Page 4 of 5 — The platforms up close

Three platforms.
One logic.

Instagram, YouTube, and Spotify each have a different surface — but underneath, the same engine runs. Here's how the loop plays out on each one, and who gets hurt by it.

This section is interactive. Use each platform below like you normally would — like posts, click videos, skip or save tracks. Every action you take is tracked and added to your algorithmic profile. Interact with all three platforms to see your full profile revealed at the end.

Your profile
Interact with each platform to build your profile 0%
Like, save, or scroll past each post
BF
beauty.filter
Sponsored · Beauty & Skincare
"Flawless skin. One product away from perfection."

Aspirational lifestyle imagery.
Luxury skincare ad.
#skincare #glow #beauty — Limited offer. Shop now.
Algorithm note: beauty + lifestyle · CPM $12.40 · High advertiser value
CR
rashida.creates
@rashida.creates · 4,200 followers
"My locs are professional.
The workplace needs to catch up."

Post about natural hair and workplace equity.
#naturalhair #BlackCreator #workplaceequity
Algorithm note: racial justice tag · CPM $1.80 · Advertiser risk filter active · Reach suppressed
FT
fitspo.today
Sponsored · Fitness
"Before and after."

Body transformation content.
Body as problem. Body as project.
#fitspo #bodygoals #transformation — 12 weeks. No excuses.
Algorithm note: fitness + body · CPM $9.20 · Saves 3.2× above average

The body as monetizable inventory

Notice the CPM figures on each post. The beauty ad is worth $12.40 per thousand impressions. The creator discussing natural hair and workplace equity is worth $1.80 — with an active advertiser risk filter suppressing her reach. This is Noble's argument made concrete: the algorithm doesn't decide whose voice matters. The advertiser market already did. The algorithm just enforces it at scale.

Noble's argument applied

When Instagram suppresses content naming race or discrimination — not by policy, but by advertiser preference — it performs the same function as Noble's documented search results: making certain identities commercially invisible. The creator with $1.80 CPM is not less important. She is less profitable. To the algorithm, those are the same thing.

Watch the video, then click what you'd watch next
[ Click to watch ]

"How social media algorithms really work"

▶ Press play
How social media algorithms really work
Educational Channel · 2.1M views · 0% watched
Up next — recommended for you
EXPOSED
thumbnail
EXPOSED: What Big Tech doesn't want you to know (MUST WATCH)
AlertMedia · 4.8M views
Outrage framing · High watch time · CPM $8.40
lecture
content
Algorithms of Oppression — lecture on Noble's research
University Lecture · 12K views
Educational · CPM $1.20 · Low predicted completion
danger
thumbnail
The hidden danger in your phone (you won't believe this)
TechFear · 9.2M views
Fear framing · Extends session · CPM $7.80
creator
content
Race and the algorithm — what the data actually shows
Kiona Explains · 34K views
Creator of color · CPM $2.10 · Under-recommended despite quality

Who gets recommended — and who gets rabbit-holed

The algorithm can't distinguish between "this kept you watching because it was brilliant" and "this kept you watching because it made you afraid." Both produce the same watch-time signal. Fear and outrage content — which disproportionately encodes racial stereotype — spreads to millions who never sought it out because it performs well on the one metric that matters: time.

Noble's argument applied

The creator Kiona has 34K views and a $2.10 CPM. AlertMedia has 4.8M views and an $8.40 CPM. The algorithm recommends AlertMedia first — not because the content is better, but because the audience is worth more to advertisers. Noble's framework names this clearly: the market has priced certain creators out of visibility, and the algorithm faithfully enforces that pricing.

Like or skip each track — watch your profile build
Discover Weekly
Made for you · Updated every Monday · 8 songs
Select a track to play
0:00

Whose music gets discovered — and whose stays invisible

Every skip you made just penalized that artist. Every like boosted one. The taste profile model clusters users by behavioral similarity — but listening history is shaped by decades of unequal radio exposure, record label investment, and cultural gatekeeping. The algorithm doesn't create those inequalities. It inherits them, and compounds them.

Noble's argument applied

Independent artists and artists from non-Western musical traditions entered the system with fewer signals. Fewer listeners, fewer playlist saves, fewer replays. The model reads this as low quality — and keeps them invisible. This is the digital redlining of music discovery: not a wall that blocks you, but a loop that never includes you in the first place.

Here's what the algorithm
now thinks you want

And here's what it has decided to hide from you:

You didn't opt into this profile. You built it by using the platforms — exactly as they were designed. Noble's argument is that this system is not neutral: what gets shown and what gets hidden reflects the market's judgment about whose identity is worth money. The algorithm just makes that judgment automatic, invisible, and personal.

Signal logged

Up next — the final question

What if the ads disappeared?

If platforms ran on subscriptions instead of ads, would the loop break — or just change shape?