Blog Viewer

Correlating NPS to CLV

By Crystal Montague, CCXP posted 05-18-2023 03:02 PM


Hello All!

I am working on a project correlating historical NPS performance to Customer Lifetime Value (CLV). I am not getting a clear correlation (although we know NPS relates to increased revenue). Has anyone had any success in demonstrating this correlation, and if so, how?




06-07-2023 10:58 AM

Hi @Crystal Montague, CCXP . I look forward to it m

05-31-2023 04:24 PM

Thank you Michelle for your input. So far we have done some regression analysis and are not seeing huge differences between the NPS groups. I am doing research to see if other metrics, i.e. likely to renew have a stronger correlation. The goal of this project is to see if NPS can predicting business growth. Happy to share my results once I get them :) 

05-31-2023 04:20 PM

Thank you Valerie for you input. 

05-22-2023 08:26 AM

Another interesting correlation is doing a shrinking/growing study and compare that to NPS.  Just either pick your top 10% or a statistical sample of who you think your best customers are.  Always fascinating to see if those 'best customers' are growing or shrinking even though their NPS might be high.    I have found little correlation between NPS or CSAT and LTV or repurchase.  We chase the wrong metrics in CX I fear.

05-19-2023 01:37 AM

Hi Crystal, it is a great idea to understand the relationship between NPS and Customer Lifetime Value. But..

At the risk of sounding curmudgeonly, we don't know how the correlation between NPS and revenue was demonstrated, let alone causation. My suspicion is that your signal has too much noise.

What statistical test have you use to check for correlation? What was the actual answer? What other influences on CLV have you tested? Might you pick a different measure to test against NPS? Revenue, perhaps repeat purchases, upsell and cross-sell or churn may have fewer other influences.

And why are you doing this? It sounds intriguing and a great thing to learn, but how will it benefit you? Knowing this could help you design a neater test.