custom_label_4

`custom_label_4` is conventionally the experiment slot — reserved for A/B tests, short-term tactical campaigns, and anything you want to be able to clean up without disturbing the stable structure built on labels 0-3. Treating one slot as the "scratch space" keeps the rest of your segmentation discipline intact when you need to try something quickly.

Practical examples: testing a new product set definition before promoting it to a permanent label; tagging a creative-test cohort during a campaign experiment; flagging items that need manual review without removing them from active serving. The value is that experiments don't pollute your stable segmentation — when the test ends, the label clears, and your label 0-3 hierarchy is untouched.

Some teams skip this convention and use all five labels for stable dimensions. That works for catalogs with five orthogonal segmentation needs, which is rare. More common is three or four stable dimensions plus the occasional need for tactical labelling — which is what custom_label_4 makes clean.