For example, say I'm paying for a connection with a 20 Mbps upload speed but when I upload to backblaze my upload only runs at 10 Mbps. The problem could be my hard disk, filesystem, CPU, software, wifi, router/modem, ISP, or backblaze themselves.
Currently we don't do a particularly good job of giving nontechnical users the information they need to make rational decisions about what they need to change or upgrade.
One of the ISPs would charge lower prices while offering crappier service. Some customers would still stick with it because it's cheaper. That ISP then strikes a deal with some smaller websites to make them faster than their competitors. Then the ISP offers a sweetheart deal to the larger competitors (at first) which they'll take to avoid losing ground, so then everybody is paying them for faster service.
Now that ISP has the lowest prices and their service is suddenly better because it's being subsidized by the websites, so they get more customers. Then they use the leverage of more customers to negotiate higher prices from the websites ad infinitum. The competitive advantage forces other websites to adopt the same business model.
It isn't a problem that competition solves because the people who have to pay can't choose between competing ISPs.
There are two responses to this. The first is OK do it and then we'll get rid of the regulations as soon as there actually are 5 or more competitive ISPs.
But the more important point is that it doesn't actually solve the problem in question. There could be 10 ISPs that every consumer could choose from but every website would still have to pay all of them if they were allowed to charge websites, because the site can't get faster access to Verizon's customers by paying AT&T or vice versa.