The break-even point: when custom software starts costing less than your subscriptions

Most people compare the wrong two numbers. They put a build quote next to one month of SaaS and conclude the build is expensive. Of course it looks expensive. One is a purchase and the other is a rent payment.

The honest comparison is total cost of ownership over a fixed window. Pick five years. Five is fair because most business systems live longer than that, and because five years is roughly how long a SaaS vendor needs to quietly double your bill through price increases and seat growth.

On one side you have rent that never ends and rises every renewal. On the other you have a purchase, plus hosting, plus upkeep. Ownership means the asset is yours. You can sell the company with it, change it without asking anyone, and stop paying anyone who annoys you.

Open your card statement, not your memory. Almost every team we talk to underestimates their stack by 30 percent or more, because the spend is spread across departments and nobody owns the total.

Here's what to count. Project management seats. CRM seats, including the pricier sales tier that only six people need. ERP or accounting, plus its modules. The integration platform holding it all together. E-signature. File storage overages. Reporting or BI add-ons. Sandbox environments. Premium support. The implementation partner you still pay for admin work.

Run a hypothetical. A 40-person company pays $25 per seat for project management, which is $1,000 a month. Twenty-five people are on a CRM at $100 a seat, another $2,500. ERP and accounting run $1,800. An integration tool costs $500. Add-ons and storage are $400. That's $6,200 a month, or $74,400 a year.

Now apply a 7 percent annual increase, which is conservative for this market. Year one is $74,400. Year five is about $97,500. The five-year total is roughly $428,000. And that assumes you never hire anyone.

Amortization means spreading a one-time cost across the months you'll use the thing. It's how you compare a purchase to a rent payment without lying to yourself.

Take whatever a build is quoted at. Divide it by 60 months. That's your equivalent monthly cost over five years. Then add hosting, which for an internal business system is usually hundreds of dollars a month, not thousands. Then add an upkeep budget. A common planning rule across the industry is 15 to 20 percent of the original build cost per year for changes, dependency updates, and support.

Add those three numbers together and you have a monthly figure that sits next to your subscription line honestly. If the combined number is lower than what you pay now, you've already crossed break-even on day one. More often it's higher for the first year or two, then falls below as the build cost finishes amortizing and your SaaS bill keeps climbing.

The crossover point is the number that matters. Ask any vendor or builder to show it to you as a chart with two lines. If they can't, they haven't done the math either.

Per-seat pricing is the single biggest reason companies outgrow their stack. It ties your software cost to your headcount, which is backwards. Your fiftieth employee doesn't make your CRM more expensive to run. The vendor just charges you more because they can.

A system you own costs the same at 40 users as it does at 90. Hosting creeps up with data volume and traffic, but that curve is gentle. The line is nearly flat while your subscription line climbs with every hire.

Back to the hypothetical company. Say they grow from 40 to 65 people over five years, which is normal for a healthy mid-sized firm. Their per-seat lines grow about 60 percent on top of the annual price increases. That $428,000 five-year projection becomes closer to $600,000. The build didn't get more expensive. Nothing about it changed.

There's a second saving that's harder to see on a statement. Every connector between two SaaS tools is a thing that breaks, and somebody on your team spends hours a month babysitting it. When the data lives in one place, those hours disappear. Count them at a loaded hourly rate and they're usually worth more than the integration platform subscription itself.

It has maintenance costs, and they're not hidden if the person selling you the build is honest about them. Software needs updates. Libraries get security patches. Browsers change. Your business changes and the system has to follow.

What ownership changes is who controls that spend. With a subscription, maintenance is bundled into a price you don't set and can't refuse. The vendor decides when to raise it, when to sunset a feature you depend on, and when to force you onto a new interface your team has to relearn.

With a system you own, maintenance is a decision. Some quarters you spend nothing because nothing needs changing. Some quarters you spend a lot because you're adding a whole workflow. You can also change who does the work. If your builder disappoints you, hire someone else, because the code and the data are yours.

The failure mode is real though. Companies that budget nothing for upkeep end up with a system that slowly rots, then blame custom software. Put a line in the annual budget. Treat it like maintaining a building you own instead of a lease you renew.

Plenty of situations where building is the wrong call, and pretending otherwise would waste your money.

Don't build accounting, payroll, or tax. Those vendors spend enormous sums keeping up with regulation across jurisdictions, and that compliance work is the actual product. Same with email delivery, payment processing, and e-signature. Buy those, connect them, move on.

Don't build if the platform is how your clients work with you. If half your customers already live in a shared portal from a vendor and like it, migrating them is a real cost with real relationship risk. Sometimes the right move is a custom system internally that publishes to the tool your clients use.

Don't build if your team is under about 15 to 20 people on cheap tools. The subscription total is too small for the arithmetic to work. Break-even at that size might be eight years out, which is too far to plan against.

Don't build if your process changes every quarter. Custom systems encode how you work. If how you work isn't settled, you'll pay to rebuild the same thing three times. Get the process stable on paper first.

And be honest about migration. Moving twelve years of records, file attachments, and audit history out of a hosted platform is a scripted export and a cleanup project, not a copy and paste. Budget for it as a line item. It's usually the most underestimated part of a switch.

Here's the arithmetic in five steps.

1. Total your current monthly recurring software spend. Every line, every department.
2. Multiply by 12, then grow it 7 percent a year for five years and sum the total.
3. Add your expected headcount growth to any per-seat lines. That's your five-year rent number.
4. Take the quoted build cost, add 60 months of hosting, and add five years of upkeep at 15 to 20 percent of build cost annually. That's your five-year ownership number.
5. Subtract. If ownership is lower, find the month the two lines cross and decide whether you'll still be in business then.

One more adjustment most people skip. Add the labor cost of the manual work your current stack forces. Rekeying data between systems, building the monthly report by hand, chasing status updates in three places. Estimate the hours, multiply by loaded rate, and add it to the rent side. That's the number that usually decides it.

Start when your renewal is six months out, because that's when you still have options. Start when a vendor raises your price without adding anything you asked for. Start when you're paying an integration platform more than one of the tools it connects. Start when someone on your team has become a full-time spreadsheet operator for systems that were supposed to prevent spreadsheets.

Bring your actual numbers, not estimates. A real break-even conversation takes 30 minutes when both sides have statements in front of them. If the math says keep renewing, that's a fine answer and you'll have saved yourself a bad decision.

If you want help running the arithmetic against your own stack, book a free strategy session and bring your last three software invoices.

If you're weighing this decision, read our breakdown of what actually happens to your data and workflows when you migrate off a hosted platform. It covers the export, the cleanup, and the parts teams routinely underestimate.

• What you really own when you own your software
• Per-seat pricing and why it punishes growing teams
• Custom build or off-the-shelf, how to decide by function
• Planning a phased migration off a multi-tool stack