Nobody advertises the number that matters

Every broadband advertisement is a download figure. Six megabits, ten, twenty if the fibre has reached you. Nobody puts the other number on the poster, and the other number is the one that is about to matter.

A typical cable connection gives you several megabits down and a few hundred kilobits up. The ratio is not an accident or an oversight. It was designed in, on an assumption that has quietly stopped being true.

The assumption is that a household consumes. You request pages and they arrive, you request video and it arrives, and what travels the other way is a handful of clicks and the occasional email. Under that model, allocating almost all the capacity downstream is simply good engineering. There is a finite amount of spectrum on the cable and it should go where the traffic is.

Now look at what people are actually doing.

Uploading video to a sharing site. Putting a few hundred photographs somewhere the family can see them. Making calls over the internet, where the upstream is what carries your voice and where a saturated upstream is why you sound like you are underwater. Backing up a machine to a remote service, which involves pushing tens of gigabytes through a pipe that was sized for clicks.

Every one of those is a person producing rather than receiving, and every one of them runs into the same wall. A connection that feels quick when you are reading feels broken the moment you try to contribute anything.

There is a technical constraint here and I do not want to pretend otherwise. Rebalancing a cable plant toward upstream means taking spectrum from downstream, and downstream is what the customer was sold and what gets measured in reviews.

But it is worth noticing what upstream capacity would enable. It is the thing you need to run a service from your house rather than subscribe to one. Symmetric connections at consumer prices would mean a household could host, and hosting is a business the people selling you the line are also in, or would like to be.

I am not claiming the ratio was chosen for that reason. I am saying that when the engineering constraint and the commercial interest point the same way, the constraint tends to get discussed a great deal more than the interest, and nobody has to decide anything for the outcome to hold.