Filtration is the most widespread stabilisation operation for extra virgin olive oil, preventing microbial and enzymatic changes. However, during the harvest, the workload of olive mills is at its peak. This results in two approaches to filtration: (i) delays it until after harvesting, increasing the risk of degraded oil quality, and (ii) filters it immediately, increasing the workload. The aim of our experiment is to assess the risk of delaying filtration and establish a safe delay time. Changes in the sensory profile and volatile compound contents were evaluated during 30 days in filtered and unfiltered samples. Significant differences were related to filtration: both turbidity grade and microbial contamination; no differences for the legal parameters were found. Two, contrasting, results were obtained with respect to oil quality: (i) the fusty defect, appearing in less than five days in unfiltered oils, leading to the downgrade of the oil’s commercial category, and (ii) filtration removing some lipoxygenase volatile compounds. Consequently, a fruity attribute was more pronounced in unfiltered samples until day five of storage; it seems that, from this point, the fusty defect masked a fruity attribute. Hence, filtering within a few days strongly reduced the risk of degraded oil quality compared to a delayed filtration.