Background: Metastatic colorectal cancer remains among the most common causes for cancer death and is a major public health problem. While overall survival improvements have been achieved in clinical trials, it is unknown how much this progress translated into increased survival on a population-based level. Methods: The SEER (Surveillance Epidemiology and End Results) registry was queried from 1988 to 2008. Patients with metastatic colorectal cancer were included. Multivariable adjusted Cox proportional hazards regression analyses were used. Results: In total, 89,543 patients were included. Median age was 67.1 years, 47.2% were female, 79.7% Caucasian. Median overall survival was 11 months, which increased from 8 months in 1988 to 14 months in 2008 (75% increase). In patients <= 65 years, overall survival increased from 9 months in 1988 to 20 months in 2008 (122% increase), in patients above 65 years from 7 months in 1988 to 9 months in 2008 (29% increase). In Cox proportional analyses, later year of treatment (hazard ratio=HR: 0.973 per year [95% CI: 0.972-0.975], p<0.001), tumor location in the rectum (HR: 0.72 [95% CI: 0.71 - 0.74]; p<0.001), married civil status (HR: 0.85, [95% CI: 0.84 - 0.86]; p<0.001), and cancer-directed surgery (HR: 0.44, [95% CI: 0.43 – 0.45]; p<0.001) were associated with better overall survival, while patients above 65 years (HR: 1.48, [95% CI: 1.46 - 1.51]; p<0.001) and African-Americans (HR: 1.08, [95%CI: 1.06 - 1.11], p<0.001) had worse survival. Conclusions: Tumor location in the rectum, married civil status and cancer-directed surgery are associated with significantly improved overall survival while elderly and African-Americans have worse outcomes. Moreover, a statistically significant and clinically relevant overall survival increase from 1988 to 2008 was found for metastatic colorectal cancer patients. However, most progress is observed in patients aged below 65 years, while the overall survival increase in the elderly is modest. Further studies among elderly patients assessing the reasons for the lack of a parallel survival increase are warranted.