Animal pollination is critical for wild plant reproduction and crop yield. However, not all flower visitors are necessarily pollinators. Using cocoa (Theobroma cacao L.) as a case study, we explored a combination of indicators to assess the effectiveness of flower visitors as pollinators. We first investigated the potential trade-off between the quality and quantity of pollen deposited on the limitation of pollination success in a hand pollination trial under controlled conditions. We found that the fruit set and ovule fecundation rate are mainly driven by the quantity of pollen grain deposited. We then conducted transects and video monitoring in the field in Ecuador to compare the performance of cocoa flower visitors as pollinators using several indicators. We found that small crawling arthropods, in particular ants and aphids, proved to be promising pollinators, with a higher level of pollination with viable pollen being observed on the flowers they visited. Ants were even more efficient than other visitors in depositing significant numbers of pollen grains on the styles. Instead, stingless bees and microdiptera were less efficient when depositing pollen, but complementary to small crawling arthropods in the time of flower visits. Interestingly, we found that others cocoa flower visitors were detrimental to the pollination service, such as hemipteran nymphs and caterpillars. We conclude that cocoa plantation management focused on increasing the complementarity of pollinator visits, acting at different periods of the day and with distinct behaviours, could improve pollination services, and that the use of multiple indicators should be advocated in future research to assess pollinator effectiveness.