HER2 testing in metastatic gastric or gastroesophageal junction cancer (mGC/mGEJC) is standard practice. Variations in HER2-positivity rates suggest factors affecting test quality; however, the influence of patient-, tumor-, and laboratory-related factors on HER2-positivity rates remains unknown. This observational, prospective study collected routine HER2 testing data from 50 pathology centers in Germany (January 2013–December 2015). For each sample, HER2 status, primary tumor location, method of sample retrieval, and other patient- and tumor-related parameters were recorded. A model for predicting the probability of HER2-positivity was developed using stepwise multiple logistic regression to identify influencing factors. Documented positivity rates and corresponding predicted HER2-positivity probabilities were compared to identify institutes with deviations in HER2-positivity. Data from 2761 mGC/mGEJC routine diagnostic specimens included 2033 with HER2 test results (1554 mGC, 479 mGEJC); overall HER2-positivity rates across centers were 19.8% and 30.5%, respectively. HER2-positivity correlated most with Lauren classification, then HER2 testing rate, primary tumor location, sample type, and testing method (all p  < 0.05). Three institutes had model-predicted HER2-positivity rates outside the 95% confidence interval of their documented rate, which could not be explained by sample and center characteristics. Results demonstrated the high quality of routine HER2 testing in the mGC/mGEJC cohort analyzed. This is the first study investigating parameters impacting on HER2-positivity rates in mGC/mGEJC in routine practice and suggests that assessment of HER2 testing quality should consider primary tumor location, testing method and rate, and tumor characteristics. Accurate identification of patients with HER2-positive mGC/mGEJC is essential for appropriate use of HER2-targeted therapies.