After some messing about, I got things working. Initially, the WFS210 wouldn't be recognized using USB. Windows wouldn't even indicate that something had been plugged in. I swapped the USB cable and tested the new one with another device before use. Still nothing. After trying all the possible permutations of installing the software, rebooting and powering up the device and starting the program, I decided to test yet another USB cable. This time I took one from a microcontroller development kit, so it should be a decent quality one. Suddenly, the oscilloscope shows up! It's either quite picky about the cable used, or I just got lucky this time. I did notice the quality cable snapped into the socket in a much more convincing fashion. After this, the WFS210 still didn't yield any traces, but a quick self calibration fixed that problem. It actually works!
First impressions aren't terrible. The hardware seems to be better made than feared. I see little wrong there. The software isn't as ugly as I feared it to be. However, functionality is very, very basic. Lots of data I would expect to be available is missing, or requires manually measuring it. As people have discovered, viewing frequencies at even one tenth of the specifications is a challenge. All in all, it so far seems to be a nice-to-have-toy. I don't think I would be happy if I paid $100+ for this. At this price, it's not terrible.
I just wish Velleman didn't lie so blatantly about the bandwidth. If waveforms don't turn out looking nice at higher frequencies, that's not really great. But the fact that the equipment can't even tell 720 KHz from 8 MHz means the specifications are completely fabricated.