The Product Was the Fixed Thing

Last month I built a capital matching engine. Not a spreadsheet with formulas, an actual tool that matches investor profiles against specific criteria, then maps them to allocation opportunities based on a dozen variables unique to how our fintech marketplace operates. Two years ago, this would have required a dedicated engineer working for several weeks. I built it in a few days as a side project. I'm not an engineer.

Before that, I built tools that augment our CRM to ingest unstructured data from sources that no off-the-shelf platform knows how to parse. Last year this workflow automation would have required a significant integration project and expense. None of this is our core product. Crib Equity is a fintech marketplace. Building internal tools isn't our core competency and it's not the product we're selling. But the off-the-shelf tools we're paying real money for can't do what we specifically needed. So we built it. Because building custom software became cheaper than paying a manual labor tax or spending months making it sort-of-work.

I'm not special. This is happening everywhere, right now. The barrier between idea and execution has collapsed. And the consequences of that collapse are far bigger than most people realize.

The Savile Row Suit at Off-the-Rack Prices

For most of commercial history, there have been two options for almost anything. Bespoke — perfect fit, exorbitant cost, reserved for people who could afford a tailor on Savile Row. Or off-the-rack — imperfect fit, accessible price, available to everyone. Then an entire middle market exists to close that gap. Made-to-measure. Semi-custom. “Tailored fit.” All of it is an attempt to approximate bespoke without paying for it.

Software has worked similarly. On one end: Goldman Sachs, with 10,000 engineers and a $1.9 billion annual technology budget building internal tools that do exactly what Goldman needs. Custom-fit. Billions of dollars a year. On the other end: Salesforce, HubSpot, Workday provide off-the-rack products designed for the broadest possible market. You buy them, configure them, hire consultants to make them sort-of-work, and adapt your workflows to match their constraints. Not because they fit well. Because building custom was too expensive or too time intensive.

AI collapses that spectrum. The Goldman Sachs internal tool built by a dedicated engineering org with unlimited budget is now available to a 5-person company at the cost of an API subscription and a few days of focused work. Bespoke capability at off-the-rack economics. Not a slightly better fitting off-the-rack suit. An actual Savile Row suit, cut to your exact measurements, at a price anyone can afford.

This isn't a marginal improvement. It's a category collapse. And it changes who software is built for, how it's built, and most importantly — how it's found.

The Inversion

The conventional read on this moment is: “AI makes building easy, so distribution matters more.” That's true. It's also shallow. It skips the structural shift underneath.

For thirty years, software has operated on a simple premise: the product is the fixed thing, the user is the flexible thing. A company builds a product for the broadest possible market. You buy it. Then you spend weeks configuring it, learning its quirks, building workflows around its limitations. You adapt to the software. This was always a compromise, but the economics were clear. Building custom was expensive, so you lived with imperfect fit. You bought the off-the-rack suit and took it to a tailor who could only adjust the hem.

That premise is broken.

When building custom becomes cheaper than configuring generic, the cost of fit shifts. It used to be borne by the user. Now it's borne by the product. Software adapts to you, not the other way around. That's not a feature improvement. It's a structural inversion of the entire product-user relationship.

Watch what this breaks.

The configuration tax disappears. HubSpot charges $800 a month and then you spend three months hiring a consultant to make it sort-of-work for your specific sales process. That entire model: the implementation partner, the onboarding specialist, the admin you need on staff to maintain it was a workaround for the fact that custom software was too expensive. The workaround is no longer necessary.

The “everything app” loses its moat. Salesforce, ServiceNow, Workday. These companies became giants because integration was expensive, customization was prohibitive, and switching costs were high. It was rational to pick one sprawling platform and live within its constraints. But when building is cheap and AI can handle integration and data migration, those structural advantages evaporate. “Does everything adequately” loses to “does your specific thing perfectly.”

Enterprise procurement gets disrupted. An 18-month buying cycle for a platform that needs another 6 months of implementation and training starts to look absurd when a team lead can describe what they need and have a working tool by Friday.

Here's what this looks like in practice. Take a craft brewery running a taproom. They need point-of-sale for the tasting room, mug club and subscription management, direct-to-consumer shipping across states with different alcohol regulations, federal alcohol compliance reporting, and inventory tracking tied to production batches. No horizontal platform does all of this well. Toast was built for restaurants. Square is generic retail. Shopify handles e-commerce but doesn't understand the regulated distribution networks or state-by-state shipping compliance that govern alcohol sales. So the brewery stitches together four or five tools, builds workarounds in spreadsheets, and lives with the gaps.

Now someone with domain expertise in craft beverage operations can build the exact system these producers need. One tool. Every workflow understood. No configuration required. The Savile Row suit, cut to the brewer's exact measurements, at a price a 10-person operation can afford.

Multiply that by every underserved vertical, every industry with specific regulatory requirements, every customer segment that's been forcing itself into software that was designed for someone else. That's the scale of what's shifting.

The Discovery Problem

But here's where the consensus take that “building is easy now, so just build” misses the real challenge.

When anyone can build a perfectly tailored product for any niche, the supply of products explodes. And we've seen this movie before. It's exactly what happened to content.

When YouTube made video publishing free, the bottleneck wasn't production. It shifted to discovery. When Substack and podcasting tools made it trivially easy to publish written and audio content, the problem wasn't “can I create?” It became “can anyone find what I created?” The App Store has nearly two million apps. A quarter of them have fewer than a hundred downloads. The signal-to-noise ratio collapsed, and the winners are the people who have distribution independent of the platform.

The same thing is about to happen to software. AI makes it possible for anyone with domain expertise to build the perfect tool for a specific customer segment. That's the supply-side revolution. But it also means the market is about to be flooded with products. Many of them good, solving real problems for real segments. The question now shifts from “can you build the right product?” to “can the right customer find it?”

This is the part that most builders haven't thought through. The content explosion didn't just mean more content. It meant that the economics of attention fundamentally changed. The traditional discovery channels are simultaneously eroding as users shift to AI-powered answer engines. The playbook for getting found is breaking at the exact moment when getting found matters more than ever.

So if distribution is the new bottleneck, what does the distribution playbook look like in a world of infinite product supply?

The Creator Playbook

It looks like content creators. Not enterprise SaaS go-to-market.

That might sound like a stretch. But hear me out.

When content creation tools became free, the creators who won weren't the ones who made the “best” content in some absolute sense. They were the ones who owned a niche and built trust within that niche before they ever tried to monetize. The underlying lesson: in a world of infinite supply, the relationship with your audience is the moat. Not the content itself.

Two specific mechanics transfer directly to software builders.

Niche authority is the distribution channel. A content creator with 50,000 deeply engaged subscribers in a specific domain consistently outperforms one with two million casual followers. Micro-influencers with niche audiences generate up to seven times the engagement of accounts with millions of followers. The reason is trust. When someone has demonstrated real expertise in your specific world, their recommendations carry weight that no ad campaign can replicate.

The same dynamic applies to product builders. The person who's spent years in craft beverage operations and can speak fluently about federal alcohol compliance and taproom-to-DTC fulfillment logistics has a distribution advantage that no horizontal SaaS company can buy. Their domain expertise isn't just what makes them able to build the right product — it's what makes the right customers find and trust the product. The expertise is the signal that cuts through a market full of noise.

Audience before product. The traditional SaaS playbook runs: build product, then find customers, then iterate based on feedback. The creator economy proved a different sequencing works better in saturated markets: build credibility in a specific domain, develop a deep understanding of what that audience needs, then ship the product directly into a community that already trusts you. The audience is the moat. The product is the expression of the moat, not the moat itself.

This isn't about “personal branding” in the influencer sense. It's about the structural reality that when product quality becomes table stakes the differentiator is whether the right customer knows you exist and trusts that you understand their problem. Building that trust takes time. It compounds. And it's very hard to replicate once established.

The sequencing implication is significant. Product builders in this new landscape need to think more like domain experts building a following. Write about the problems in your industry. Build a community around a specific operational challenge. Demonstrate that you understand the pain at a level that generic software companies never will. Then build the product. The distribution is already solved.

What This Means for Big Software

This is where incumbents should be paying attention.

The honest version of the incumbent argument goes like this: “We have data moats and switching costs that protect us.” And that's partially true. Incumbents sit on years of customer data that create real advantages. But AI can now extract insights from data in ways that weren't possible before, combining sources, finding patterns and making connections that used to require the kind of proprietary dataset only an incumbent had. The moat isn't gone, but it's starting to dry up.

When the data advantage narrows, what's left is lock-in. But these barriers also erode when the cost of switching drops. When an AI can migrate your data, replicate your workflows, and stand up a replacement system in days instead of months, the switching cost argument starts to look like a countdown.

Large enterprise software companies face a specific structural challenge: their architectures were built for configuration, not adaptation. HubSpot's product is literally a configuration engine — settings, workflows, custom fields, automation rules, all designed for the user to shape the product to their needs through menus and toggles. That worked when the alternative was building from scratch. It doesn't work when the alternative is a product that already fits without any configuration at all.

Their go-to-market motion has the same problem. Enterprise sales, implementation partners, training programs, customer success managers who help you configure the thing you already bought. That entire structure exists because the product doesn't work out of the box.

But incumbents aren't helpless. They have real advantages in feature depth, massive distribution, brand trust, and existing customer relationships. The path forward isn't to pretend AI isn't happening. It's to use these same tools to build the AI layer that eliminates the configuration tax from the inside. Make the product adapt to the user automatically instead of asking the user to configure it manually.

The ones who don't will watch their most valuable customer segments peel away to domain experts with AI tools who build exactly what a specific segment needs, and who already have that segment's trust because they've been operating in that world for years.

The Window

There's a timing dimension to this that matters.

Early YouTube creators had a structural advantage over later entrants. Not because they were better at making videos, but because they built audiences before the platform was saturated. They established authority in specific niches when the competition for attention was low. By the time everyone else showed up, the early movers had compounding advantages. Subscriber bases, algorithmic familiarity, and brand recognition all functioned as durable distribution.

The same window is open right now for product builders. AI has made it possible to build bespoke software for any niche. Most people haven't fully internalized that yet. The supply explosion is coming but hasn't peaked. The builders who establish domain credibility and community trust now — before the market is flooded — will have a distribution advantage that compounds over time and becomes very difficult to displace.

This is the Savile Row problem. The tailored suit is available to everyone. The fabric is essentially free. The craftsmanship can be replicated by machines. When everyone can make a perfect suit, the tailor who wins isn't the one with the best stitching. It's the one who has a relationship with their clients before they ever walk through the door.

The question for every product builder, every software company, every enterprise incumbent is the same: in a world where anyone can build the right product, can you reach the right customer? Are you building that relationship now?