Discussion about this post

User's avatar
Konstantinos Tsilimparis's avatar

That was a very nice read! Looking forward for more

Abhishek Pan's avatar

For the Strands Decider swap I'd also diff the label distribution week over week, alongside per-ticket shadow accuracy. When a new product line lands, a decision model keeps scoring fine on the old categories while quietly dumping the new ones into the nearest bucket.

Ready for more?