When Custom Software Beats Paying for Subscriptions

Most businesses accumulate software the way garages accumulate tools. One thing for scheduling, another for quoting, a third for reporting, each with per-seat pricing and an annual increase nobody negotiates.

At some point the question arises: would it be cheaper to build our own? Usually the answer is no. Occasionally it is emphatically yes, and the gap between those cases is worth understanding.

Start with your actual number

Before anything else, list every subscription, its per-seat cost, seat count, and annual increase.

Most businesses doing this exercise find two things. First, the total is higher than anyone thought — nobody adds it up because each line looks small. Second, several tools overlap, with two or three products doing versions of the same job because different people chose different things.

That total is the number the build has to beat.

Where the arithmetic favours building

You are paying for seats you barely use

Per-seat pricing punishes businesses where many people need occasional access. Paying $80 a month for someone who logs in twice is where the waste concentrates.

A build has no per-seat cost. That is the single biggest structural difference.

You are paying for a fraction of the product

Enterprise tools are priced for enterprise feature sets. If you use fifteen percent of a platform, you are subsidising the other eighty-five.

Several tools overlap

The strongest case. Replacing three overlapping subscriptions is a much better proposition than replacing one, because the saving compounds while the build cost does not triple.

A worked example. One client was running three tools with overlapping functions, each per-seat, each increasing annually. We replaced them with a single system built to their process. The saving was $130,000 a year in licence fees alone, before counting the time previously lost to moving data between them.

That is at the upper end of what we see, and it was possible because of the overlap. A single tool replacement almost never produces that.

Your process genuinely differs

If you are fighting your software — maintaining spreadsheets alongside it, or paying for consultants to configure it into an unnatural shape — you are already paying for custom software, just inefficiently.

The vendor has you

Annual increases well above inflation, features moving to a higher tier, pricing changes you cannot refuse. Once switching cost is high enough that you tolerate this, a build is partly buying negotiating position.

Where building is the wrong answer

Under about $20,000 a year in total licence spend. The build will not pay back in a reasonable period. Negotiate your renewals instead — genuinely, most vendors will discount for a multi-year commitment and almost nobody asks.

The tool is good and you use most of it. Well-fitted software is worth paying for. Xero is a good product; almost nobody should build their own accounting system.

Your process is still changing. Building against a moving target means building twice.

You have nobody to own it. This is the one people underestimate. Custom software needs someone internally who owns decisions about it. Without that, it decays.

You want it to do everything. Builds that try to replace an entire stack at once fail far more often than focused ones.

The real cost of building

Anyone quoting only the build cost is misleading you.

That last row is real and routinely ignored. Someone has to answer questions, test, and make decisions. Projects stall on this more than on engineering.

Ongoing cost is typically ten to twenty percent of build cost per year. Software is not a fixed asset — it needs updating as dependencies, integrations and regulations change. A build with no maintenance budget becomes a liability in about three years.

Working out payback

Annual licence spend, minus ongoing cost of the build, divided into the build cost.

A business spending $60,000 a year on subscriptions, commissioning an $80,000 build with $12,000 annual running cost, saves $48,000 a year and pays back in about twenty months.

Under three years, worth considering. Over five, probably not — the risk of your requirements changing is too high.

The middle path most businesses should take first

Between "keep paying" and "build everything" sits building the one piece that does not exist and integrating it with what you keep.

Keep Xero. Keep the CRM if it fits. Build the thing in the middle that currently lives in a spreadsheet and three people's heads.

That is usually a $20,000 to $40,000 project rather than $150,000, it delivers in weeks, and it removes most of the pain. We recommend it more often than a full replacement, and it is the version most likely to still be in use in five years.