The article discusses a method to figure out lost sales during a disruption by analyzing time series data with seasonal adjustments. By creating optimized seasonal indexes and using quarterly time series models, the researchers found that estimating lost sales is more accurate and less variable compared to using monthly models. This approach helps account for seasonal trends and makes the estimation process simpler.