When you build software, you add little hooks into the code so that, as users open a window, tap a picture, upload a file, the code tattles on them, sending some of their data to another company’s server. Log data is sowed; reports are reaped. This is known as “analytics” or, if there are people with advanced degrees involved, “data science.”

I cofounded a software company, like a doofus, so I attend a weekly analytics meeting. I nod at the Zoom camera and say things like “Not surprising to see that drop-off” or “Promising, but let’s keep our focus on the people who aren’t engaged.” My cofounder runs “product”; I’m loosely in charge of “the funnel.” The funnel, in case you are blessed not to know, is an inverted triangle with horizontal stripes. The top of the funnel, the first stripe, is the stuff that brings in visitors—advertising, YouTube tutorials, LinkedIn posts, newsletters, blog posts, all the endless churn of content. Some fraction of those visitors sign up for the newsletter (that’s stripe two), and a fraction of those sign up for the product (three), and a fraction of those turn into customers—conversion! So it’s not really a funnel, more of a juicer.

In our analytics meetings, we measure the human juice as it dribbles in: pages visited, sign-ups, actions taken. We talk about how to squeeze more. We behave ourselves and don’t track the people who ask us not to track them. We say things like “Team, 98 percent of our beloved users never click the gray button. Have we considered red?” No one ever says, “Eileen in apartment 4A is saving links about fentanyl—let’s tell her insurer.” They’re good meetings. I’ve done them for years. But this past summer, something felt off.

The summer was very warm. I don’t need to tell you, do I? One of our managers came back from vacation and told us that they rarely left their hotel because it was too hot in the daytime. Birds and humans alike changed migration patterns, sometimes to route around floods and fires. Protesters questioned humanity’s endless focus on growth; arrests were made. Climate Week came and went, presumably leaving 51 other weeks without climate.

As the leaves turned, my wilted brain figured it out. Here I was, looking for growth—how to get from 10,000 users to 10,001—while outside the company, people were marching about how it was time to focus on absolutely anything else. I tend to agree with them. Our startup has a small carbon footprint, so we’re not the problem in that sense. But was growth the right metric, the only metric, for us to obsess about?

My mind drifted to a book called Lean Logic. It’s a big red book. I took it off the shelf and skimmed through it. It’s the life’s work of a British economist named David Fleming, published after his death in 2010. Fleming was of his time—big into predicting peak oil and very against nuclear power—but the book is a one-person hypertext of surprising depth, and it is very helpful framing as things get a little worse. (The whole thing is available for free at Leanlogic.online.)

Fleming believed that growth has natural limits. Things grow to maturity—kids into adults, saplings into trees, startups into full-fledged companies—but growth beyond that point is, in his words, a “pathology” and an “affliction.” The bigger and more productive an economy gets, he argued, the more resources it needs to burn to maintain its own infrastructure. It becomes less and less efficient at keeping any one person clothed, fed, and sheltered. He called this the “intensification paradox”: The harder everyone works to make the GDP line point up, the harder everyone has to work to make the GDP line point up. Inevitably, Fleming believed, growth will turn to degrowth, intensification to deintensification. These are things to prepare for, plan for, and the way to do that is with the missing metric: resilience.

Fleming offers several definitions of resilience, the briefest of which is “the ability of a system to cope with shock.” He describes two kinds: preventive resilience, which helps you maintain an existing state in spite of shocks, and recovery-elastic resilience, which helps you adapt quickly to a new post-shock state. Growth won’t help you with resilience, Fleming argues. Only community will. He’s big on the “informal economy”—think Craigslist and Buy Nothing, not Amazon. People helping people.

So I began to imagine, in my hypocritical heart, an analytics platform that would measure resilience in those terms. As growth shot too high, notifications would fire off to your phone: Slow down! Stop selling! Instead of revenue, it would measure relationships formed, barters fulfilled, products loaned and reused. It would reflect all sorts of non-transactional activities that make a company resilient: Is the sales team doing enough yoga? Are the office dogs getting enough pets? In the analytics meeting, we would ask questions like “Is the product cheap enough for everyone?” I even tried to sketch out a resilience funnel, where the juice that drips down is people checking in on their neighbors. It was an interesting exercise, but what I ended up imagining was basically HR software for Burning Man, which, well, I’m not sure that’s the world I want to live in either. If you come up with a good resilience funnel, let me know. Such a product would perform very badly in the marketplace (assuming you could even measure that).

The fundamental problem is that the stuff that creates resilience won’t ever show up in the analytics. Let’s say you were building a chat app. If people chat more using your app, that’s good, right? That’s community! But the really good number, from a resilience perspective, is how often they put down the app and meet up in person to hash things out. Because that will lead to someone coming by the house with lasagna when someone else has Covid, or someone giving someone’s kid an old acoustic guitar from the attic in exchange for, I don’t know, a beehive. Whole Earth stuff. You know how it works.

All of this somewhat guilty running around led me back to the simplest answer: I can’t measure resilience. I mean, sure, I could wing a bunch of vague, abstract stats and make pronouncements. God knows I’ve done a lot of that before. But there’s no metric, really, that can capture it. Which means I have to talk to strangers, politely, about problems they’re trying to solve.

I hate this conclusion. I want to push out content and see lines move and make no more small talk. I want my freaking charts. That’s why I like tech. Benchmarks, CPU speeds, hard drive sizes, bandwidth, users, point releases, revenue. I love when the number goes up. It’s almost impossible to imagine a world where it doesn’t. Or rather it used to be.

This article appears in the November 2023 issue. Subscribe now.