I hit create too soon... so the measurement is actually how the scores in the treatment group compare to the placebo, which is essentially what you are saying but this approach allows for a measurement even if both groups improve.

If treatment improves faster, or more significantly than placebo group (on average) there is still something to measure.