Person Product Fit
My predictions for 2025 have led me to something I think is... kinda a big deal. At least it is for me—as someone who's been making a career trying to build product businesses.
So if you're like me, aiming to build a product-based business, then your primary aim has been to achieve "product-market-fit".
In Marc Andreessen's 2007 post, The only thing that matters, he defines product-market-fit as:
Product/market fit means being in a good market with a product that can satisfy that market.
My favorite piece on product-market-fit is Jason Cohen's, where he defines it as having all of the following:
"Easy growth (pull, not push)... High retention... Critical mass."
I've always understood the concept to be: Identify a large group of people who share a very similar problem and solve that problem with a one-size-fits-all solution (a product).
Spoiler alert #1: I'm not here to say PMF is a thing of the past (I'll get to why not).
Spoiler alert #2: I'm not claiming that all software products will go extinct (SaaS—especially already established ones with distribution advantages aint going anywhere)
Spoiler alert #3: If you're a software developer, your career security is just fine (there are opportunities that a lot of us are, frankly, sleeping on).
What's changing?
AI doesn't just make a few tedious tasks easier or faster. It fundamentally changes the way problems get solved.
We already see the pace at which AI (and the tooling around it) is advancing. At some point in the not-so-distant future, AI (and its tooling) will enter mainstream adoption.
Think: Small business owners. School teachers. Local governments. Health workers. Teams and individuals working on all floors of large companies. AI hits the mainstream when every worker in every sector is using AI in their daily workflow just like they reach for the smartphone that sits in their pocket every waking hour.
When people's jobs depend on their competency with using AI at work, that's when, I believe, our game of chasing product-market-fit will noticeably change.
Is good enough... good enough?
Today, when faced with a problem, the business owner (or employee) searches for the best-fit product on the market to buy to solve their problem.
As product makers, we like to think that once we've hit product-market-fit, we've perfectly crushed the problem for every single one of our paying customers. But deep down, we know that's not true.
No—at best, we've mostly solved a problem. We've had the good fortune of stumbling into a market of many customers who experience that pain sharply enough that they're willing to pay for a product that's good enough.
No product is perfect. The products I pay for are always 80% useful, 20% annoying or incomplete. As product makers, we're lucky if we can hit that 80/20 mark. It probably means we've got a damn good business on our hands.
But at some point, "good enough" will become... a problem.
At BigCo, operations manager, Michael, buys a SaaS product called "ProcessRocket" to manage his team's workflow for producing widgets. It's an upgrade from the messy spreadsheets they used to use. But there are things specific to BigCo's operations that "ProcessRocket" doesn't quite nail (and won't because they're too specific). These papercuts frustrate Mike's teammates and slow them down. Customer orders still fall through the cracks.
BigCo recently hired another ops person, Jasmine. Jasmine used AI to architect an alternative to "ProcessRocket". She gave it an internal name, "BigCoOperator". It's designed to manage BigCo's production line in the specific unique way that BigCo creates widget (their special sauce, included).
But Jasmine's vision for "BigCoOperator" ran into a snag. It turns out, some of the requirements are more technical and Jasmine (without programming experience) couldn't figure them out on her own. So she hires a freelance product engineer to fill those gaps (or someday, hires an AI agent instead) to get it to the finish line.
Still—there's friction. Jasmine has 2 teammates, Ben and Lacy. Ben is a more visual thinker, while Lacy is more analytical. When Jasmine rolled out her "BigCoOperator" tool to her teammates, it clicked for Lacy but not for Ben. Easy fix: Tweak the instance running on Ben's machine to fit his personal use-case.
Not only was Jasmine's solution a better overall fit from a workflow perspective, it also costs BigCo a lot less to build and run. And it's future-proof: It can continuously evolve as BigCo grows.
Mike's solution of buying an off-the-shelf product got BigCo's operation to good enough (for now). Jasmine's solution of architecting a custom solution got BigCo's ops to perfect (and future-proof).
So who's getting the promotion? Jasmine, of course.
And which company is winning their market? The company with more Jasmines than Mikes.
Person Product Fit
I believe we, as product people, we need to start to consider a world where products don't serve a market of many. Products serve a market of one.
Individuals will build their own perfect products to solve their own unique problems—even if there are many other people who share a similar-looking problem. Choosing to buy a tool vs. building a tool will become harder to justify from a cost, efficiency, and growth perspective.
The burden of solving a problem will shift away from the products companies and to the individual. Individuals will find their own person-product-fit.
The opportunities
I'm hearing a lot of fear out there, especially from career developers, and aspiring product founders. If even half of what I (and many others) are expecting from the AI wave comes true, what will that mean for our careers or the businesses we hope to start?
I personally feel this uncertainty. I've spent the past 20 years developing my technical coding skills with the aim of building my own products. Where does someone like me, with my building skills, fit into this AI-powered future?
Today, I'm happy to report that my uncertainty has turned into excitement.
I'm not here to tell you where to invest your time, or what business you should or shouldn't build.
I'll only speak to what's piquing my interest in 2025: I see a few waves coming:
Wave #1: More (not less) people will learn to build tools.
As the build-your-own-tools era gets into full swing, more professionals will need to skill up on building tools. Building (actually useful) tools will still require a good amount of technical product chops to pull off.
The difference will be that you don't need to learn how to code by hand. Learning programming syntax and all the internal plumbing of applications becomes far less important than foundational architecture and product management skills.
Yesterday's workplaces were filled with people with intermediate to advanced spreadsheet skills. Tomorrow's workplaces will be filled with intermediate to advanced builder skills.
Wave #2: Technical builders will be in more (not less) demand.
This is where I think the "AI is coming for everyone's jobs!" people get it wrong.
The learning curve and generational turnover of the workforce will be very slow. Before every person and company has the in-house skills to build their own tools, there will be a huge wave of demand to hire people who can help build highly personalized tools.
It's not a new wave. Businesses have hired software firms to build custom internal software for decades. But that used to be only an enterprise thing.
The wave will get (a lot) bigger as it comes to the masses. Small businesses, even individual professionals will hire their own personal assistant, freelancer, or boutique agency to help them build private, personalized tools.
Product-market-fit, in the traditional sense, will become even harder to achieve, particularly for smaller, bootstrapped product makers.
The companies that achieve the next wave of product-market-fit will be those that help more people achieve person-product-fit.