If an app has a new feature that is to be released after internal testing, canary method decreases the risk of impact of unpredicted behaviour (hidden bug or misconfiguration). When the new feature is confirmed to work well, the feature may give different performance depending on parameters of that feature. To find out which set of parameters gives better performance we use A/B testing.

A typicall example in e-commerce is releasing new web page with new campaign of a product (canary deployment), then rearanging images, links or even modifying the contents of an advertisement on that web page to find out how viewers respond to the campaign (A/B testing).