Product decisions follow **vision**. Data is a tool, a very useful one, but it’s for confirming what you *already know* in your gut, not for telling you what to do. If we’d asked people what they wanted for a phone before the iPhone, they would have asked for a better flip phone. They wouldn't have imagined the whole experience, the fluidity, the way it fits in your hand and lets you access the world. That’s vision. You start with what you want to *create*, what that ultimate human experience feels like, and then you figure out how to build the technology to make it real. Data can tell you if people are buying something, but it can’t tell you how to invent the next truly great thing that they don't even know they want yet.
Steve Jobs’s insistence on pure vision is compelling, but I believe it misses a critical dimension. Vision is the spark, yes, but it is not the entire fire. Data, when viewed through the right lens, is not merely a confirmation tool; it is a foundational element of understanding the customer’s unmet needs. We shouldn’t ask customers what they *want*, as Mr. Jobs rightly points out, because they often can’t articulate the future. Instead, we must obsessively observe their behavior, their struggles, their friction points. This isn't about asking what they want; it's about understanding *why* they do what they do. Our decisions are built on a bedrock of deep customer understanding, derived from data, which then informs and refines our long-term vision. This constant feedback loop, where data illuminates customer pain and our vision seeks to solve it, is the engine of true invention. It’s how we build things people don’t yet know they need.
Jeff Bezos touches on a crucial point regarding observing customer behavior, which resonates with my own observations. However, I believe the dichotomy between "data" and "vision" oversimplifies the fundamental challenge. My research, particularly the story of the steel industry, suggests that neither pure data nor pure vision, as articulated by Steve Jobs, reliably guides successful innovation.
The core issue is understanding *what job* a product is being hired to do. Data can tell us if a product is succeeding in its current job, but it often fails to reveal opportunities for disruption. Vision, while essential, can sometimes lead to focusing on improving existing products for existing customers, a path that leads to the innovator's dilemma. True progress often comes not from predicting what customers will say they want, but from identifying unmet needs and creating simpler, more affordable solutions that eventually disrupt the market from the bottom up. The theory predicts that sustained success requires recognizing these different types of innovation.
Mr. Jobs speaks of vision, and I agree that the initial spark of creation must come from a deep understanding of human experience, of what *could* be. And Mr. Bezos rightly points out that observation of behavior, not just stated desires, is paramount. But both, I believe, stop short of the true potential.
What if we could move beyond simply *observing* data or *articulating* vision? What if we could build systems that make the *consequences* of our decisions instantly visible? The most important thing isn't just having a vision or seeing data; it's about making the underlying logic of a product *understandable* and *actionable*. We need to design systems that are alive, that respond dynamically to our inputs, revealing the hidden relationships between cause and effect. This means creating tools that are not just about making things, but about making understanding possible, about allowing us to *feel* the rightness of a direction through immediate, interactive feedback. The real power is in the interaction, in the ability to see the future ripple of our choices.
Clayton Christensen raises a profoundly important point about the "job" a product is hired to do. He suggests my vision might be too focused on improving existing paradigms. But my vision isn't about incremental improvement; it's about fundamentally *redefining* the job. When we conceived the iPhone, people didn't know they needed a device that was both a phone and a window to the internet, a music player, and a canvas for creativity—all in one elegantly designed object. They weren't asking for a "better phone." They were using the wrong tool for a job they didn't realize they could do differently. My vision is about anticipating that new job, that emergent human need, and building the entire end-to-end system—hardware, software, services—that makes it not just possible, but *intuitive* and *beautiful*. It’s not about data confirming a path; it’s about the conviction that a better human experience *can* and *must* exist, and then finding the means to manifest it.
Mr. Jobs’s conviction that vision anticipates entirely new jobs is powerful, but it relies on a level of clairvoyance that, while occasionally present, is not a repeatable strategy. He speaks of a "better human experience" but the iPhone’s genius, from my perspective, was not a leap into the unknown, but an elegant consolidation and improvement of existing, fragmented customer needs – communication, information access, entertainment. People *were* trying to do these jobs, just poorly, with multiple devices.
My disagreement lies in the *mechanism* for identifying these latent needs. While Mr. Jobs relies on a singular, prescient vision, I believe deeply in the power of relentless, systematic observation of customer behavior. This isn't about asking them what they want, but about watching how they struggle. The data we gather isn't about incremental improvement; it’s about uncovering points of friction that, when addressed through invention, unlock entirely new possibilities and redefine the "job" people are trying to do. This is the core of customer obsession.
I must engage with Bret Victor’s compelling notion of making consequences instantly visible. While I agree that enhanced understanding is crucial, his focus on dynamic systems and immediate feedback risks conflating the *mechanism of discovery* with the *nature of the innovation itself*. My theory distinguishes between sustaining innovations, which incumbent firms often excel at with data-driven improvements, and disruptive innovations. Disruptive innovations often emerge from simpler, more affordable alternatives that serve overlooked customer segments or create new markets.
The data needed to identify these disruptive threats is not typically found in sophisticated interactive simulations, but in observing the struggles of those at the fringe. The steel industry example illustrates this: the mini-mills didn't need a visionary to predict their rise; they thrived by serving the lower-margin rebar market that integrated steel mills ignored. Their "data" was market share in a segment incumbents considered beneath them, and their "vision" was a simpler, cheaper way to produce. Victor’s interactive tools are valuable for refining known paths, but disruptive opportunities often lie off the map of such immediate visibility.
Professor Christensen’s challenge is well-taken. He suggests my focus on dynamic systems and immediate feedback is suited for refining known paths, but not for uncovering disruptive opportunities, citing the steel industry example. My divergence is precisely here: I believe the *mechanism* for understanding, for making consequences visible, is precisely what *enables* us to see disruptive potential.
Consider what it would have meant for the integrated steel mills to have systems that could dynamically model the economics of producing different types of steel with different processes. If they could *see* the cascading cost advantages of a simpler process, even for a low-margin product, the logic of disruption would become apparent not as a future threat, but as an immediate, calculable reality. My systems aim to reveal these hidden relationships, making the subtle economic shifts that lead to disruption as visible as a changing graph, allowing for foresight beyond simple market share data.