ProblemEvaluation of influenza surveillance systems is poor, especially in Africa.ApproachIn 2007, the Institut Pasteur de Madagascar and the Malagasy Ministry of Public Health implemented a countrywide system for the prospective syndromic and virological surveillance of influenza-like illnesses. In assessing this system’s performance, we identified gaps and ways to promote the best use of resources. We investigated acceptability, data quality, flexibility, representativeness, simplicity, stability, timeliness and usefulness and developed qualitative and/or quantitative indicators for each of these attributes.Local settingUntil 2007, the influenza surveillance system in Madagascar was only operational in Antananarivo and the observations made could not be extrapolated to the entire country.Relevant changesBy 2014, the system covered 34 sentinel sites across the country. At 12 sites, nasopharyngeal and/or oropharyngeal samples were collected and tested for influenza virus. Between 2009 and 2014, 177 718 fever cases were detected, 25 809 (14.5%) of these fever cases were classified as cases of influenza-like illness. Of the 9192 samples from patients with influenza-like illness that were tested for influenza viruses, 3573 (38.9%) tested positive. Data quality for all evaluated indicators was categorized as above 90% and the system also appeared to be strong in terms of its acceptability, simplicity and stability. However, sample collection needed improvement.Lessons learntThe influenza surveillance system in Madagascar performed well and provided reliable and timely data for public health interventions. Given its flexibility and overall moderate cost, this system may become a useful platform for syndromic and laboratory-based surveillance in other low-resource settings.