Electron transport layers (ETLs) play a fundamental role in perovskite solar cells (PSCs) through charge extraction.Here, we developed flexible PSCs on 12 different kinds of ETLs based on SnO 2 . We show that ETLs need to be specifically developed for plastic substrates in order to attain 15% efficient flexible cells. Recipes developed for glass substrates do not typically transfer directly. Among all the ETLs, ZnO/SnO 2 double layers delivered the highest average power conversion efficiency of 14.6% (best cell 14.8%), 39% higher than that of flexible cells of the same batch based on SnO 2 -only ETLs. However, the cells with a single ETL made of SnO 2 nanoparticles were found to be more stable as well as more efficient and reproducible than SnO 2 formed from a liquid precursor (SnO 2 -LP). We aimed at increasing the understanding of what makes a good ETL on polyethylene terephthalate (PET) substrates. More so than ensuring electron transport (as seen from on-current and series resistance analysis), delivering high shunt resistances (R SH ) and lower recombination currents (I off ) is key to obtain high efficiency. In fact, R SH of PSCs fabricated on glass was twice as large, and I off was 76% lower in relative terms, on average, than those on PET, indicating considerably better blocking behavior of ETLs on glass, which to a large extent explains the differences in average PCE (+29% in relative terms for glass vs PET) between these two types of devices. Importantly, we also found a clear trend for all ETLs and for different substrates between the wetting behavior of each surface and the final performance of the device, with efficiencies increasing with lower contact angles (ranging between ∼50 and 80°). Better wetting, with average contact angles being lower by 25% on glass versus PET, was conducive to delivering higher-quality layers and interfaces. This cognizance can help further optimize flexible devices and close the efficiency gap that still exists with their glass counterparts.