Build vs Buy SaaS: 5 Subscriptions I Deleted This Year (Simply Explained)

Most advice about software subscriptions falls into two camps.

One side says: just buy everything, never build your own, your time is too valuable. The other side says: build everything yourself, own it all, control your destiny.

Both are wrong.

This year I cancelled five software subscriptions from the DTC fashion brand I run in San Diego. Together they were costing me a little over $18,000 a year. I replaced each one with custom tools I own.

Here is the rule I followed: own the parts that are specific to my business, keep renting the parts that are hard for everyone.

I did not build my own credit card processor. I did not build my own email system. Those are like electricity or running water. A company that has spent ten years perfecting them will always beat me. I kept paying for those.

What I stopped paying for was software that took my own business rules, wrapped them in a nice screen, and charged me every month to use them.

How to know when to stop paying

Before the examples, here is the simple test. Three things tell you it might be time to build instead of buy. You usually need two of them happening at once.

First: you are fighting the software every month. You bought a tool but you keep fixing its mistakes by hand. That means it does not actually fit your business.

Second: the price keeps climbing but the work does not. The bill grows as you grow, even though the actual job stays the same size. You are paying for their profit, not your result.

Third: they are holding your own data hostage. Your customer records and your history live inside their system, and you have to ask permission to get them out. You do not own your business at that point. You are renting it.

Two or three of those happening? Time to look at building.

The five I cancelled

The bookkeeping app. It could not handle how my business actually tracks money. Inventory, multiple currencies, pre-orders that ship weeks after the customer pays. Every month I exported everything and fixed it by hand. When you pay for software and then correct its work yourself, you are paying twice. I replaced it with a system where the numbers literally cannot go out of balance. The computer rejects any entry that does not add up. And to be clear, I did not fire my accountant. I still pay a human for tax strategy and judgment. I just stopped paying software to do data entry badly.

The sales-tax filing service. It charged me more for every state I sold into. But when I looked at the actual work, it was three simple steps. Pull the sales numbers, sort them by state, file them. That is a small program, not a $200-a-month service. Honest downside: tax rules change, and now I am the one who has to keep up. I decided that was manageable at my size. At ten times the complexity, I might have kept paying.

The returns app. This one held my own customer return history hostage. I wanted returns to connect directly to my inventory, so a returned shirt goes back into available stock instantly. I could not get my own data out cleanly to make that happen. I replaced it with an AI-assisted returns system that writes straight to my own records. That saved about $3,600 a year, but the real win was having everything in one place. No more "the returns tool says one thing and inventory says another."

The loyalty points app. It forced my rewards program into its templates. But how I reward my best customers is a competitive decision, not something to hand to a vendor's settings page. A reward point is basically an IOU on my books. Price it wrong and I am either giving away profit or offering rewards nobody wants. I built my own version tied to my real numbers. Honest downside: the app had a polished, trustworthy look for customers. Building something my customers trusted just as much took real design work and a few failed tries.

The fulfillment middleman. This was the easy one. A tool sat between me and my warehouse, charging me a fee on every single order, slowing things down, and breaking on anything unusual. So I deleted it and connected directly to the warehouse myself. Sometimes the software is not something you need to rebuild. It is just a tollbooth between two systems that could talk to each other directly.

The part nobody tells you

Cancelling those five saved me over $18,000 a year. That money stays in the business now, forever.

But building is not free. The moment I turned on each replacement, I owned it. The bugs are mine. The security is mine. If my tax filing goes wrong now, that is my problem, not a vendor's.

There is another hidden cost. If I disappeared tomorrow, someone has to understand 22,000 lines of code I wrote. Part of what a subscription buys you is that the knowledge lives inside a company, not inside one person's head.

Here is the honest truth this whole story depends on. The math only worked because I can build fast with AI, and because I run my own brand. I understand the business rules because I wrote them. I can build the tools because that is my job.

For a 50-person company with nobody in-house who can build, at least three of these five should have stayed as subscriptions. The bookkeeping, the tax filing, and the loyalty program carry too much risk without someone who can own the code. I deleted them because I could carry them. Most companies cannot, and I am not going to pretend otherwise.

How to run this on your own business

Go down your list of software subscriptions and ask three questions about each one.

Are we fixing its mistakes by hand every month? Is the price climbing while the work stays the same? Can we get our own data out, or are we renting access to it?

Two or three yeses means it is worth a closer look. One yes, leave it alone. Zero, you are happily married. Keep paying.

But here is the thing that matters most. Knowing what to cancel is the easy part. The hard part is having someone who can build the replacement and keep it running without creating a bigger mess than the subscription you killed. I have seen companies replace a $300-a-month tool with a $40,000 mistake.

That gap is exactly what I close. I go through a company's software, find the subscriptions genuinely worth cancelling, and build replacements that hold up in the real world. Not a slideshow of ideas. Working systems.

Ready to bring AI leadership into your company?

I work with a small number of companies at a time. If you're serious about AI, apply to work together and I'll review your application personally.

Apply to Work Together