Why do people buy custom suits?

If you’re here for fashion advice, I’m sorry. This won’t be helpful, but you might as well stick around and learn something new about another sexy world: enterprise software. To answer your immediate question “Then what the hell is this going to be about?” I will get to that. Just humor me while I do my best Jon Hamm impression from Mad Men to tackle the question in the title.

People buy custom suits because they make them feel good. Off-the-rack suits are designed for theoretical people who don’t exist: average shoulders, average arms, average everything. But you’re not theoretical. Your shoulders are a bit broader, your right arm is slightly longer than your left, and your knees are a little bulgy, who cares? You choose a custom suit because it honors these nuances that make you you, instead of forcing you to squeeze into someone else’s idea of how a person should look. The suit becomes proof that you notice details others ignore. It tells clients you respect their time, tells friends you showed up on purpose, and tells you, in the mirror each morning, that you control your day.

So, how does this relate to enterprise software? The same problem that plagues suits appears when you buy software off the rack. Your CRM provider may say, “Don’t worry, we have thousands of configurations; it’ll feel like your own software!” Come on! That’s no different from the sales guy at Men’s Wearhouse insisting, “We have that suit in black or red.”

This, my friends, is the Software Fit Problem. Before those cheeky folks at OpenAI popularized the LLM (remember the Plinko-machine thingy?), we all had to live with it. Not anymore. This piece explains why the problem exists, what it affects without our noticing, and how it will change in the next ten years, complete with another analogy or two.

How did we get here?

Every software starts as a sharp tool built for one clear pain. A few engineers work shoulder-to-shoulder with early users, so every control lands where it should and each screen feels obvious.

Then growth hits. Investors push for a larger TAM, marketing invents personas, and sales sells “flexibility.” Features spread, toggles multiply, and the codebase stretches to cover edge cases no one on the original team will ever see. Intuitive flows get buried under onboarding videos, which soon hand off to customer-success reps. By the time the logo shows up in an enterprise quadrant, the product is a bundle of loosely stitched add-ons.

For the vendor, that’s success. They’ve reached product-market fit, meaning the average user in a big market can muddle through well enough to keep paying. Revenue climbs, new segments open, and the roadmap tilts toward scale over specificity. Meanwhile, each new customer shows up with the problem they actually have, not the brutally generalized one featured on the vendor’s product page.

You see, software has to be built to fit the mean, and as the surface area grows, the mean drifts further from the specific problem you're actually facing.

Why does it matter?

This matters for one reason, and sorry, Venture Capitalists, but I don’t want to live in a world where product-market fit is the finish line. I want a world where product-firm fit is the standard. You should too!

I promise this isn’t just me riffing on custom-suit metaphors or product-market-fit puns. Misfit software is inflicting real issues across organizations, problems we’ve either learned to tolerate or chosen to ignore.

Things like when software doesn’t match real work, and teams re-key the same data across tools, over and over and over. Sure, this wastes time, but have you ever done it? It is so fucking boring! I would rather listen to my sister sing Umbrella by Beyonce. (shout-out to Bella; for reference, last night the police were called on her for singing on her apartment balcony).

So you say, “Who cares? You need to earn your keep!” Well, what about shadow IT? Yeah, you know the impatient operator who spins up unapproved SaaS tools and runs your proprietary data through them. Oh, you don’t have any of those at your company? Sure, John from Marketing definitely doesn't have ChatGPT pulled up on his private laptop…

What happens next?

First, I trademark Product-Firm fit, because start-up twitter is coming for me when they read this.

Next, and far more important, I finish this because it is 2 AM, and my mom would be pissed if she knew I was up this late.

Then, everything changes. It will not happen overnight, and maybe not this year or the next, but the buggers at OpenAI and the other foundation labs (the ones who build the Plinko boards) have released technology that makes per-company, per-team tools feel inevitable.

In the long tail, as AI continues to improve, we will eventually evolve to the point where models can create these tools from scratch using your domain expertise, and your context. For now, we can spark the same magic by focusing on a platform built for a specific industry.

You see, that matters because if we limit ourselves to a small, constrained set of use cases, we can equip AI with the right skills for the right industry and make real magic. I promised one more analogy, so here is one to make this click.

Right now, to achieve this, you need to build a football. Try playing tennis with that football and you’ll end up with, well, probably a broken racket. But put that football in the hands of twenty-two guys on a hundred yards of green turf and you get a whole society of people who make fantasy rosters of other men.

This is where the industry is going, and it should excite you; in fact, it should bring thoughts of grandeur. Imagine you open up your laptop and you have an experience built for exactly how you do work. There is no tab hopping, no copy and pasting, just you in flow, doing better, happier work than ever before.

Thanks for reading! That was quite fun to write, and I hope it was at least half as fun to read. If you’re wondering why I kept talking about Plinko boards, check out last week’s piece. If this one entertained you, stick around for the next: I’m torn between writing about how bad my sister’s singing really is and sharing how to squeeze the most out of ChatGPT. We’ll see which topic wins!