What are some of the *worst* KPIs to commit to achieving?

I've found two KPIs to be difficult to commit to:

1. Customer Health. If you have a robust algorithm to measure customer health (influenced by a number of inputs ), it can be hard commit to a certain outcome. To frame this another way, I've often observed customer health scores as being a bit of a black box where it's hard to tie the actions you take to specific outcomes when there could be a number of variables outside of your control that influence the ultimate score. I much prefer to commit to lead measures that are directly within the control of the team. KPIs related to customer engagement are a good example of things that are more directly within the team's control.  
2. Upgrade rate. Many CSM teams are measured on Net Revenue Retention. As part of this, your CSMs may be responsible for identifying growth opportunities within the install base of customers. I find it's effective to measure the team on how many growth opportunities the team identifies but not the close rate or upgrade rate, especially if the Sales or Account Management team owns the closing motion. 

Committing to certain KPIs in customer success without considering their broader impact or relevance to overall business objectives can lead to misguided efforts and low outcomes.

The metrics needs to be aligned in every phase of customer lifecycle, reviewed and alter accordingly to the goals of the organisation.

Example:

1. Attain 0 Churns - GRR is important but it doesn't mean that you can over emphasize on 100% GRR, instead take it slow and have a projection considering all aspects, be it macro or any as such.

2. The same with exemplary NPS scores, expansions too. Have it fair in the business.

Do not be in a rush to achieve the impossible. Always have a structured process and a significant increase quarter by quarter instead.

It's essential to prioritize metrics that truly reflect customer value, satisfaction, and loyalty, while also driving sustainable business growth and profitability.