Model Drift
Definition
Model drift refers to relevant changes compared to the expected operational baseline of a model.
A relevant change in inputs, relationships, or performance of a production model compared to the expected operational baseline.
Simply explained
Like a practiced plan that no longer fits the changed environment; individual fluctuations alone are not sufficient for diagnosis.
Why it matters in practice
Helps realistically assess model quality, limits, and technical decisions.
Example from business
The error rate of a demand forecast increases after assortment and market changes.
Limits and pitfalls
- Reporting every fluctuation as drift or observing only input drift without performance reference.
Origin & history
Drift monitoring has become a core task of model risk management with productive ML systems.
Use in mid-sized companies
- Possible application: Business and operations teams monitor inputs and performance with defined thresholds. In case of anomalies, they check data and environment changes before deciding on adjustment or retraining.
Frequently asked questions
What does model drift mean?
A relevant change in inputs, relationships, or performance of a production model compared to the expected operational baseline.
Which mistakes should be avoided?
Reporting every fluctuation as drift or observing only input drift without performance reference.
Sources
Source checked on 10.10.2026. Editorial use cases are hypothetical; no evidence of real customer results.
Editorially reviewed on: