Measuring CSM Performance: Beyond NPS

The question of how to measure CSM performance is one most CS leaders find uncomfortable. Not because there's no answer, but because the answer requires being honest about a gap that's been papered over for a long time.

Most CS teams measure CSM performance on a combination of NPS or CSAT scores, renewal rates, and a handful of activity metrics. The renewal rate tends to dominate. And the underlying logic is clean enough: CSMs are responsible for retaining customers, so if renewals are happening, CSMs are doing their job.

The problem is that a CSM's influence on renewal is real but partial. It's mediated by product quality, pricing, sales promises made during the deal, the customer's own business performance, and factors the CSM had no role in and no control over. Holding CSMs accountable for renewal rates without controlling for these factors is a bit like holding a ship's first mate accountable for whether the ship arrives on schedule, without accounting for the weather.

The result is a performance framework that creates the wrong incentives: CSMs focused on managing the renewal outcome rather than on the upstream activities that drive it, and that makes it genuinely difficult to distinguish between a CSM who is excellent and was given a difficult book, and one who is average and got lucky.

---

What CSMs actually control

A useful CSM performance framework starts from the question: what can a CSM actually influence? Not what outcomes are they nominally responsible for, but what activities, decisions, and relationship investments are genuinely within their control and have a causal relationship with the things that matter?

Four areas stand out.

Engagement quality and depth. Does the CSM build genuine relationships across the customer organisation, or only with the champion? Are they having conversations that surface useful information, about the customer's goals, their business pressures, the internal conversation about the product? Are they multi-threaded, or single-threaded? This is qualitative, but it can be operationalised: relationship map depth, number of meaningful contacts in an account, evidence of conversations with stakeholders beyond the primary contact.

Outcome delivery. Is the CSM actively driving customers toward their agreed success milestones, or managing the relationship without it? This requires the value milestone framework to exist first, but once it does, it's possible to track whether CSMs are moving accounts toward outcome achievement rather than just maintaining engagement.

Risk identification. How early does the CSM surface account risk? A CSM who consistently flags accounts before they show up in the health score as red is more valuable than one who reacts after the fact. Early risk identification is the kind of thing that only shows up in performance data if you're deliberately tracking the gap between when a CSM flags something and when the system would have caught it.

Playbook adherence and quality. Is the CSM running the plays the team has agreed on, at the right moments? Are the outcomes from those plays (the executive business review that generated a clear next step, the recovery conversation that turned an amber account) being documented in a way that's useful? CSMs who adapt the playbook intelligently and document it well make the whole team better over time.

---

Renewal rate as context, not as score

Renewal rate should not be removed from CSM performance measurement. But it should be contextualised.

The most useful framing is to separate the renewal into components: what was within the CSM's sphere of influence, and what wasn't? If an account churned because the product didn't deliver on a commitment made in the sale, that's a signal about the sales-to-CS handover, not about the CSM's effectiveness. If an account churned because the customer's business had a structural change that made the product irrelevant, that's not a performance issue at all.

Some CS teams use a qualitative classification alongside renewal data (flagging churns as "preventable" or "structural") as a way of giving leadership a more honest picture of what renewal rates are actually telling them. This isn't about giving CSMs an easy excuse for every lost account. It's about building a measurement system that reflects reality rather than one that treats all churn as equivalent.

---

The problem with activity metrics

Activity metrics (number of calls logged, QBRs conducted, emails sent) are the other common component of CSM scorecards, and they carry their own risks.

The risk is Goodhart's Law: when a measure becomes a target, it ceases to be a good measure. A CSM who is being tracked on number of QBRs conducted will conduct more QBRs. Those QBRs may or may not be valuable. The customer may or may not find them useful. The metric is moving in the right direction; the underlying purpose of the QBR may not be.

Activity metrics are most useful as baseline guardrails rather than as performance targets. They tell you whether a CSM is engaging at a minimum level with their book, which is useful information, especially for CS leaders who want to catch under-engagement early. But they don't tell you whether the engagement is good, and optimising for the metric creates perverse incentives.

A better approach is to track activity quality alongside activity quantity, using a small number of input-output pairs: not "number of QBRs conducted" but "number of QBRs conducted that resulted in a documented next step." Not "number of accounts contacted this week" but "number of accounts where the last contact moved something forward."

---

Building the framework

A CSM performance framework that works tends to have four to six metrics, covering:

• Relationship depth (multi-threading score, number of meaningful contacts)
• Outcome progress (percentage of assigned accounts that have hit their agreed value milestone within timeframe)
• Risk capture rate (accounts surfaced before reaching automated red status)
• Playbook quality (documented plays with clear outcomes, retrospective participation)
• Renewal performance in context (renewal rate, with qualitative classification of churns)

None of these is perfect. All of them are better than NPS as a measure of what a CSM is actually contributing to the business.

The harder work is building the infrastructure that makes these metrics trackable. That requires the value milestone framework to exist. It requires risk flagging to be something the system can detect. It requires playbook documentation to be a habit. All of which points back to CS Ops as the enabling function.

---

Questions worth sitting with

If you had to explain to a CSM exactly why they were performing at their current level, what would you say, and how much of that explanation would rely on their renewal numbers versus everything else?

How do you currently distinguish between a CSM who is excellent and happened to have a hard book, and one who is average and happened to have an easy one?

What would change in how your CSMs spend their time if the performance framework shifted toward the metrics above and away from renewal rate as the primary signal?

---

The short version

CSM performance measured primarily on renewal rate creates the wrong incentives and doesn't reflect what CSMs actually control. A more honest framework tracks engagement depth, outcome delivery, risk capture, and playbook quality, with renewal rate contextualised by the factors that were and weren't within the CSM's sphere of influence.

Building this framework requires the underlying CS Ops infrastructure to exist first: value milestones, risk detection, playbook documentation.

---

Where Pivotal Path comes in

Pivotal Path helps CS leaders design performance frameworks that measure what CSMs actually contribute, and build the CS Ops infrastructure that makes those metrics trackable. If your CSM scorecards are heavy on renewal rate and light on everything else, we can help you build something more useful.