--- title: "Micro-SaaS is dead, so I built custom software to replace a client's $200 a month tool" description: "The \"micro-SaaS is dead\" debate is real. Here's how I ripped out a $200 a month tool and replaced it with custom software a client now owns, built in a week." author: "Josh Green" section: "Build It Yourself" published: 2026-07-29T16:31:19.996Z updated: 2026-07-29T16:31:19.996Z canonical: https://lettherobotsdoit.com/articles/build-it-yourself/micro-saas-is-dead-so-i-built-a-client-custom-software-to-replace-their-200-a-month-tool --- # Micro-SaaS is dead, so I built custom software to replace a client's $200 a month tool There's an essay going around called "Micro-SaaS Is Dead. Service with a Software Replaces It," by Adrien Gonin. It hit Hacker News and picked up 54 points and 79 comments, which for a niche business-model take is a real fight. Half the room nodding, half the room calling it nonsense. I'm on the nodding side, and I've got a receipt for it. Last month I ripped out a $200 a month tool for a client and replaced it with a small piece of custom software they now own outright. Built it in a week. So let me walk you through that instead of arguing in the abstract. ## What "micro-SaaS is dead" actually means The claim isn't that software is over. It's that the tiny standalone tool, the $15 to $200 a month app that does one narrow thing, has lost its footing. Two things happened at once. Anyone can now prompt a working version of your clever little tool by the weekend, so supply exploded. And the people who used to pay for those tools are quietly canceling them and building their own instead. Gonin's phrase for what replaces it is "Service with a Software." The software isn't the product you sell. It's a private tool built so tightly around one business that nobody can compete with it. That's basically my whole job, so the essay just put a clean name on something I already do. ## The $200 a month tool I ripped out The client runs a home services business. They were paying a booking-and-reminders SaaS a little over $200 a month on their plan tier. It did three things they cared about: an online booking page, automated text reminders before each appointment, and a review request text after the job. Useful stuff. But they were on the tier that unlocked those three features, and that tier also came with a marketing suite, a "reputation dashboard," lead scoring, and a dozen other things they never touched. Classic bloat. They were renting a whole office building to use one desk. And none of it was theirs. Their customer list, their booking history, their appointment data, all of it lived inside someone else's app, exportable only when the vendor felt like it. ## What I built instead, in a week I built them a small custom app that does exactly their three things and nothing else. A booking page that matches their brand. A tiny scheduled job that texts appointment reminders through Twilio. A follow-up text that asks for a review after the job is marked done. That's it. No dashboard nobody opens. No modules they'll never touch. It runs on a cheap server for about 15 dollars a month, plus a few cents per text they actually send. So a $200 a month line item turned into something closer to 25 or 30 dollars all in. And the important part: they own the code and the data. If they ever fire me, they keep everything. It doesn't shut off because a subscription lapsed. To be clear about what this is, because people love to call anything with a trigger "automation." I did not wire up some no-code app that snaps other apps together. I wrote them real software, right-sized to their actual workflow, with the reminder logic built in. ## Why the custom one wins On paper, the SaaS was the safer bet. Battle tested, supported, feature rich. But that's exactly the trap. A product has to fit a thousand businesses approximately, which means it fits none of them exactly. My little tool only has to fit one business perfectly, and that's the whole edge. Their reminder text sounds like them. The booking flow asks the two questions they actually need answered and skips the ten they don't. When they want a change, it's an afternoon, not a support ticket into a void. The thing is overfit to one business on purpose, and overfitting used to be a sin because it didn't scale. It scales fine now. I standardize how I build these, not what comes out, so each client gets a one-off that took a fraction of the old effort. ## If you're paying for a micro-SaaS right now Pull up your subscriptions and look for the ones where you use maybe 20 percent of the features and the bill still climbs every year. Those are the candidates. If a tool does two or three things you actually need, there's a good chance it can be a small custom app you own instead of a fee you rent forever. Micro-SaaS as a business to launch might be dying. But micro software built for one company, quietly making that company impossible to compete with, is very much alive. That's the part of the debate I'd put money on.