What if you’re wrong? How to design for uncertainty
Relying on rigid plans and predictions ignores the unpredictable nature of humans. Instead of trying to be right all the time, we should focus on what happens when we’re wrong.
I have a friend who’s obsessed with Bitcoin. During the crypto craze of 2022, he was convinced the price was going “to the moon.” He was so sure he almost took out a bank loan to buy more. Thankfully, I talked him out of it. Not long after, Bitcoin crashed.
This wasn’t the first time he’d misjudged market trends, but somehow, he’s still absolutely confident in his predictions. He’s got all the charts, market analyses, and financial pundits on his side. And yet, I can’t help but wonder: how can he be so sure when so many of these forecasts are more luck than science?
Financial and political pundits may sound confident, but their predictions are about as reliable as horoscopes.
There’s a reason no one can consistently beat the market, why nobody saw Trump coming in 2016, and why the Amazon Fire Phone flopped despite the hype. All of these involve one unpredictable variable: people. Human behavior is messy, emotional, and impossible to forecast with certainty.
As product designers, we often fall into the same trap. We want to believe our users are rational beings who will behave exactly as we expect. With enough interviews and data, we convince ourselves we’ve cracked the code. But users aren’t robots. They don’t follow logic like an algorithm — they’re driven by emotions, biases, and the chaos of being human.
The illusion of rationality
A shirt is just a shirt, right? Well, not quite. Imagine I told you this shirt was worn by Lionel Messi when he won the World Cup. Suddenly, it’s not just fabric — it’s a priceless piece of history.
Now, what if the same shirt was worn by Jeffrey Dahmer during one of his crimes? It instantly becomes repulsive (unless you’re into true-crime memorabilia). Context changes everything, making us value — or devalue — things in irrational ways.
In Thinking, Fast and Slow, Daniel Kahneman offers plenty of examples of how humans defy logic. We overvalue things simply because we own them. We give more weight to anecdotal stories than to cold, hard statistics. And we’re champions at confirming our beliefs instead of challenging them with opposing evidence. As Dan Ariely puts it:
“We’re not thinking machines. We’re feeling machines that happen to think.”
So, if we can’t reliably predict user behavior, what should we do? Recognize that our assumptions are just educated guesses. Instead of trying to be right, focus on minimizing risk and reducing uncertainty. The best way to do that? Experimentation. Trial and error. Luckily, we already have a framework for this: the scientific method.
Unlike pundits or fortune tellers, scientists don’t claim to “know” the future. They form hypotheses, test them, and learn from the results. It’s an iterative process that drives progress, one small insight at a time.
Experimentation provides valuable data to guide decisions, but data isn’t a crystal ball. Numbers can identify problems, but they won’t hand you solutions on a silver platter. Metrics are best used as a compass, not a map. They point you in the right direction, but the path forward is yours to figure out.
Metrics are a compass, not a map
We love metrics. They give us the comforting illusion of control. User retention, time on page, click-through rates — these numbers feel objective, reliable, and actionable. But when it comes to human behavior, metrics can trick us into thinking we know more than we do.
Metrics are great for spotting patterns in customer behavior, but they lack the context behind those behaviors. Numbers can’t capture the nuances of human decisions. How do you measure a user’s emotions?
People aren’t just complicated — they’re complex. Not like a rocket ship (which, with enough expertise, can be replicated), but like the Amazon rainforest. You could remove a part from a rocket ship and still predict how it will function. But mess with a species in the rainforest, and you have no idea how the entire ecosystem will react.
Creating products for customers is a lot like creating ecosystems. Remove one feature or service, and the ripple effects can be unpredictable, no matter how much data you have.
Snapchat’s 2018 redesign is a perfect example. Hoping to improve engagement, Snapchat redesigned its interface and removed some underused features — relying heavily on metrics to guide these changes. The result? A massive drop in user engagement and a $1.3 billion hit to its market value. Not exactly what they were aiming for.
Now, I’m not saying we should ignore data and just trust our gut. Far from it. The more information we have, the better we can make informed decisions. But data isn’t enough. We need to respect the complexity of human behavior and understand that no metric can predict every outcome. Metrics are tools — not oracles. They should guide us, not dictate our decisions.
Trial and error is the only reliable method
Imagine climbing a mountain no one has ever climbed before. You don’t know the best path to the top, and every step is a gamble. You’ll hit dead ends, face tough terrain, and backtrack often. But with each misstep, you learn more about the landscape. By the time you reach the summit, you’ve mapped out the best route.
This is what creating value for customers looks like. Whether climbing a mountain or building a product, the journey always starts with assumptions. The key question is: What happens if we’re wrong? How big can we afford to fail?
Relying on assumptions to predict customer behavior is risky. It can lead to expensive mistakes when those assumptions don’t match reality. Yet companies often rely on detailed long-term roadmaps, packed with milestones and deadlines. This is a minefield. The further ahead we plan, the more assumptions we make, and the less prepared we are for unexpected changes.
As the great philosopher Mike Tyson said “Everyone has a plan until they get punched in the mouth.” Anything can be that punch. A one-year plan for a vacation rental service seems solid — until COVID happens. AI innovations might render your tools obsolete. Geopolitical conflicts or market crashes could derail everything. The world is unpredictable, and long-term plans are brittle in the face of change.
We need to stop obsessing about shipping on schedule and start obsessing about going in the right direction. Methodologies like The Lean Startup or Lean UX encourage a learning culture that helps us validate if we are on the right track. Instead of following a detailed plan, you take small steps. You build, measure, and learn. This approach allows you to adjust as you go, minimizing risk and maximizing learning.
Instead of obsessing over shipping on schedule, we need to focus on moving in the right direction. Methodologies like The Lean Startup or Lean UX teach us to embrace a culture of learning. Instead of following rigid plans, we take small, deliberate steps: build, measure, and learn. This iterative approach minimizes risk and maximizes adaptability.
We improve not by sticking to assumptions but by testing them. Small experiments let us validate ideas before making big investments. Progress isn’t about guessing right — it’s about learning as you go.
Don’t believe in fortune tellers
The illusion of certainty is comforting. It’s easy to feel confident in our beliefs, while admitting doubt or uncertainty can seem like a weakness. But pretending we know more than we do doesn’t change reality: humans are unpredictable, and the future is anything but certain.
There’s no shame in admitting we don’t have all the answers. In fact, that’s where innovation starts. The more experiments we run, the more we learn. And the more we learn, the more comfortable we become with uncertainty.
The best way to face the unknown is to think like scientists. Instead of betting everything on being right, we experiment, adapt, and grow. It’s a lot like evolution by natural selection. Products should evolve through trial and error — not through intelligent design.
So, the next time a pundit, analyst, or manager claims to predict the future, remember: they’re just guessing. Think like a scientist. Get comfortable with being a little wrong — it’s the only way to eventually get it right.
