Prior studies have demonstrated an association between excision repair cross-complementation group 1 (ERCC1) expression level and outcomes in patients with advanced non-small cell lung cancer (NSCLC) treated with platinum-based chemotherapy. The aim of this study was to assess the impact of ERCC1 on survival for patients with stage IIIB/IV non-squamous NSCLC (NS-NSCLC) enrolled in the INNOVATIONS trial, thus receiving as treatment either erlotinib/bevacizumab (EB) or cisplatin/gemcitabine/bevacizumab (PGB). We retrospectively analyzed tumor tissue of 72 patients using immunohistochemistry to assess the expression of ERCC1. The distribution between treatment arms was equal (36 patients each). Two different H scores were calculated and correlated with survival. In ERCC1-positive patients, no significant difference in terms of progression-free survival (PFS) between treatment arms has been detected. ERCC1-negative patients benefited from PGB compared to EB arm ( H score: HR = 0.377, 95% CI 0.167–0.849, p  = 0.0151; modified H score: HR = 0.484, 95% CI 0.234–1.004, p  = 0.0468). With respect to the scoring system, in the EB-arm, a significant superior PFS turned out in ERCC1-positive patients when employing the H -score (HR = 0.430, 95% CI 0.188–0.981, p  = 0.0397; median 4.9 vs. 3.9 months), but not with the modified H -score. Our findings support the hypothesis that NS-NSCLC displaying a low ERCC1 expression might benefit from cisplatin-based chemotherapy. High expression indicated better PFS in the EB arm supporting the prognostic impact. However, as impact of ERCC1-assessment even might depend on scoring systems differences, the need in standardization of assessment methodology is emphasized.