Your title is doing two jobs at once, and most creators only optimise for one of them. It has to tell YouTube's search and recommendation systems what the video is about, and it has to convince a human being scrolling at speed that this particular rectangle is worth eleven minutes of their evening.
Get one right and lose the other, and the video stalls. A keyword-stuffed title ranks for a query nobody clicks. A clever, mysterious title gets clicked by nobody because it surfaces for nothing.
Here is the framework we use, the constraints you are actually working inside, and the patterns that keep earning clicks.
The constraints that actually matter
YouTube allows 100 characters in a title. That is the technical limit, not the useful one.
- Around 60 characters is where desktop search results start truncating.
- Around 40 characters is where mobile home-feed cards start truncating, and mobile is the majority of most channels' impressions.
- The first 40 characters therefore carry almost all of the weight. Front-load them.
This does not mean every title must be 40 characters. It means the part a viewer must read to decide has to fit in the first 40. Anything after that is a bonus that some surfaces will show and others will cut.
Check your own analytics before trusting any rule of thumb. In YouTube Studio, open Analytics → Content → Impressions and look at the traffic-source split. If most impressions come from Browse and Suggested rather than Search, your title is competing on curiosity against a thumbnail, not on keyword match against a query.
Step 1: Start from the query, not the video
Before writing anything, decide what a person would type to find this video. Not what the video is about — what someone who does not know your video exists would search for.
There is a real difference between:
- "My Thoughts On The New Camera" — describes the video
- "Sony A7 V After 30 Days: What Nobody Mentions" — describes the search
The second one contains a phrase people actually type, and adds a reason to click. The first one is only findable by people who already follow you.
Run your candidate phrase through the tag extractor on the top three videos already ranking for it. If the same two or three phrases keep appearing across competitors, that is the vocabulary the topic uses. Match it — do not invent your own synonym for a term the audience already has a word for.
Step 2: Put the keyword where it survives truncation
Once you have the phrase, place it as close to the front as the sentence allows.
| Weaker | Stronger |
|---|---|
| "Everything I learned about YouTube keyword research this year" | "YouTube Keyword Research: What Actually Works in 2026" |
| "Let me show you how I edit my podcast in Premiere Pro" | "Premiere Pro Podcast Editing: My Full Workflow" |
| "This is the best budget microphone I have ever used" | "Best Budget Microphone 2026: 6 Months of Testing" |
The pattern in the right column is the same each time: keyword phrase, colon, reason to click. It is not the only pattern, but it is the most reliable one, because it satisfies the search system before the colon and the human after it.
Step 3: Add one specific promise
Vague titles lose to specific ones almost every time. Specificity is what separates a title that sounds like every other video on the topic from one that sounds like it contains information.
Three cheap ways to add it:
- A number. "6 Months of Testing" beats "Long-Term Review". "3 Settings" beats "Some Settings".
- A constraint. "Under $200", "In 10 Minutes", "Without a Second Camera" — constraints signal that the advice fits a real situation.
- A named result. "…That Doubled My Watch Time" is stronger than "…That Improved My Channel", because the outcome is measurable.
Six title patterns worth stealing
The plain-search pattern
[Keyword]: [Specific promise]
Best for tutorials and reviews where search is the primary traffic source. It is boring and it works.
The number pattern
[N] [Things] That [Outcome]
Works because the viewer can predict the structure of the video. They know what they are getting and roughly how long it takes.
The mistake pattern
[N] [Topic] Mistakes (And How to Fix Them)
Loss aversion is a stronger motivator than gain. People click to check whether they are making the mistake.
The comparison pattern
[A] vs [B]: Which Should You Actually Use?
Excellent for search intent — people compare before buying or committing. See our breakdown of tags versus hashtags for what a comparison post looks like when the answer is "both, differently".
The timeframe pattern
I [Did Thing] for [Time]. Here's What Happened.
The commitment is the hook. It only works if the timeframe is genuinely notable — 30 days is interesting, 2 days is not.
The contrarian pattern
Stop [Common Advice]. Do This Instead.
High ceiling, high risk. It needs the video to actually deliver a defensible counter-argument, or the comments will say so.
Every one of these patterns fails when the video does not deliver on it. A title that overpromises raises click-through rate for a week and depresses average view duration permanently — and duration is the metric that decides whether YouTube keeps showing the video after the first push.
Step 4: Write ten, then cut to three
The single biggest improvement most creators can make is volume. The first title you write is almost never the best one — it is just the first one your brain reached for.
Write ten variations. Then judge them against three questions:
- Would this surface for the query I picked in step 1?
- Does the decisive part fit inside 40 characters?
- Is there one specific, checkable promise?
If generating ten from scratch is the part that stalls you, the title generator will produce a batch from your topic and keyword so you are editing options instead of staring at an empty box. Treat its output as raw material — pick the two strongest, then rewrite them in your own voice.
Step 5: Match the title to the thumbnail
The title and thumbnail are one unit. They should not say the same thing twice.
If the thumbnail shows the result, the title should say why it matters. If the thumbnail shows the setup, the title should name the payoff. Redundancy wastes half your surface area — a thumbnail that reads "SONY A7 V" next to a title that starts "Sony A7 V" has spent two slots on one fact.
A fast way to audit this: pull the thumbnails of the videos currently ranking for your query with the thumbnail downloader, lay them out next to their titles, and look for the pairs that split the work rather than repeating it. Our thumbnail guide covers the sizing and legibility side.
Step 6: Measure, then rewrite the losers
Titles are not permanent. Renaming a video does not reset its performance, and YouTube re-evaluates a video's click-through rate continuously.
The workflow:
- Publish, then leave it alone for 7–14 days so the data settles.
- In Studio, compare the video's impressions click-through rate against your channel's median for the same traffic source.
- If it is meaningfully below, change the title — one variable at a time — and give it another two weeks.
- Track whether the video's ranking for your target query moved using the rank tracker.
Changing the title and the thumbnail simultaneously tells you nothing about which one mattered. Change one.
A worked example
Say the video is a walkthrough of setting up a second channel for Shorts.
- First draft: "Setting Up My Second Channel" — no keyword, no promise, findable by nobody.
- Add the query: "How to Set Up a Second YouTube Channel" — findable, but generic; a dozen videos say this.
- Add specificity: "How to Set Up a Second YouTube Channel for Shorts" — narrower query, less competition, clearer audience.
- Add the promise: "Second Channel for Shorts: Setup + What I'd Do Differently" — the keyword is front-loaded, and the second half offers something the other dozen videos do not have.
Final length: 58 characters, with the decisive phrase inside the first 30.
The short version
- Pick the query first, write the title second.
- Front-load the keyword; assume everything after 40 characters may be cut.
- One number, constraint, or named outcome per title.
- Write ten, keep three, publish one.
- Split the work between title and thumbnail instead of repeating it.
- Revisit underperformers after two weeks and change one variable.
None of this is about tricking anyone. A good title is an accurate description of a video someone wants to watch, written in the words they would use to look for it.
