In 2018, the first randomized controlled trial of Overcome Social Anxiety was published in the Journal of Medical Internet Research, run in partnership with researchers at the University of British Columbia . It compared the program against a wait list control condition among university undergraduates and found a between groups effect size (Cohen's d = 0.97), meaning the group using the program improved substantially more over the study period than the group that received nothing. That number sat close to the average effect size of computer aided interventions that include therapist support (0.96), despite Overcome Social Anxiety having none. A self guided program performing at that level, with no clinician on the other end, was not the expected result.
A trial like that one is necessary, but it has a known limitation. The people in it were recruited, screened, and often taking part as course credit, which is not how someone actually finds and uses a program in real life. That gap between a study population and real world use is what researchers call ecological validity, or external validity, and it's one of the harder things to establish in psychology research. You cannot randomize your way into it, you either have real world evidence or you don't.
A follow up analysis answered that question directly. It looked at real world users, people who found the program entirely on their own, with no recruitment, no course credit, and no researcher checking in. Among those who completed it, the effect size was even larger (Cohen's d = 1.76), evidence the results held up, and if anything strengthened, outside a research setting. That real world population turned out to be more international than we expected too, spanning 30 countries by 2019, a pattern of organic global reach we've written about separately.
Neither result was the finish line. They were the first two data points in what's now 13 published trials across three programs.