Because the map is so small and speeds higher than that make Viper (mostly) useless/normal imp, and many reasons too but curious to know what you all think!
I am building AE plugins in the $0-15 range and I want to hear from professionals if there is something specific they want next, an extension/plugin that has never been implemented before.
Pulled a dataset of 67,715 long-form YouTube videos and split them by the pronouns in the title. The result genuinely surprised me.
my findings:
- Titles using first-person words (I, my, we): +48% median views
- Titles using second-person words (you, your): −11% median views
That's a 59-point swing between two choices every copywriting guide tells you to make the other way ("make it about the viewer, use 'you'"). On YouTube, that advice is backwards.
Broken down by word, titles containing "I" earn +65%. The worst word is "your" at −23%, it always precedes an instruction ("fix your audio", "your titles are wrong"), and an instruction is the opposite of a story.
Also, 2 checks I ran so it's not nonsense:
- "It's not a length thing". First-person titles aren't just shorter, I re-ran it inside every title-length bucket and first-person stays positive at all of them.
- "It holds across niches" - positive in 11 of 13. The one real exception is Comedy, where direct address (you won't believe) actually works.
Best guess at the mechanism: a title is a promise about the video's shape. "I tried X" promises a story with a protagonist. "You should try X" promises a lecture. People usually click the story.
Click here for Full write-up with the per-niche data, charts, and methodology
What's new in v2:
How it works:
Select a layer or property in the timeline → pick an expression → hit Apply. If the expression has parameters it asks you first, otherwise it fires instantly. Works on multiple selected layers at once.
It's free, download link in comments.
Would love to see what expression packs people build!
Seriously I think this would matter a lot so ppl stop saying "vent cd?" Or get confused about a specific role count etc. It would help a lot imo
I built VersionCraft. Drop in a master comp, pick which aspects/resolutions you want, add language variants (or paste a CSV), and hit one button. It builds every variant comp, scales/centers content with your chosen reframe mode, and queues them all to AE's Render Queue or Adobe Media Encoder.
What it does:
• Multi-aspect generation - 16:9, 9:16, 1:1, 4:5, 21:9 + custom WxH
• Language variants - name your text layers, fill replacement values per variant, or import a CSV (one column per text layer, one row per language)
• Smart reframe - Fit / Fill / Stretch per batch
• Filename templating - {name}_{w}x{h}_{lang} tokens
• Render preset selector - picks up any Output Module / Render Settings
template you've saved in AfterEffects, applied to every queued item
• Direct dispatch to AME or AE's render queue
• Source nests instead of copies - edits to the master flow into every variant automatically
link in comments.
Happy to answer questions or take feature requests.
Hey everyone, so I was working on a SaaS the last few weeks and finally made ViewsKit, made it because there was not any single 1 tool that does everything from scripting to SEO. Most were standalone tools. Also the AI is trained specifically by best YouTube advice so it isn't generic. Hopefully I get people to try it!
Hello everyone, so I've been working on specific prompts for Claude to be a better scriptwriter for YouTube specially since retention falls in the first 30 seconds, and I want to know if it sounds great. Of these 3 generation one is VidIQ, one is ChatGPT and one is my tool shuffled.
Pulled a dataset of 60,693 YouTube videos across the main content categories and tried to find out what day should you actually upload.
my findings were: Monday's median views are 27% below Saturday's. No other day comes close, Wednesday is next-worst at -19%, then Tuesday at -16%. In 7 of 12 niches, therefore Monday is the single worst publishing day, *except for some niches like Gaming. You can also find by niche in the second image.
Wrote up the full analysis with methodology and per-niche charts here if anyone wants the deep version: https://viewskit.com/blog/the-monday-penalty