the Title Is Written Last
A title written before the script imposes a promise on the video that it may not be able to keep, so you end up either with a video that lets its title down or a title that doesn't describe what you made.
Writing it after the script is finished makes you choose from what you actually have. And "last" here means after the script, not at the moment you press upload.
Four Title Formats That Work
The specific promise: "how to write a video script in one hour". It states the result and the time.
The question the viewer is already asking: "why do channels stop after six months?". On the condition that the video actually answers it.
The paradox: "the best decision I made for my channel was publishing less". An apparent contradiction that demands explanation.
Strict specificity: "three mistakes in the first thirty seconds". Numbers and boundaries make the promise measurable.
What Ruins a Title
Overstatement: a title promising more than the video delivers raises clicks and raises drop-off with them, and the net result is negative.
Clever vagueness: a title that hides its subject to provoke curiosity stops anyone searching for the subject from finding it.
Excessive length: the last words get truncated in most interfaces, so put the decisive part at the start.
Padding: "a very important video you have to see" adds no meaning and takes up space.
The Description: Three Layers
The first two lines are what's visible before someone presses "more", and their job is to complete the title's promise rather than repeat it. Write them as if they're your last chance to persuade someone hesitating.
The body explains what the video covers, and carries the words people search the subject with, naturally, inside readable sentences. Piled-up keywords hurt rather than help.
Links and timestamps come last: a chapter index, the sources you relied on, and any references you mentioned in the video. Listing sources here is a practice worth committing to, especially if your video carries figures or claims, as in: research before writing: gathering your material and verifying it.
Write More Than One Title
The common mistake is writing one title and uploading, which is exactly what happened at the top of this article.
Write five, then read them in sequence and ask: which would I click if I weren't the person who made this? And which describes the video honestly? The title that answers both is your candidate.
And keep the rejected titles inside the video project, because some of them will suit a short clip derived from it, or a future video.
Consistency With the Thumbnail
The title and the image work together rather than separately: if the image repeats the title's text you've wasted half your space.
Let the image show what the title doesn't say, and the title say what the image doesn't show.
Before Uploading
Read the title aloud, then read the first two lines of the description, then ask yourself one question: can the video deliver this?
If you hesitate, the problem is either in the promise or in the video, and fixing it now is cheaper than fixing it after viewers leave halfway through, as explained in: after publishing: reading a video's performance and feeding it back into your ideas.
And give this step half an hour out of those two weeks, not forty seconds.