Custom versus off the shelf: the calculation has changed

Last week I wrote about digital foundations and mentioned that the custom versus off the shelf question deserved a post of its own. This is that post, and the short version is this: the advice most businesses are still following was written for a world that no longer exists.

For a long time the sensible default was clear. Don't build custom software. Buy the established product, pay the subscription, and let someone else worry about hosting, security and maintenance. That advice was sound because building was expensive, slow and risky, while SaaS was cheap, fast and someone else's problem when it broke.

What changed on the buying side

SaaS stopped being cheap. Per seat pricing on the major CRM and marketing platforms has climbed steadily, and the pricing models have grown more creative: usage tiers, feature gates, add-ons that used to be included, contacts limits that punish you for growing your own audience.

Then there's the shelfware problem. Most businesses I look at are paying for platforms where the team genuinely uses a small fraction of the features. The rest is paid for, ignored, and quietly renewed every year because nobody wants to own the migration conversation. The subscription was justified by everything the platform could do. The value is determined by what you actually use.

And there's a quieter cost that rarely makes it onto the spreadsheet: the off the shelf tool shapes your process rather than the other way around. Every workaround your team performs to fit the tool is friction, and as I argued last week, friction tends to land exactly where customers feel it.

What changed on the building side

Meanwhile, the cost of building something fit for purpose has fallen further and faster than most people outside the industry realise. Mature open source foundations, modern hosting that costs pounds rather than thousands, and AI assisted development have compressed timelines that used to be measured in months into weeks. A lean internal tool that would once have needed a development team and a six figure budget is now within reach of a small business working with the right partner.

That doesn't mean custom is free. It means the gap has narrowed enough that the comparison is worth making honestly, which most businesses haven't done since they first signed up.

The honest case for off the shelf

Let me argue against myself for a moment, because the established product is still often the right answer. You get security patches without thinking about them. You get support, documentation, an ecosystem of integrations, and a product that improves without you paying for the improvement directly. If your needs are genuinely standard, the standard tool serves them well, and the boring choice is the right choice.

Custom builds carry real obligations: someone has to maintain them, document them, and own them when the person who built them moves on. A custom system without a maintenance plan is a liability with a delay on it.

The honest case for custom

The case for building is strongest when your process is genuinely distinctive, when you're paying enterprise prices for a fraction of an enterprise feature set, or when data ownership matters. With a custom build, your customer data sits where you decide, structured the way your business actually works, and no pricing change or product sunset can hold it hostage.

It's also rarely all or nothing. Some of the best outcomes I see are composable: a solid established core with lean custom pieces around it, doing exactly what the customer journey needs and nothing else. The skill is in deciding where the joins go.

How to re-run the calculation

If you haven't looked at this since you first signed the subscription, here's the exercise. Take a three to five year view, not an annual one. Count the full subscription cost including the seats you'll add as you grow. List the features you actually use against the features you pay for. Add the cost of the workarounds, the integrations that don't quite work, and the data you can't easily get out. Then price the alternative honestly, including build, maintenance and ownership.

Sometimes the spreadsheet says stay put, and that's a good outcome, because now it's a decision rather than a default. Sometimes it says something else entirely, and that conversation is worth having before the next renewal date, not after it.