You open YouTube to watch a tutorial. Instead, the homepage is full of gaming videos, vlogs, and content you watched once three months ago.
Or the opposite: you sit down to relax on a Sunday and YouTube serves you programming tutorials and business content.
The algorithm isn't broken. It's doing exactly what it was designed to do — and that design doesn't account for the fact that you use YouTube for completely different purposes at different times.
How YouTube's recommendation algorithm actually works
YouTube builds a single interest profile for your account. Everything you watch, like, search for, and spend time on contributes to this profile. The algorithm then uses this profile to surface content it predicts you'll watch.
The problem: your profile is an average of everything you've ever watched, not a reflection of what you want right now.
- ✕No context awareness — The algorithm doesn't know it's Monday morning and you're trying to learn something. It only knows what you've watched before.
- ✕Recency bias amplification — If you watch several videos in a new category in one session, the algorithm over-weights that session in your recommendations. One rabbit hole can shift your feed for weeks.
- ✕No guest mode — If a family member uses your account, their viewing becomes part of your profile. The algorithm can't tell it was someone else.
- ✕Engagement optimization — YouTube optimizes for total watch time, not relevance to your current needs. A video that got you to watch for 40 minutes will be surfaced again, even if it was entertainment during a work session.
Why the usual fixes don't work
Clicking 'Not interested'
Works for that video. The algorithm finds another angle to surface the same category. You'd need to do this hundreds of times to meaningfully shift your feed.
Clearing your watch history
Resets everything — including all the personalization you actually want. You lose years of signal just to undo a few bad sessions.
Multiple accounts
Works in theory. In practice, you forget to switch, use the wrong account, and end up mixing recommendations anyway. Managing multiple YouTube logins is friction most people abandon within a week.
Incognito mode
No recommendations at all. You lose the entire personalized experience just to avoid the wrong personalization.
A better approach: tell YouTube what context you're in
Instead of fighting the algorithm or abandoning personalization, you can filter what it shows you based on your current context — before you start watching.
The idea is simple: switch modes before you open YouTube, the same way you might switch from a work browser profile to a personal one.
Work Mode
Filters recommendations to tech, programming, science, and business. Entertainment content is hidden.
Hobby Mode
Gaming, music, and personal interest content comes through. Work content is filtered out.
All Mode
Normal YouTube. No filtering. Switch here when you want everything.
ContextTube: context switching for YouTube
ContextTube is a free Chrome extension that adds Work, Hobby, and All modes to YouTube. It uses client-side keyword filtering — no YouTube API, no data collection, no account required. You click the extension icon, choose your context, and your recommendations filter instantly.
It doesn't change your watch history or try to teach the algorithm anything new. It just filters what you see while you're in that mode — so your programming tutorials are there when you need them, and gone when you don't.
Free · Chrome Extension · No account
Try ContextTube
Switch YouTube to the right context in one click.
See how it works →Frequently asked questions
Why are my YouTube recommendations suddenly different?
YouTube's algorithm updates recommendations based on recent watch history. A single session of watching content outside your usual topics — even by someone else on your account — can shift your recommendations significantly. The algorithm doesn't know which viewing session was typical for you and which wasn't.
How do I reset my YouTube recommendations?
You can clear your watch history in YouTube settings, but this removes all personalization — including content you actually want recommended. It's a blunt fix. A better approach is context-based filtering, which shows you relevant content without erasing your history.
Why does YouTube keep recommending videos I don't want?
YouTube optimizes for watch time across your entire account. If you've ever watched something — even once — the algorithm may keep surfacing similar content because it associates it with engagement. There's no built-in way to tell YouTube 'I watched this once, don't infer anything from it.'
Can someone else's viewing ruin my YouTube recommendations?
Yes. If someone else watches content on your account — a family member, a friend — the algorithm treats it as your preference. YouTube doesn't distinguish between account holders and guests.
Is there a YouTube work mode?
YouTube doesn't have a built-in work mode. Extensions like ContextTube add this functionality by filtering recommendations based on content categories — so when you're in work mode, entertainment content is filtered out, and vice versa.