
Almost every software decision is framed as a choice between building something custom and buying a product. There is usually a third option that gets less attention: extending a platform you already own. Making the call deliberately, and writing it down, avoids both the custom system nobody needed and the SaaS tool that is bent out of shape to fit.
Start with three questions
First, is this a source of advantage? If the process is how you win customers or how you run operations differently from competitors, it deserves more control. If everyone does it the same way, as with payroll or ticketing, it does not. Second, how much of the need does a product cover without workarounds? Third, do you already own a platform, such as an ERP, CRM or data platform, that could take this on through configuration or integration?
When to buy
Buy when the process is common, the products on the market fit most of your requirements, and your differences are small enough to handle through configuration. You gain a tested product, a roadmap you do not fund directly and a faster start. Check the exit path before you sign: how you export your data, how the pricing grows with usage and how deeply the product would be embedded in your operations.
When to build
Build when the workflow is central to how you operate, when the data or rules are unusual, or when stitching products together costs more than owning the result. Custom software also makes sense when you need to bring several systems together in one coherent experience. Keep the first release narrow, put it in front of real users quickly, and plan for the ongoing cost of ownership: hosting, security updates, support and the people who understand it.
Build what makes you different, buy what everyone needs, and extend what you already own.
When to extend
If you already run a platform that covers most of the need, extending it is often the cheapest and lowest-risk route. Use its APIs, add automation around it, or build a thin application on top for the parts it does not cover. This keeps one system of record, avoids another subscription and shortens the path to value. The limit is the platform itself: if you are fighting its design on every feature, it is time to reconsider.
Compare the full cost over several years
Subscription fees and development estimates are only part of the picture. Include implementation, integration, training, change management, support and the cost of switching later. Compare options over three to five years rather than the first twelve months, and be honest about the assumptions that could change the answer.
Write the decision down
Capture the decision on a single page: the problem, the options considered, the choice, the assumptions behind it and a date to review it. It makes the reasoning clear to new team members and makes it easy to revisit when circumstances change. Few decisions in software are permanent, and a recorded one is far easier to change than an implicit one.
Want to apply this to your own project?
Talk to our Product Engineering team about where to start.


