CSM Relationship Score Best Practices

Hello @donnamcnamara 

The use case here is  more along the lines on quantifying CSM sentiment or Client relationship measure which is usually a Manual measure. The general recommendation here can be, having 3 levels(limited) of scoring on this measure only, like green yellow or Red instead of putting a score which can 85 or 95 for green. If you have three levels (best practice)or 5 (at max) levels it would be easy to define internally what they stand for and how we evaluate a customer. A risky customer may fall in Red, a non engaging or customer getting less value falls in Yellow and engaging and happy customer may get Green. 

This again varies company to company. I am just putting out my thoughts on this.

Like @aktripathi , I’ve found a best practice for customer sentiment is not to use the 0-100 scale, because it can descend into unproductive questions like “Is this customer a 33, or a 35?” and to your point, can lead to some inconsistency. Simplicity leads to value here, and having buckets of Red, Yellow, Green or A, B, C, D, F is typically sufficient, and sometimes actually more clear and actionable than a 0-100 range. However, it can be hard to transition off a 0-100 scale, so I wouldn’t sweat it if you’re saddled with it.

I have seen plenty of off-line matrices which describe which customers should slot into which scores. They often come in an Excel- or Sheets-style format, with columns like Score, Description, Examples, Likely Next Steps. A key here is enabling CSMs to see examples or stories of customers with certain scores. Expect it to be rocky to start, and for the CSM teams to get better at this over time, as they develop the collective muscle of sentiment scoring across a team.

Another area to watch on manually-set sentiment is stale data. The most successful CS teams ensure there’s a solid process, workflow or check/balance that helps CSMs regularly input and update the scores. It’s a waste of time, and erodes tool confidence, to find outdated or obviously incorrect manually-set sentiment scores. Some CS leaders put a validity time on this (at minimum), a report with last score updated information (in the middle) or a rule to clear stale scores (for the more aggressive) to ensure a score isn’t outdated or forgotten.

And lastly, if you do use manual sentiment scoring, I strongly promote a culture of transparency. A manually set score should contribute to communication and customer health, not to a “gaming” world in which the CSM has to hide the true story to attain some ill-conceived KPI metric. The last scenario you want is a CSM coming under undue fire for being honest and truthful about customer sentiment.

Hi @matthew_lind - Great information, super helpful. I am pretty much saddled with a 0-100 scale and was trying to find some benchmarks around examples of customers with certain scores. You mention having seen some off line matrices - any chance you have one/some you can share or can point me in a direction to find one? I’m getting questions from our newer folks on what the scores mean and seeing a bit of “gaming” from seasoned folks who don’t want their overall scores to fall below a certain threshold. Lots of work ahead of me on that front but hoping I can pull something together with examples to leverage of where to set the relationship scores appropriately. TIA!

@donnamcnamara I’m glad this is helpful.

I don’t have my hands on an off-line customer sentiment matrix just now. However, this Gitlab site has some great examples and some significant insight into how some of their functionality works.

https://about.gitlab.com/handbook/customer-success/tam/health-score-triage/#health-assessment-guidelines