A product goes viral on social media, creating high demand. Distribution channels are facing supply chain issues because the testing and training models that are used for sales forecasting have not…

Model bias refers to systematic errors in the model's predictions, often due to unrepresentative training data or flawed assumptions, but not necessarily a change in data characteristics after deployment.

Data drift occurs when the distribution of the input data or the relationship between the input data and the output variable changes significantly over time, making the existing model less accurate. In this scenario, the "viral" demand represents a new data pattern that the sales forecasting models, trained on past data, have not encountered, thus leading to inaccurate predictions due to the shift in data characteristics.

Incorrect sizing generally refers to resource allocation (e.g., server capacity) and is not a statistical concept related to model accuracy due to changing data patterns.

Skewing refers to an asymmetry in a statistical distribution, which can be part of data drift but doesn't fully encompass the concept of the model encountering fundamentally new data patterns it wasn't trained on.