This Australian human epidermal growth factor receptor 2 (HER2) testing program aimed to analyse >800 cases tested in a coordinated setting to further evaluate the criteria to establish HER2 status in advanced gastric and gastro-oesophageal junction (GOJ) cancer. Heterogeneity, and minimum number of biopsy fragments for reliable HER2 assessment were also examined in a subset of samples. Five laboratories tested 891 samples referred to determine HER2 status for potential anti-HER2 treatment. Cancer site, specimen type (endoscopic biopsy/resection/metastases), immunohistochemistry (IHC) score, HER2 gene and CEP17 copy number (CN) and HER2:CEP17 ratios were recorded. Samples were derived from stomach (53.1%), GOJ (28.2%) or metastases (18.5%). IHC for HER2 and dual probe HER2:CEP17 in situ hybridisation (ISH) were performed in parallel. A stringent definition (SD) of HER2 positivity was used (IHC2+/3+ plus CN>6 and ratio>2) and compared with other published criteria. HER2 positive rate was 13.9% (114/820) by SD, and 12.9–16.0% using other definitions. There was higher concordance between IHC and HER2 CN by ISH than with ratio. The HER2 positive rate was significantly higher in GOJ samples than others (p = 0.03) and in endoscopic biopsies than resections (p = 0.047). In a subset of 98 positive cases, 39 (39.8%) showed heterogeneity, and in 282 endoscopic biopsies positivity rate plateaued at five tumour fragments, suggesting this is the minimum number of biopsies that should be examined.