Right, you're all quick to blame the ISP, but it might not be (directly) their fault.
The ISPs only run small networks with links generally to the large hubs such as Telehouse and Telecity, and rely on larger transit providers, such as Level 3, from there on.
The transit providers tend to have resilient links, but as with all IT, generally relies on all the equipment cooperating properly. As always with IT, under failure conditions, reality doesn't always match the test conditions when the infrastructure was designed and tested. Been there, done that. Enjoyed it so much, its almost a weekly pleasure

Entwood - if the names are resolving to IPs, that (almost) rules out DNS issues. You need to study the traceroutes to see if they are all going via the same transit, and stopping at same point.
Additionally, remember virtually all infrastructure routers will prioritise real traffic over icmp echo-request and icmp echo-reply, making ping and tracert less useful than some people believe.