Custom software vs off-the-shelf software, decided by the numbers

Why the custom software vs off-the-shelf software question is the wrong one#

Every ranking page frames this as one big switch. Custom on one side, off-the-shelf on the other, pick a lane. That framing is what gets teams stuck. No real business runs on a single tool, so the choice is never all or nothing. Your workflow is a set of slices. Some are commodities every competitor runs the same way. Others are the specific things customers choose you for.

The useful move is to test each slice on its own. Buy the commodity slices, since a mature vendor already sells them well. Build only the differentiator slices, the ones you cannot rent. The answer is almost always a mix, not a side. For the pre-read on where a build even makes sense, see what custom software development actually involves. This page picks up after that, at the money question.

Which parts of your workflow should you build, and which should you buy?#

This is a two-move idea. First, a procedure to test any slice. Second, the traits that tell a commodity apart from a differentiator. Take them in order.

The differentiator test: does this slice win you customers?#

Run each slice through one question. Does it win or keep customers? If not, it is table stakes, so check whether a mature vendor sells it well. When a vendor does, buy it. When no vendor fits, configure or extend a platform and build only the gap. If the slice does win customers, ask whether leaning on a vendor would put that edge at risk. High lock-in risk points to a build. So does a rule set that changes often as you learn.

Commodity vs differentiator, trait by trait#

The decision tree resolves fast once you can name the traits. Here is the attribute matrix. Read each row and place your slice on one side or the other. A slice that sits mostly in the left column is a buy. One that sits mostly on the right is a build.

Notice how few slices truly land on the right. For most businesses, one or two parts of the workflow are real differentiators. The rest are commodities a subscription handles well. That imbalance is the whole reason the honest answer is a mix.

The question your CFO actually asks: when does building pay back?#

A CFO does not care which tool is nicer. The CFO asks one thing. At what seat count and time horizon does owning a build cost less than renting a subscription? That is a break-even question, and it has a real answer. The custom software vs off-the-shelf software decision comes down to a crossover year. The section below hands you the calculator to find it.

Run your own numbers#

Set your own inputs below. Seats, seat growth, subscription price, renewal escalation, a one-time build estimate, and annual maintenance all feed the model. The chart plots cumulative cost for both paths and marks the crossover year. You get a plain recommendation plus the US financial-lens deltas. Every number is an illustrative planning figure, not a quote.

Move the sliders and watch the crossover shift. Scale stops being a hunch and becomes a calculation you can defend. A high seat count and steep renewal escalation pull the crossover earlier. A small team on a cheap tool never reaches it, so the recommendation says buy. This is the same crossover we build on a client's real seats, growth, and renewal terms in a discovery engagement, before a US team commits capital to either path.

How the crossover math works#

The tool is not a black box. Here is the arithmetic behind it, in plain terms. The subscription path compounds twice. Your seat count grows, and the per-seat price renews upward every year. The build path is a large upfront cost plus a flat maintenance line. Where the running subscription total passes the running build total, custom starts to win.

The US financial lens most comparisons skip#

Most comparison pages stop at a cost table. A US buyer has to defend the decision to a finance team that thinks in accounting treatment and tax timing, not headline price. Here is the lens most of them skip. It has three parts, and each one moves the real number.

OpEx vs CapEx: how each path hits your books#

The two paths land on your books in opposite ways. A subscription is operating expense. You deduct it as you pay it, and it never becomes an asset. A custom build is capital expense. You carry it as an asset, then write it down over time. The same dollar looks very different to a CFO depending on the path. Here is the line-by-line contrast.

IRC Section 174: why a custom build no longer expenses in year one#

This is the single most decision-changing fact, and almost no comparison page mentions it. Before 2022, a US firm could deduct software development costs in the year it spent them. Then a provision of the 2017 tax law took effect. Now those costs must be capitalized and amortized over five years for domestic work, so only about one fifth of a custom build is deductible in year one. The rest arrives across the following years.

SaaS renewal escalation: the compounding line item nobody budgets#

Buyers budget the sticker price and forget the renewal. Yet most subscriptions rise every year, often 6 to 10 percent, before you add a single seat. That compounding is quiet in year one and loud by year seven. So the chart below isolates it. Hold the seats flat and watch what an 8 percent renewal escalation does to one seat over seven years.

Now add seat growth on top of that curve. The two compoundings stack, and the subscription line bends sharply upward. The build line stays nearly flat after launch. That divergence is exactly what the calculator turns into a crossover year.

Worked example: a 140-seat US logistics operator#

For instance, here is the method on real numbers. Picture an illustrative US logistics operator, 140 seats and growing 12 percent a year. Its off-the-shelf platform costs $52 per seat each month, with 7 percent renewal escalation. A custom build to replace it is estimated at $310,000, with 16 percent annual maintenance. This buyer is a fictional example, drawn to show the math, not a real client.

Feed those numbers in and the crossover lands just before the end of year 4. Below that point, the subscription is cheaper and building would be vanity. Above it, owning the software pulls ahead and keeps widening. By year five, the gap is real money. The calculator above is pre-seeded with these exact inputs, so you can watch the case run live.

Change one input and the story changes. For instance, drop the seat count to 40 and the crossover disappears inside the horizon, so the recommendation flips to buy. That sensitivity is the point. A defensible decision comes from your own numbers, not a generic verdict. If you also need to size the build itself, our guide to how much a custom build actually costs gives you the estimate to plug in here.

When NOT to build custom software#

A custom-software firm that only ever says build is not worth trusting. So here is the honest counterweight. In several common cases, off-the-shelf is plainly the right call, and we will tell you so.

In short, the custom software vs off-the-shelf software answer is a portfolio, not a verdict. Custom software earns its cost in the opposite conditions. First, build when the slice is a genuine differentiator customers feel. Second, build when integration or control is a hard requirement a vendor cannot meet. Finally, build when scale economics make per-seat pricing hurt more each year. For the enterprise-scale version of this decision, run through the five-year total cost of ownership view for CTOs. When you are ready to turn a rough idea into a scoped brief, a completed software development RFP is the document that gets you there.