February 11, 2013, 11:40 AM — The other day, after speaking to an audience of technical people about the importance of building good relationships at work, a young help desk manager asked, "What metrics should I monitor for that?" I realized that he was trying to reconcile the importance
of relationships with a conflicting tenet of his managerial faith that says, "If you can't measure it, you can't manage it."
We geeks love that idea. It meshes with our preference for objectively verifiable facts over squishy subjectivity. We find metrics comforting, having been taught to expect constant feedback on the question "How am I doing?" We get grades on tests and papers. We get scores in video games. We get raises and promotions at work.
But our ardor for this explicit feedback loop can lead us astray. We take it too far, transforming the not-meant-to-be-taken-literally "If you can't measure it, you can't manage it" into the absurd dogma "If you can't measure it, it doesn't matter."
With that mantra as our lodestar, we not only ignore hard-to-measure aspects of work such as relationships, but also self-righteously see that approach as a virtue. You can see this mindset play out in numerous destructive ways. Project leaders become obsessed with the process-prescribed approvals while ignoring the mutual trust required to get them. Technical leaders abrasively condescend to users or sponsors, since "they don't get it." Developers adhere to the precise content of a requirements document -- as if ticking off 100% of the requirements is more important than making sure the product makes sense -- and they end up delivering unusable code.
A more appropriate mantra would be, "Even if you can't measure it, you're still responsible for it." That's because some things, like relationships, are too important to be ignored but too subjective to be measured.
For those important things that metrics can't quantify, you need to use indicators. Indicators are metrics' cousins. Where metrics quantify observable facts that are very closely related to what you want to measure, such as production rates, indicators are less precise and more indirect. They are observable facts, but they can serve only as a proxy to suggest how you're doing on something fundamentally unmeasurable.
For example, there really is no metric for directly measuring the quality of your relationship with your project sponsor. But there are indirect indicators that help you assess that relationship. You could get a reasonable impression by asking questions such as these:
• Does she respond to my emails?
• Does she show up at our meetings prepared?
• Does she seek my opinion before making important decisions?
• Does she share information with me?
• Does she invite me to meetings that I feel I should attend?