“My manager needs to put an ROI on everything,” an SRE leader told me at KubeCon this week. Yeah, I feel that. We know the value of SRE, smoothing all other work so that production can produce. Putting that into dollars though? It’s hard to count the cut costs of incidents that didn’t happen. Yet someone has to justify each effort in ROI terms to keep the team funded.
On the way to KubeCon I read about Louis Agassiz, a famous naturalist in the nineteenth century. Back then, people didn’t value studying nature for its own sake. Learning the names of plants, observing the world in detail, and keeping detailed notes were frivolous activities, shiftless, wasteful. A better person engaged in industry every hour of the day, with busy hands crafting objects of obvious use. But Agassiz made taxonomy sound important! He justified it in religious terms: by studying nature, he could learn about God’s hierarchy. He could find moral instruction in the differences between species.
This sounds ridiculous now, but it spoke to the values of the day. It let Agassiz fund teams and expeditions full of scientists to catalog the world, increasing human knowledge. Today, we value increased knowledge for its own sake. Back then, knowledge was not respeected, but virtue was. Agassiz put an ROI on science, in the currency of the day.
The taxonomists and botanist, fans of fishes and lovers of leafs, felt the importance of their work. It took a charismatic leader like Agassiz to make it relevant to people who controlled resources.
SREs, developers, and the managers and directors who establish and protect those programs feel intuitively that this work is crucial. Their work safeguarding the software systems lets everyone keep working, lets the business move forward with options. It takes a leader to know this, and work it into the kind of narrative that VPs need to convince the executives, that executives need to satisfy the board.
Cultural value systems change, personal perspectives shift. I have advanced from “code quality is a virtue” to “appropriate attention to malleability keeps our options open.” While the people I respect value system health within complexity, capitalist business culture demands we linearize into an ROI story.
Someone has say that upgrading our software observability is going to save money in terms of shorter incidents and increased development speed, with a bonus of lowering the risk of AI-generated code. Find some ways to put numbers around that. No, it is not really measurable. From developers to CTO and CEO, we know system health is not a linear outcome. We know there’s more to the story, and still we need the story.
Agassiz believed his story, though. After Origin of the Species came out, scientists around the world gradually shifted to work from evolution, to find family trees in the fishes and ferns. Agassiz never came around. Every species was a thought of God, and he would not give that up.
We don’t have to believe that the real reason to upgrade our libraries and architecture is some specific dollar savings. We know that it eases everything we build on top. We can choose our timing and latch on to some feature that is perceived as profitable, we can point to our upgrade as facilitating that in particular. Keep both stories in our heads, the deep benefits and the nominal ones.
When we understand the deeper value of what we’re trying to do, then expressing what we need in terms relevant to people who control the resources is not lying. It’s leadership.