All the speed-reading tests I've tried seem to have the same problem. There's nothing stopping you from just clicking through the text really fast and claiming you can read 900 words per minute. It sounds impressive, sure, but it doesn't actually mean anything.

So, I decided to build my own, but with a comprehension check built in. You read a passage, then answer some questions about it. If your score is at or around what you'd get just by guessing, you don't get a score. Instead, you get an explanation that there's no real way to know if you actually read it.

There were a couple of outcomes I didn't really anticipate. First, some people fail the test and get upset. I guess you could say that's actually the tool doing its job. Second, it made the results more useful, not less. Since comprehension is tracked as the speed increases, the test can tell you "you were still understanding the material at 332 wpm," instead of just giving you a theoretical top speed. That's a number you can actually do something with.

The thing I keep wrestling with is this: is it good product design to refuse to give someone a number they're looking for, or am I just annoying people and making them leave?