A SaaS is a Set of Promises

I built a SaaS application I was hoping to supplement my income during retirement. It's called SuiteCollector, and it's the most ambitious project I've ever completed. Yes. I think it's ready for production. Ready for customers. Ready for general availability.

And I'm not going to do anything more with it at the moment because running a SaaS application is a series of promises you make to your customers. As a solo developer thinking of turning this project into a business, I don't think I want to make those promises during retirement, when I'm supposed to be enjoying friends, travel, golf, reading, writing, games, personal projects, life.

What are these promises? They're all of the things it takes to run a business. SaaS companies often have entire teams dedicated to each of these promises so thinking of tackling all of them by myself was quite an eye-opener.

Let's get into it.

The domain

It all starts with the domain. It seems simple at first. You find a cool name which suits your project and you buy it. But it's the front door for your entire business. It's the face of the application, the API, and email, all of which resolve through DNS I own personally. Miss a renewal, fat-finger a DNS record, let a registrar flag your account, and the product doesn't degrade gracefully. It vanishes.

When it vanishes, it isn't my weekend project that's down, it's someone's accounts-receivable operation. One missed calendar reminder for me is all that stands between a customer and an outage.

That's a strange amount of weight to hang on a line item that costs $15 a year.

The website

And the site itself is never really finished. It runs beautifully on my laptop, and it runs beautifully for the first handful of customers, which is exactly the trap. The query that returns instantly against a few hundred records crawls against a few hundred thousand.

The background job that sipped memory in testing balloons in production until the process gets killed and quietly stops doing its work. And nothing tells me it stopped. Caches go stale, sessions pile up, a dependency I patched for a security fix changes behavior three layers down where I wasn't looking.

Scaling isn't a single event I prepare for once; it's a moving target I'd be chasing for as long as the product lives, profiling slow endpoints and hunting leaks at exactly the moments the app is busiest, which is to say, the moments it matters most.

Every new customer makes the software more valuable and more fragile at the same time.

Hosting

Once you get your first customer, uptime changes from a metric to an obligation. The server doesn't care that it's 2am my time or that I'm out on the golf course. Patching, security updates, certificate renewals, the disk that quietly fills up, the dependency that ships a breaking change.

I'm the entire operations team...on call...forever. I've worked in full ops environments before supporting hundreds of servers. Infrastructure outages never seem to happen when I'm sitting around doing nothing. Cloud computing and managed platforms help, but they don't make the problem not mine.

Payment providers

The moment I start taking money, I've signed up for a second business I didn't really want to run. Failed charges, expired cards, disputes, chargebacks, refunds, proration, the subscription which didn't renew correctly, the sales tax nexus I didn't know I had tripped.

Services like Stripe make the happy path easy but I still have to deal with all of those exceptions. When the money touches some other company's books, mistakes aren't bugs. They're finance problems, and finance problems have a way of becoming legal ones.

Email delivery

This one is deceptively brutal, and it's worse when the whole point of SuiteCollector is to send email on someone else's behalf. Deliverability is a dark art which includes things like SPF, DKIM, DMARC, IP reputation, bounce handling, or the blocklist I'll land on for reasons I'll burn a day diagnosing.

A dunning email that lands in a spam folder isn't a missed notification. It's a customer's invoice that didn't get paid and now it's my fault. It was sent through a mail server I don't control, governed by rules Google and Microsoft seem to change without telling anyone.

Support

For a solo dev, this is the one that could actually break me. I would be the entire support desk. Every question, every confused user, every "it's urgent" at 4:45 on a Friday. That's me.

There's no tier one to absorb the easy stuff and no colleague to cover while I'm out. Support doesn't respect my timezone, my vacation, or my capacity to have a bad day. A single frustrated customer can quietly eat an afternoon I planned to spend elsewhere.

Backups

Underneath it all sits the thing that would keep me up at night. I am now the custodian of another company's data. Many companies. Backups that run aren't backups unless I have actually restored data from those backups on a schedule with a tested recovery path.

Retention policies, offsite copies, the quiet dread of wondering whether last night's snapshot is any good. Lose a customer's history and I haven't had an incident, I've had a catastrophe with my name on it.

The LLC that doesn't do what you think

I set up an LLC, and for a while that felt like the answer to all of this. It isn't. An LLC is a container for liability, not an eraser. It changes what's on the hook in some situations, but it doesn't make the obligations disappear, and as the solo owner I'm standing behind most of them anyway.

Start with the contracts. When you sell to real companies, their legal and procurement teams don't sign my friendly little terms of service. They send back an MSA with indemnification clauses, warranty language, a liability cap, security commitments, sometimes an SLA with actual penalties attached. I'm just one person. I either agree to their contract modifications or risk losing that customer. The LLC caps some of that in theory, but the promises have my signature under them.

Then there's the veil. Courts can look straight through an LLC when the owner doesn't keep it scrupulously separate. Commingled funds, one bank account doing double duty, the business run as an extension of myself. Solo operators are exactly the profile that drifts there without noticing. And the shield doesn't cover my conduct anyway. If I'm negligent, in any way, mishandled data, missed a duty I owed, I can be personally on the hook no matter whose name is on the door.

And sitting over all of it is the data. Hold other companies' records and you inherit breach-notification laws, privacy regulations, and contractual obligations that trigger the instant something leaks. The cost of a real breach (forensics, notification, legal defense, regulatory exposure) can dwarf everything the product ever earned. Cyber and errors-and-omissions insurance help, but they cost money, carry exclusions, and are one more thing to keep current, alone.

Here's the part that finally landed for me: I don't have to lose to lose. Even a dispute I'd eventually win still costs me the legal bills, the months, and the stress of defending it. The LLC is never going to hand me a clean night's sleep. It just moved where the worry lives.

The part that made it all clear

None of these is hard on its own. I've done all of them. That's sort of the point. Individually they're manageable, and collectively they form a job that never clocks out, carried by one person, on behalf of businesses whose operations depend on you keeping every promise at once.

The upside has a limit, one modest product's revenue. The downside doesn't, burnout, liability, and a single point of failure that turns out to be a human being who occasionally wants to go play golf.

None of this is an argument against SaaS itself. Every burden on that list weighs what it weighs because it lands on one person. When you split them across a team, they turn back into ordinary jobs. Someone owns the infrastructure, someone answers support, someone runs billing, a lawyer sits on retainer, and a real policy absorbs the tail risk instead of my personal balance sheet. That's what a company is for. It spreads the entire load of work that would eventually flatten a solo owner. Plenty of people should start a SaaS, and the economics can be extraordinary precisely because software scales in a way an hour of my time never will. The ones who do it well just don't do it alone.

I loved building SuiteCollector. I just don't want to be the thing it depends on. And once I could say that out loud, a lot of decisions got easier.