While many applications do scale things to multiple threads, a lot don't. VPN tunnelled traffic would be one such example (there are VPNs that support multiple threads, but they're not the most common). Frequently software updates won't, neither will video streaming, etc.
But here it's being used as a ready-reckoner gauge for how likely a provider network is to be well-implemented. If they've got high single threaded performance then it's unlikely to be the weak link. If single threaded performance is lower then you might expect there are bottlenecks somewhere in the network that you're connected to, which might or might not become a factor in your usage.
My main reason for having a 900Mbps is not because I want to push >200 terabytes a month down a piece of glass. In terms of the amount of data we use in this house; I could get away with an 80Mbps connection. It's because I'm wanting to pay not to have to wait even a second more for things to happen than I have to. Software updates, video streams starting up, large downloads, etc. In the case of the Zen GEA connection I've noted that I often have to scale to an obscene number of connections to achieve close to line rate for a given download (eg the TBB large files). So even allowing for a web browser applying some multithreading, it's still much slower than you'd expect.
Given that with the exception of a couple of notable outliers that pricing is broadly similar for FTTP900 providers, if there is opportunity to have a better service by understanding the behaviour of the provider's networks then it's a consumer win, and if awareness of the performance of provider networks stirs providers into looking after their bits a little more, then that's also a win.
It's hard to see there being any downside other than the cost in my own time / effort to investigate.
