O'Reilly logo
  • Asad Quraishi thinks this is interesting:

In particular, the need to scan a vast number of potential neighbors makes it impossible to compute predictions in real time. Large-scale e-commerce sites, therefore, often implement a different technique, item-based recommendation, which is more apt for offline preprocessing1 and thus allows for the computation of recommendations in real time even for a very large rating matrix

From

Cover of Recommender Systems

Note

item-based recommendation better for real-time prediction when there are millions of users