Building Products During Uncertainty

Every startup article written before 2020 treats uncertainty as a manageable risk. "Plan for contingencies." "Build a buffer." "Stay lean."

2020 turned uncertainty from a planning exercise into a lived reality. Not the abstract uncertainty of "maybe the market will shift," but the concrete uncertainty of "will my industry exist next month?"

Building products during this kind of uncertainty requires a different mindset than building during growth. Here's what I learned.

The Default State Is Uncertain

The pre-COVID mental model: stability is the default, disruption is the exception. You plan for the expected case and prepare contingencies for the unlikely case.

The post-COVID mental model: uncertainty is the default, stability is a lucky streak. You build systems that work under a range of conditions, not systems optimized for one condition.

This mindset shift changes every decision:

Architecture decisions favor simplicity over optimization. Simple systems are easier to adapt. Optimized systems are tuned for specific conditions that might change.

Product decisions favor versatility over specialization. A product that serves multiple use cases (even if it serves each one less perfectly) is more resilient than a product optimized for a single use case that might disappear.

Financial decisions favor cash reserves over growth investment. Cash is the ultimate hedge against uncertainty. Growth investments assume a future that might not materialize.

What Works During Uncertainty

Low burn rate. The products that survive downturns are the ones that can operate on minimal revenue. My products have minimal fixed costs: serverless infrastructure scales down as traffic drops, there's no office lease, no employee payroll. When revenue dips, costs dip proportionally.

Direct customer relationships. During uncertainty, direct communication with customers reveals what's changing in real time. Email conversations with pilot students told me how training was evolving months before industry reports confirmed it.

Rapid iteration. Uncertainty means the ground is shifting. Products that iterate quickly (weekly releases, fast response to market changes) adapt to new realities. Products on quarterly release cycles are always one quarter behind the current reality.

Diversified revenue. Multiple products, multiple markets, multiple customer segments. When aviation contracted, the contraction was specific to aviation. Other markets were less affected or even growing.

What Doesn't Work

Long-term planning. Detailed twelve-month roadmaps are fantasies during uncertainty. I replaced annual plans with quarterly experiments, three-month bets that I evaluate and adjust.

Aggressive expansion. Expanding into new markets, hiring ahead of demand, and investing in growth assumes the future is predictable. During uncertainty, expansion bets can turn into liabilities.

Waiting for clarity. "I'll build it when the market settles" means never building it. Uncertainty doesn't resolve into clarity. It evolves into different uncertainty. Waiting for perfect conditions is waiting forever.

The Paradox of Building During Crisis

Here's the paradox: some of the best products are built during downturns.

Competition retreats. Competitors who were growing aggressively pre-crisis pull back, cut budgets, and pause development. The competitive landscape thins.

Users are more receptive. Users during a crisis are actively looking for better solutions. The pain points are sharper, the willingness to try new tools is higher, and the feedback is more specific.

Focus is forced. Uncertainty eliminates distraction. There's no shiny new market opportunity to chase. There's only the work: build the product, serve the users, survive to build another day.

The products I built and improved during 2020, while the world was uncertain and my primary market was in crisis, became the foundation for everything that followed. Building during uncertainty isn't comfortable. But the work you do when things are hard compounds when things get better.