Transparency
Measured first. No made-up scores.
Every HookSignals tool starts from public YouTube data — the most-viewed videos on your topic, from similar creators — and counts how your packaging differs. The AI verdict comes on top of that, with its reason. There are no 0–100 packaging scores.
Methodology
What is measured, what the AI reads — and what we never claim.
Measured
- ◆Public counts: each video's views and upload date, and its channel's subscriber count
- ◆Comparison videos: the most-viewed videos on your topic and format, from public YouTube data
- ◆Counted differences: title length, numbers, questions, words on thumbnails, greetings in the first 15 seconds
AI read on top
- ✓A verdict — publish, change one thing, or rework — with the reason
- ✓Which part is weak next to the comparison videos, citing them by number
- ✓Rewrites for the title, the first line and the thumbnail text
Never claimed
- ✗Guaranteed views
- ✗Actual YouTube CTR
- ✗Actual retention data
- ✗A 0–100 score dressed up as precise
Actual CTR, retention curves and watch time are only available inside YouTube Studio, so HookSignals does not estimate them.
Comparison videos — public counts, read at the time of the search
Raw views mostly reflect channel size: a million views on a ten-million-subscriber channel is a normal day, while fifty thousand views on a channel with two thousand subscribers deserves a closer look. So every comparison video is shown with its public view count and its channel's public subscriber count, and you judge the gap yourself. Here is how the comparison videos are chosen and counted:
- ◆Videos come from a YouTube Data API search on your topic. Their view counts, upload dates, durations and channel subscriber counts are the public numbers at the time of the search, cached for up to 7 days per topic and dated on the page.
- ◆Videos are ordered by public view count. Videos with very few views, other languages and obvious off-topic matches are left out, along with official music videos, trailers, livestreams and auto-generated channels.
- ◆Comparison videos are limited to the format you are checking: up to 3 minutes or over 3 minutes. Duration alone cannot confirm whether YouTube classifies a video as a Short.
- ◆Lifetime views are not age-matched. A newer upload has had less time to accumulate views; release timing, topic, traffic sources and promotion can all affect a count, so a view count is not a measure of why a video worked.
- ◆Counted differences (title length, numbers, questions, formats, words on thumbnails) are tallies over the videos shown; nothing is estimated.
A view count describes how a published video has performed so far — it is a public observation, not a prediction, and it can change as views accrue. It uses only public view counts and upload dates; no channel login, no Studio data.
What the analyzer actually does
- ◆Fetches public video metadata via the YouTube Data API — title, views, likes, duration and thumbnail URL.
- ◆Shows the video's public numbers: views, likes, comments and duration.
- ◆Puts the title next to the most-viewed videos on the same topic and counts the differences.
- ◆Gives a verdict with the reason, citing the videos it compared with.
- ◆Returns improvement suggestions — alternative titles, hook rewrites, thumbnail text and a description angle.
What the analyzer does not do
- ✗Does not access your YouTube account or YouTube Studio.
- ✗Does not read private analytics — actual CTR, watch time or audience retention curves.
- ✗Does not guarantee views, clicks or channel growth.
- ✗Does not access private or unlisted video data.
How to use a verdict
A verdict is not a prediction. It tells you which part of your packaging — the title, the first line or the thumbnail text — is weak next to the most-viewed videos on your topic, and what to change before you publish. Look at the videos it cites and judge for yourself.