Human beings are extraordinary problem-solvers. Give us an obstacle, an inconvenience, or something that takes longer than we think it should, and we’ll immediately begin inventing ways around it. We have spent hundreds of thousands of years doing precisely that, and the results have been remarkable.
We’ve learned to cultivate food, harness electricity, cure diseases, fly across oceans, and send people into space. We’ve created machines capable of performing calculations that would have occupied entire generations of mathematicians. We’ve developed technologies that allow us to communicate instantly with someone standing on the opposite side of the planet.
And yet, despite all this ingenuity, we seem to have developed a peculiar talent for solving one problem only to discover that we’ve created three more.
I sometimes wonder whether humanity’s greatest invention isn’t the wheel, the printing press, or the computer, but the ability to make our own lives progressively more complicated.
Consider the automobile.
At the turn of the twentieth century, transportation was a messy business. Quite literally. Horses were the primary means of moving people and goods through many cities, and while they were certainly useful, they came with considerable disadvantages. They needed food, shelter, veterinary care, and regular opportunities to relieve themselves, which they did with an enthusiasm that urban sanitation departments probably found less than charming.
Along came the automobile, promising greater speed, independence, and convenience. It was an extraordinary innovation, and like so many great inventions, it succeeded beyond anything its early adopters could have imagined.
People could travel farther. Businesses could transport goods more efficiently. Entire industries emerged around manufacturing, maintenance, fuel, and infrastructure. The automobile transformed economies, expanded personal freedom, and reshaped the modern world.
And then we started building everything around it.
Roads became highways. Highways became interstate systems. Cities expanded outward. Homes moved farther from workplaces. Shopping centers appeared miles from residential neighborhoods. What had once been a remarkable convenience gradually became a practical necessity.
Today, millions of people spend substantial portions of their lives sitting in traffic, surrounded by other people who purchased automobiles for essentially the same reason: to get somewhere more efficiently.
We invented a machine to make transportation easier, then built a society in which getting from one place to another can be extraordinarily difficult.
And what have we done to address the problem?
We’ve built more roads, expanded highways, developed sophisticated traffic-management systems, and invested in increasingly intelligent vehicles. Some of these solutions have produced genuine improvements. Others have introduced additional costs, dependencies, and complications.
The automobile wasn’t a failure. Quite the opposite.
Its success changed the world so profoundly that it created problems the original invention was never intended to address.
And that’s where our story begins.
The Peculiar Consequences of Success
We’ve all heard the familiar observation that we learn more from failure than success.
It sounds sensible. Failure is uncomfortable, occasionally embarrassing, and remarkably effective at getting our attention. When something goes wrong, we tend to examine it.
What happened? Why did it happen? What could we have done differently?
In business, we’ve developed entire disciplines devoted to answering those questions. Root-cause analysis, process improvement, incident investigations, corrective-action plans, and continuous improvement methodologies all help us understand what went wrong and how to prevent it from happening again.
I have spent much of my professional career working with organizations attempting to improve complex systems, particularly in healthcare and technology. I’ve seen how much time, energy, and expertise can go into understanding a failed process or correcting an unsuccessful implementation.
And that’s appropriate. Failure deserves investigation.
But I’ve also noticed something interesting about success.
When something succeeds, our questions often change.
Instead of asking what happened and why, we ask how quickly we can expand it.
How do we replicate this? How do we implement it across the organization? How do we increase production? How much faster can we go?
Success becomes an invitation to accelerate.
Failure encourages investigation. Success encourages expansion.
Of course, this isn’t universally true. Successful organizations frequently evaluate their achievements, and many conduct extensive post-implementation reviews. But the distinction between measuring whether a solution worked and understanding everything its success changed is important.
A project can meet every stated objective while creating consequences that weren’t included in those objectives.
A technology can reduce costs while increasing dependency.
A process can improve productivity while introducing new administrative burdens.
An innovation can make something easier while fundamentally changing what people are expected to accomplish.
None of these outcomes necessarily means the original solution was wrong.
They mean that success is more complicated than we sometimes acknowledge.
Perhaps we don’t learn less from success because success has less to teach us. Perhaps we learn less because we stop asking questions too soon.
Three Hundred Thousand Years of Getting Better at Things
Our species, Homo sapiens, emerged approximately 300,000 years ago.
For much of that history, survival depended on our ability to adapt, cooperate, communicate, and develop solutions to the challenges around us.
We learned to create increasingly sophisticated tools, preserve knowledge, and pass discoveries from one generation to the next. Unlike biological evolution, which operates through inherited variation across generations, cultural and technological innovation allowed us to change our surroundings through accumulated knowledge and deliberate action.
One generation discovered something useful. The next improved upon it. Another found a new application.
Innovation became cumulative.
And cumulative innovation is among humanity’s most remarkable achievements.
The ability to build upon previous discoveries has allowed us to accomplish things that would have been unimaginable to our ancestors.
But innovation has another characteristic that deserves our attention.
Every successful innovation changes the conditions in which the next innovation occurs.
The printing press transformed the distribution of knowledge. Electricity changed how societies worked and lived. The automobile reshaped transportation and geography. The internet transformed communication, commerce, and access to information.
Each advancement created opportunities that previously didn’t exist.
Each also introduced new dependencies, expectations, and challenges.
This doesn’t mean innovation is inherently destructive. Far from it. Many innovations have produced extraordinary improvements in health, safety, prosperity, and quality of life.
But it does mean that innovation isn’t simply a collection of independent achievements.
It’s an evolving process in which every success changes the environment for what comes next.
And if innovation is a process, shouldn’t we examine how that process evolves?
Then Along Came Artificial Intelligence
If the automobile demonstrates what happens when successful innovation transforms society over generations, artificial intelligence presents a different challenge.
Speed.
The development of modern AI systems has made increasingly sophisticated capabilities accessible to enormous numbers of people and organizations.
Tasks that once required specialized expertise can now be performed, or at least assisted, by software. Information can be analyzed, documents created, code generated, and complex workflows automated with remarkable efficiency.
These capabilities offer tremendous opportunities.
In healthcare, AI may help clinicians analyze information, reduce administrative burdens, and identify patterns that might otherwise go unnoticed. In business, it can streamline operations and expand access to expertise. In education, it can support new approaches to learning.
But the speed at which software can be distributed introduces a challenge that previous generations of innovators encountered differently.
A successful technological capability can spread faster than our ability to understand all the changes it creates.
Consider a hypothetical hospital that introduces an AI-powered documentation system.
The technology works beautifully. Physicians spend less time completing administrative paperwork. Documentation becomes faster. Productivity improves.
The implementation is declared a success.
And by the measurements established at the beginning of the project, it is.
But what happens next?
Does the hospital give physicians more time with patients? Does it increase the number of appointments they are expected to complete? Does it introduce new verification requirements? Does the organization become dependent on a technology that its employees no longer know how to operate without?
And perhaps most importantly, does the innovation ultimately improve patient care?
The original success doesn’t answer those questions.
They’re questions about what success creates.
This distinction becomes increasingly important as innovation accelerates. When the time between demonstrating a successful solution and expanding it across an organization becomes shorter, the opportunity to understand its broader consequences can shrink.
We become exceptionally good at measuring how much faster something works while remaining less certain about whether the larger system has actually improved.
We’ve entered an era in which the speed of innovation may exceed the speed of our institutional learning.
And that should concern us—not because innovation is undesirable, but because innovation is too important to evaluate incompletely.
A Million Needles
Some time ago, while thinking about the challenges facing modern society, I began developing an idea I called A Million Needles.
The metaphor originally emerged from a simple observation about the human condition.
We often imagine that significant problems arise from significant events. A catastrophic failure. A disastrous decision. A single moment when everything goes wrong.
But many of the challenges we experience don’t arrive that way.
They accumulate.
A small compromise here. An additional obligation there. Another process, another dependency, another expectation.
Each addition seems manageable.
Until one day, we discover that the collective weight of those decisions has become something entirely different from what anyone intended.
A million needles.
Not one enormous catastrophe, but countless small changes whose combined consequences become increasingly difficult to recognize, much less resolve.
The metaphor applies surprisingly well to innovation.
Every new technology, process, policy, and system is introduced for a reason. Most are intended to solve legitimate problems.
But as those solutions accumulate, they begin interacting with one another.
A process designed to improve accountability introduces additional reporting requirements. New reporting requirements encourage the adoption of another technology. That technology creates new maintenance responsibilities. Those responsibilities require new policies, training, and oversight.
Each decision may be reasonable when examined individually.
Yet the resulting system can become extraordinarily complicated.
And when the complexity itself becomes a problem, our instinct is often to introduce another solution.
Another needle.
What makes this particularly difficult is that no single decision necessarily caused the problem.
The challenge emerges from the accumulation of otherwise sensible choices.
It’s tempting to respond by arguing that we should simply stop adding complexity.
But that would be too easy, and often wrong.
Some complex systems are necessarily complex. Modern healthcare, aviation, telecommunications, and infrastructure require sophisticated coordination because the problems they address are genuinely difficult.
The objective shouldn’t be to eliminate complexity indiscriminately.
It should be to understand whether the complexity we introduce produces benefits that justify its consequences.
And that requires us to evaluate success differently.
The Second Success Test
I believe we need to become as disciplined about investigating success as we are about investigating failure.
Not because success is suspicious.
Not because innovation should slow to a crawl.
But because successful solutions can produce consequences that extend far beyond the problems they were designed to solve.
I call this approach the Second Success Test.
The first success test is familiar.
Did our solution work?
Did we accomplish what we intended? Did we reduce costs, improve efficiency, increase access, or solve the original problem?
These are important questions. We should continue asking them.
But they shouldn’t be the end of our evaluation.
The second success test asks something different.
What did our successful solution change?
It’s a deceptively simple question.
And answering it requires looking beyond the original objectives.
We need to examine how success changes people’s behavior, expectations, and relationships. We need to identify the new dependencies that emerge and understand how those dependencies affect the larger system.
We need to investigate not only the intended benefits but also the unexpected costs and opportunities.
And we need to ask whether the benefits remain meaningful as the solution expands.
Consider the automobile again.
The first success test asks whether automobiles improved personal transportation.
They did.
The second asks how widespread automobile ownership changed cities, infrastructure, environmental conditions, economic opportunities, and the ways people organize their lives.
Those questions reveal a much larger story.
Or consider our hypothetical AI documentation system.
The first test asks whether physicians complete documentation more efficiently.
The second asks whether the resulting changes improve the experience of physicians, patients, and the healthcare system as a whole.
The first test measures whether we accomplished our objective.
The second examines whether accomplishing that objective produced the improvement we actually wanted.
This isn’t a replacement for established approaches to risk management, systems thinking, or continuous improvement. Those disciplines already provide valuable tools for evaluating consequences.
Rather, it’s a call to apply that discipline more consistently to successful innovations, particularly before expanding them.
And there’s one additional dimension we cannot ignore.
Time.
A solution that succeeds today may produce different consequences tomorrow.
An innovation that works beautifully for a hundred users may behave differently when adopted by a million.
A process that improves efficiency during its first year may introduce dependencies that become increasingly expensive to maintain.
Success isn’t a permanent condition established at the moment we declare victory.
It’s an evolving outcome that deserves continued examination.
We should evaluate success when it first occurs, when it expands, and after it becomes part of everyday life.
The faster we innovate, the more important that continuing evaluation becomes.
Learning to Ask Better Questions
There’s an understandable temptation to interpret all this as an argument against rapid innovation.
It isn’t.
I believe innovation is one of humanity’s greatest strengths.
Our ability to recognize problems, imagine alternatives, experiment with solutions, and build upon previous discoveries has transformed the human experience.
I wouldn’t want to return to a world without modern medicine, reliable transportation, electricity, or instant communication.
Nor would I want to abandon the extraordinary opportunities emerging from artificial intelligence.
But appreciating innovation doesn’t require us to assume that every successful implementation automatically represents lasting improvement.
We can celebrate an achievement while remaining curious about its consequences.
We can recognize that a solution worked while asking whether the larger system became better.
And we can acknowledge that progress sometimes requires not another addition, but reconsidering something we’ve already built.
This applies to organizations, governments, communities, and our personal lives.
We adopt technologies to save time, then discover that we’ve filled the available time with additional obligations.
We create systems to simplify decisions, then find ourselves navigating increasingly complicated processes.
We introduce conveniences that gradually become necessities.
These developments aren’t always harmful. But they’re worth examining.
Because the question isn’t simply whether we can make something faster, easier, or more efficient.
The question is whether doing so makes life meaningfully better.
Reevaluating How We Evaluate
We live in a world increasingly defined by three expectations.
More.
Better.
Faster.
We want more information, better technology, and faster results. We expect organizations to innovate continuously and individuals to adapt just as quickly.
There’s enormous value in those ambitions.
But somewhere along the way, I believe we’ve allowed the measurement of progress to become too narrowly focused on the immediate achievements of innovation.
We celebrate what a solution accomplishes.
We measure how efficiently it performs.
We calculate how quickly it can expand.
And then we move on to the next problem.
But perhaps the most important questions are the ones we haven’t been asking consistently enough.
What changed because we succeeded?
What new expectations did we create?
What dependencies now exist that didn’t exist before?
What consequences will emerge as this success expands?
And will the world actually be better because of it?
Humanity has spent approximately 300,000 years developing the extraordinary capacity to innovate. We’ve learned to create solutions, improve them, and build upon them.
That process has brought us further than our ancestors could possibly have imagined.
But the process itself must continue to evolve.
Our methods for evaluating innovation should advance alongside our ability to innovate.
We should investigate success with the same curiosity, discipline, and determination that we bring to understanding failure.
Because the greatest danger isn’t necessarily that our solutions will fail.
It’s that they will succeed so quickly, and expand so extensively, that we won’t fully understand what we’ve created until we’re already trying to solve the next problem.
And then the next.
And the next.
Another needle added to the million.
Perhaps it’s time to recognize that the successful completion of a solution isn’t the end of the investigation.
It’s the beginning of another.
In a world of more, better, faster, we need to reevaluate how we evaluate.
And before we scale what works, we must understand what working changes.
An essay by TREVOR DALE.More writing →