Health score bias ?

Options
Diana De Jesus
Diana De Jesus Member Posts: 4 Seeker
edited October 2023 in CS Org Conversations

If you group your health score between three buckets "healthy, neutral, and at-risk", how likely are you to swing in one direction?

For example, leaning more towards putting customers in the at-risk bucket but then this creates a "the boy who cried wolf scenario" where you have too many customers who are "at-risk" but not really. Or staying away from "neutral" and either having customers fall in "healthy" or "at-risk".

What's your take on this?

Comments

  • Ed Powers
    Ed Powers Member Posts: 180 Expert
    Photogenic 5 Insightfuls First Anniversary 5 Likes
    edited November 2020
    Options
    Diana, whenever possible I recommend using objective analysis of data, rather than subjective assessments by CSMs, to set customer health dashboards. Statistical techniques such as regression can do this with a certain level of confidence without investing a lot more time and money on a predictive analytics solution. Absent enough  data, which may be the case in your example, a subjective assessment may be all you can do in the short term.
  • Thomas Hussenet
    Thomas Hussenet Member Posts: 9 Contributor
    edited November 2020
    Options
    Hello Diana
    I would advice to look also at the trend (not only the bucket) which might avoid some biases due to cultural or industry difference among your clients (or CSMs) and is a good indicator of health... trend which is also a key input for CSM actions.
    Cheers
    Thomas
  • Diana De Jesus
    Diana De Jesus Member Posts: 4 Seeker
    edited November 2020
    Options
    Thank you, Ed! Yes, you're right, we need a short term solution here.

    Do you have any data on regression? 

    ------------------------------
    Diana De Jesus
    Customer Success Manager
    ------------------------------
    -------------------------------------------
    Original Message:
    Sent: 11-12-2020 14:50
    From: Ed Powers
    Subject: Health score bias ?

    Diana, whenever possible I recommend using objective analysis of data, rather than subjective assessments by CSMs, to set customer health dashboards. Statistical techniques such as regression can do this with a certain level of confidence without investing a lot more time and money on a predictive analytics solution. Absent enough  data, which may be the case in your example, a subjective assessment may be all you can do in the short term.

    ------------------------------
    Ed Powers
    Consultant
    ------------------------------

    Original Message:
    Sent: 11-11-2020 15:58
    From: Diana De Jesus
    Subject: Health score bias ?

    If you group your health score between three buckets "healthy, neutral, and at-risk", how likely are you to swing in one direction?

    For example, leaning more towards putting customers in the at-risk bucket but then this creates a "the boy who cried wolf scenario" where you have too many customers who are "at-risk" but not really. Or staying away from "neutral" and either having customers fall in "healthy" or "at-risk".

    What's your take on this?



    ------------------------------
    Diana De Jesus
    Customer Success Manager
    ------------------------------
  • Diana De Jesus
    Diana De Jesus Member Posts: 4 Seeker
    edited November 2020
    Options
    That's a really great point, Thomas! I'll see what I can find around trends. I really appreciate the help.
  • Ed Powers
    Ed Powers Member Posts: 180 Expert
    Photogenic 5 Insightfuls First Anniversary 5 Likes
    edited November 2020
    Options
    Regression is a statistical technique to analyze data and construct mathematical models that can be used to predict future results. IT can be used for variable data (e.g. money, time) and categorical data (e.g. logo churn, risk categories). Some links to learn more: