They could just embed tracking code to the streaming service? As in: count how many times the chunk of video was sent to the clients, rather than relying on the clients to work as THEY intended...

Client-side analytics must end

▲

danhau 4 days ago | parent | next [-]

This would make replays or scrubbing count as additional views. To fix that, they would need some kind of set to uniquely store all clients, and that‘s questionable from a security and moral point of view, even for YouTube.

▲

dec0dedab0de 4 days ago | parent | next [-]

local cache should handle scrubbing

▲

rasz 4 days ago | parent [-]

No such thing for YT videos. Official player will refetch video chunks if you so much as rewind 5 minutes back.

▲

inexcf 4 days ago | parent [-]

And that is incredibly annoying for the user and a problem Youtube should fix.

▲

brookst 4 days ago | parent [-]

If YouTube stored the entire video in a cache people would yell and scream about that. Oh, I’ve got 2TB of YouTube cache that didn’t get cleaned properly, how annoying.

▲

rasz 4 days ago | parent [-]

Well Im a liar. Checked just now and it changed since last time I was looking into this.

cache-control private, max-age=11722 (~3 hours) date Thu, 18 Sep 2025 14:19:15 GMT expires Thu, 18 Sep 2025 14:19:15 GMT

it once again lands in browser cache. I remember a moment when it returned no-cache.

We are back to situation where:

- google doesnt get any info if user with adblocker keeps rewinding in that ~3hour window

- player refetches if you pause for few hours and come back, or decide to rewind 3 hour video to watch again

- your SSD is hammered with gigabytes of useless browser cache writes - might be good idea for Extension overwriting those headers to no-store/max-age=0

▲

brookst 4 days ago | parent [-]

I would be surprised if browsers actually cashed the entirety of videos, even if the cash policy allows for it. That does seem like a way to thrash SSD.

	▲	rasz 4 days ago \| parent [-]
		They did before switch to no-cache, and I bet they are back at it now. Chrome used to roughly write as much as I watched at ~2-3GB per hour.

▲

thaumasiotes 4 days ago | parent | prev [-]

What? Replays already do count as additional views. Load a video one day, then load it again the next day. That's two views. There isn't a way to avoid this non-problem.

I'm not sure what you mean by "scrubbing".

▲

injidup 4 days ago | parent [-]

Scrubbing is a video editing term for going forwards and backwards over a video to find specific frames and editing them

	▲	rasz 4 days ago \| parent [-]
		and YT 'multiple times throughout a video playback' client side endpoint has been tracking this for years reporting every single minute of video you watched, thats what is powering Most Replayed Feature (scroll bar graph showing popular moments in every video)

▲

iamacyborg 4 days ago | parent | prev | next [-]

But then they’d have to report significantly lower CPM’s to content creators.

	▲	brookst 4 days ago \| parent [-]
		Lower CPMs, but it would be so easy to game that creators would all have trillions of views.

▲

cykros 4 days ago | parent | prev | next [-]

Indeed. It's essentially malware.

At least BonziBuddy sang for me.

	▲	muyuu 4 days ago \| parent [-]
		ah that takes me back, going to the Uni computers and have all them ridden with malware and browser bars so thick you could barely browse the net but still you could go home and have a reasonable setup, there is no escape from the current "open" interwebs

▲

tpxl 4 days ago | parent | prev | next [-]

They already do something like this - some videos have an indicator for how many times a chunk of a video is played.

▲

cm2187 4 days ago | parent | prev [-]

Plus that would make cheating on traffic really bandwidth expensive

	▲	brookst 4 days ago \| parent [-]
		Not for the cheater. You’d still buy 1m views on some shady site, armies of bots on hacked devices/routers would still pull down the steams at no cost to the bad guys.