An experiment is randomized when you use a random selection process to choose which people receive the active treatment, and which receive placebo. This way, you cannot bias the results by selecting certain people to receive the treatment whom you secretly believe might respond better.