Shared networks can work remarkably well. Across the U.S., AT&T, Comcast, and Verizon are in a fierce marketing battle to offer fiber to new developments. Table stakes now are a reliable gigabit network and premium TV offerings. 400,000 new apartments go up every year. They add less than 1/2 of 1% to the housing stock but are a prime growth market.
I suspect there's a slight exaggeration here, but 97-99% would make most of us very happy. U.S. & U.K. government testing (SamKnows) have long demonstrated that most cable networks are darn close to 100% delivery. In 2014, FCC tests showed 95% of Comcast customers received between 109% and 119% of advertised upload speeds.
All large carriers have fiber-based networks. Once you have the network built, the cost to add extra data capacity is usually very small. Nearly all cable networks I know are provisioned with enough reserve capacity to very rarely fall significantly below the promised speeds. Long DSL lines can fall far below the promised speed, but should be reasonably consistent peak and offpeak.
Telco claim their service is much better because it's not shared. That's bogus. Most of the perceived problem with surfing at peak is with the server, not the network. Even when there is a congestion issue, companies like Comcast have built systems to deal with it with minimal disruption. For examples, customers pulling 50 megabits down might be reduced by 1/3rd and they test every 15 minutes to see if the constraint can be lifted. Even if your speed is down to 30 meg, you can, for example, watch 5 HD programs simultaneously.
I've seen data for most of the large cablecos in England, the U.S., and Canada. I believe most of the developed world is similar but don't have data. A few years ago, SamKnows data found that Cablevision could fall by half. When that became public, they fixed their network problems and did fine the next year. Small cablecos, like small telcos, are subscale and pay so much more for backhaul their service often is compromised, especially in rural areas.