Stop making fragile decisions
Why robustness beats brilliance (and what Buffett really taught us).
Let's talk about failure. I built a failed data catalog for a company—well, my team did. The idea was sound, well-discovered, and validated. When it shipped though, one of the key engineers wasn't there anymore. No big deal, right? It was a great team.
Except it was a big deal. That 5% difference in execution made the difference between a product users could actually use and one that needed serious fixing. The catalog gathered digital dust while teams continued their manual workarounds.
Here's another story. I prioritized a big strategy paper supposed to be handed out to customers. Took weeks writing it; I supplied marketing material. But on launch, the marketing team got pulled away for another initiative, leaving only 60% of the resources. The strategy paper tanked. Weeks of effort lost.
Both failures teach the same lesson: fragile decisions break when reality interferes with your plans.
Warren Buffett understood this decades ago. In Pulak Prasad's "What I Learned About Investing from Darwin," Prasad details Buffett's approach—not investing in brilliant managers, but in businesses so fundamentally sound they could withstand almost any leadership change. I've learned to apply the same principle to product decisions.
Great decisions under uncertainty need to survive chaos.
If you only have 5 minutes: here are the key points
Robustness over brilliance: Like Buffett's investments, good decisions remain valuable despite leadership changes, resource shifts, or execution imperfections
Fragile brilliance fails: Perfect plans requiring perfect execution are disasters waiting to happen
Mathematical metaphor: Like a marble in a bowl, robust decisions return to stability after disturbances
Know your epsilon: Understand typical disruptions in your environment and build accordingly
Real examples matter: Theranos (founder-dependent), MoviePass (economics-dependent), Google+ (execution-dependent) all lacked robustness
0. Define what robustness doesn't mean
Here's where things get counterintuitive. Warren Buffett is famous for investing in companies led by exceptional managers. This creates the impression that Buffett is betting on superstar leadership.
But that's precisely the misunderstanding. Buffett isn't investing because of these individuals—these individuals created businesses so fundamentally sound they would thrive even without them.
It's easy to fool yourself into thinking you've found that "once in a lifetime opportunity." Theranos looked revolutionary until Elizabeth Holmes's charisma couldn't cover the lack of working technology. MoviePass seemed genius until basic unit economics destroyed it. WeWork appeared transformational until Adam Neumann's departure revealed the business had no sustainable model.
Robustness isn't about perfection under ideal circumstances—it's about adequacy across a wide range of circumstances.
1. Define your decision and what robustness means to you
Robustness isn't universal—it needs context-specific definition. For Buffett, a robust business generates consistent returns despite management changes, economic fluctuations, or competitive pressures. For product managers, a robust feature delivers value despite implementation variations, user behavior shifts, or resource constraints.
The key is identifying the core value proposition that must remain intact regardless of external factors.
For my failed data catalog, robustness would have meant: "Users will use it, even if the execution is 10-15% off from great."
For the strategy paper, it would have meant: "Customers can understand our value proposition even if marketing support is minimal and distribution is primarily through existing channels."
To define robustness for your decision:
Identify the core value that must be preserved
List the critical assumptions underlying your decision
Determine which outcomes would indicate failure
2. Know your epsilon—and build for it
In mathematics, epsilon represents small changes to a system. But how small is "small" in your context?
For Dropbox, epsilon meant surviving when Gmail added file sharing. For Slack, it meant thriving even when Microsoft bundled Teams with Office. For Zoom, it meant working flawlessly when the entire world suddenly needed video conferencing.
A useful epsilon is a disturbance you can reasonably expect to occur. For Buffett, it means that if a handful of key people (including the CEO) leave, the business should still perform excellently. For me as a product manager, it means that if key team members move to other projects, the feature should still deliver net positive value.
To define your epsilons:
List specific disruptions most likely in your environment
Quantify the typical magnitude of these disruptions
Identify which disruptions have historically derailed similar initiatives
In my data catalog case, the epsilon should have been: "Primary engineer leaves, documentation is incomplete, and adoption training gets cut by 50%." Had I built for that scenario, the outcome would have been different. Or I would’ve opted not to build this product at all.
3. Ruthlessly eliminate fragile options
This is the most counterintuitive aspect of robust decision-making. Many options that look promising under ideal conditions must be discarded because they lack robustness.
Let's be clear: This doesn't mean avoiding risky options. Quite the opposite. You're looking for big growth opportunities with capped downside. Robustness is how you save yourself while still finding massive growth opportunities.
Buffett famously operates with a "too hard" pile—investment opportunities that might be lucrative but whose outcomes are too sensitive to factors beyond his control. Google+ failed not because social networking was a bad bet, but because success required perfect execution across multiple dimensions while Facebook only needed to not screw up what they were already doing well.
In product management, this might mean killing feature ideas that solve real problems but require perfect execution across multiple dimensions. Choose the "good enough" solution that works reliably over the potentially brilliant solution that demands everything go right.
The uncomfortable truth: Most brilliant plans are actually fragile plans in disguise. Robust decisions might look boring to your team, but they'll still be working when the brilliant ones have crumbled under the weight of reality's inevitable chaos.