|

The only honest measure of customer health is whether the product is producing real work. Not logins. Not sentiment scores. Not survey responses. If the operator is designing, the relationship is healthy. If the operator stops designing, we have a problem, and we need to know about it immediately.

What we measure

We track design frequency — how often the agent is generating spatial designs — because active design output is the strongest signal that the product is embedded in a customer’s workflow. We measure command diversity: whether users employ one capability or many, because broad usage across design, optimization, and field operations indicates deep adoption rather than superficial experimentation. We watch user growth within each organization, because internal expansion is the most reliable leading indicator of retention. And we track domain expansion — whether a customer has loaded additional agent skills — because moving from one domain to two signals platform-level adoption, not point-solution usage. At Nexma, these are the signals that tell us whether we are earning our place in an organization or merely occupying it.

What we do not measure

Login frequency without context tells us nothing. NPS surveys and sentiment scores are vanity metrics that reveal how people feel about filling out surveys, not how much value they extract from the product. We care about production output: designs created, routes optimized, work orders generated. These are the artifacts of real work, and real work is the only metric we trust.

Intervention triggers

If design output drops for two consecutive weeks, we reach out. Not with a survey. With a direct question: what changed? Sometimes the answer is seasonal — project phases have natural lulls. Sometimes the answer reveals a usability problem we need to repair. The earlier we know, the faster we act. And we have learned that asking directly, with genuine curiosity rather than formulaic concern, produces more honest answers than any automated health scoring system.