Saas or Vibecoding?
By Mark Cohen
I keep encountering this conversation. Twice this week... One over coffee, one over teams. If a team can vibe code a working SaaS tool in a short time, why keep paying for the SaaS version. Good question. Nobody in either conversation answered it with a number.
So I built a calculator.
Yes, I vibe coded it :)
A five-year TCO calculator that compares the vibe-coded build-and-maintain path against the SaaS subscription path, using the inputs that actually drive the total cost answer: team cost, delivery time, maintenance rate, hosting, subscription price, implementation efficiency. It's live and free at tco.markcohen.com - one page, plug in your own numbers (you can download in pdf or xls too)
Run it on a sample solution build (2.5 person team, $110k loaded cost, four months to build, 60 licensed SaaS users at $50 a month) and SaaS comes out $77,757 cheaper over five years. That's a 21% gap in SaaS's favour. Change three assumptions, your subscription price, your delivery time, or your maintenance rate, and the gap closes or flips entirely.
That's the actual finding: the answer varies by scenario. It's assumption-driven, and most people arguing about this online have never written their assumptions down.
So which is cheaper? It depends... what have you priced in? Before you cancel the subscription or greenlight the build, have you considered:
Who maintains it in year three, when the person who built it has moved to another team (or will it still even be needed then?)
Whether "maintenance" for your vibe-coded tool means security patching, or just feature requests
Whether all 60 licensed users need the SaaS features, or 20 of them are paying for permissions nobody touches
What the vendor is pricing in that you're not: uptime, backups, compliance, someone else's on-call roster
Whether the four months your team spends building it is four months not spent on something with a bigger return - there's almost always an opportunity cost
I'm not going to tell you which way to go, and I'd be a bad witness anyway. I sell SaaS for a living. I also vibe code my own tools most weekends, so I've got a foot in both camps. The calculator also won't settle the argument. Writing your assumptions down will help you make a more informed decision, which is more than the argument online has managed so far.