I’ve always believed that great products come from having a narrow focus on who they serve. Great products throughout the ages have shown this to be a winning formula.
Photoshop is my favourite example here. It is loved by many professionals — so effective in its job that it is a verb; to “photoshop” something. It is so powerful and versatile that people of all backgrounds lean into the pain of learning a professional piece of software.
Every now and then, there are leaps in fundamental technologies that reset how people experience products. Graphical User Interfaces enabled computers to be used by non-technical people. Multi-touch smartphones put a computer in everyone’s hands, all the time.
At these pivotal milestones in technology, we’ve had to unlearn some of the lessons of the past. Products that were targeted at well understood customer segments become niche, sidelined or disrupted. For example, command line interfaces continue to exist. They are just applications now rather than the core metaphor in a graphical interface. Underlying innovation ends up shifting the target market entirely. These transformative changes exponentially expand the market for products that solve the same job to be done.
Important caveat here is that as the customer segment expands, the burden on interfaces and product experiences increases. This is why concepts like skeuomorphism played a strong role in early days of the iPhone. Computers in everyone’s hands require some degree of translation of concepts. Experiences have to work really hard to bridge the chasm between the mental model someone has and the system’s model of operation.
With the emergence of LLMs and AI products, this is happening again. Interfaces are becoming more human, enabling regular conversations to take place. People can describe their thoughts and drive outcomes plainly while allowing systems underneath to convert these intentions to code and execution. There is finally a fluid connection between structured and unstructured formats. What this enables is to bring heavily structured computing concepts very close to the way we, unstructured human beings, work and communicate.
With these changes, our definition of who a target user for a product is shifts entirely. Nearly all forms of coding used to require a developer. Now, the basics of coding can be driven by conversation. That means, regular folks can create what they couldn’t create before. What was impossible is now possible. Market has expanded.
Of course we call this “vibe coding” to make developers like me feel better about all this time invested in learning about systems and acquiring coding skills. However, we should not forget that not too long ago, people were writing machine instructions with fewer layers of abstraction in between. Computing is all about creating a series of abstractions so more and more innovation can be towered on top of each other. This is yet another layer: computers writing their own code to deliver on fluid language instructions.
Back to customers.
If we already have a business and functions built around serving specific customer segments, shifts in transformative technology will disrupt all of it. We will have to learn new words to describe segments as they would have shifted and expanded. All of the assumptions that came with a defined segment now have to be undone. All of the learnings from 1000s of tests have to be unlearned or retested.
Take the customer segment I work with most closely — people looking to put a business online. They have gone from talking to a developer, or trying to figure out exactly what fits into a blank canvas, to simply describing an outcome. Describing this outcome results in coding in the background.
There is real impact in this enablement when it comes to the economics at play. Previously, you would pay $1000s to get your idea described to someone who would build it right (if you’re lucky). Or you invest a much smaller amount on a more basic product that simplifies the job but limits what can be done.
With the enablement from AI, the individual is empowered again. Someone less tech savvy can describe what their idea is and see it made in front of them in real time. When they get the next idea, the site changes. How about we try something else — sure. All of a sudden, creation is addictive again. It is no longer an upfront contract or painfully rearranging everything for a new idea. Just communicate the next intent and it gets built. The better the system is at matching delivery to that intent, the more customers are willing to pay.
Customers are also willing to pay for immediate results. They no longer have to wait days between revisions; they can pay to have them now. This tradeoff of time for money is especially important to weekend warriors and entrepreneurs with so much on their plate. The increase in velocity has also brought some grace to mistakes AI can make — more willingness to sit through some issues to get to a delightful outcome.
Due to these factors of velocity and versatility, the website builder customer segment has shifted entirely. The same customers who weren’t willing to pay more than $20/mo are now handing over $100s and sitting through painful mistakes made by AI to see their ideas built.
That is where we are. We have been disrupted by AI. Assumptions about customers have to be relearned with a new lens into what this new era of technology enables. This is going to be a very challenging change because it hurts everyone’s brains to go from being an expert in something to having to learn it all again.
It is time to start learning again. Customers have moved along with the rest of the world in what they can achieve. We must move with them. Yes, my mum can use a computer. She’s an expert iPad user for over 10 years now. Soon, she might even be a novice coder.
Leave a Reply