Abstract: Recommender systems today face major challenges in keeping up with dynamic customer preferences. Disruptions or sudden changes in the environment affect customer preferences drastically and render historical data ineffective for modeling. With businesses relying heavily on Machine Learning(ML) based recommender systems for catering to customer preferences, the accuracy of timely recommendations gains prime significance. To address these challenges, we propose a novel concept, LDT (Labeled Data Threshold), a newly defined parameter to determine the sufficiency of available labeled training data. Our proposed scheme, using LDT leads to a significant reduction (  50X) in the training time for a model, thus enabling recommender systems to adapt quickly to disruptions. We illustrate the efficacy of our proposed scheme, by conducting extensive experimental analysis on six well known, structured data sets from various public domains.