Pollination ecosystem service studies across multiple crops in many countries have shown that managed pollinators often complement wild pollinators. Such studies typically compare farms with or without access to wild pollinators, without controlling for the introduction of honey bee hives. Here, we compared three pollination strategies in commercial apple orchards in the Cape Floristic Region (CFR) of South Africa: (i) orchards relying exclusively on managed pollination through the introduction of hives of the endemic Cape honey bee (Apis mellifera capensis); (ii) orchards relying solely on pollination from all wild indigenous pollinators, including the Cape honey bee; and (iii) orchards using a combination of both sources. We found that pollination across all three pollination options was almost entirely from the Cape honey bee, but with significantly fewer flower visits in wild orchards. Wild pollinated orchards produced significantly fewer seeds per apple compared to orchards with hived honey bees, resulting in fewer apples of marketable size or shape. Hived bees resulted in the highest proportion of high-quality fruit. Summed across all farms, excluding insect pollination led to a ~95% reduction in the number of apples per tree, while in terms of apple marketability, failure to rent hives at 2 hives per hectare resulted in a 17.4-percentage-point reduction in high-quality fruit. Use of hived native honey bees in the CFR should therefore be considered an essential management input and an insurance measure for maintaining high-quality apple production. Our study emphasizes why local context should take precedence over broad international policy recommendations for optimal management of pollination services.