For twenty years the sensible answer was to buy. Building custom software meant developers, a long project, and maintaining something forever afterwards.

The cost of building has fallen a long way and the calculation has moved with it. The answer is still not to build everything.

01 / 04

What actually changed

Work that needed a team and a year now takes a small group a few weeks. Tools that were only worth building at enterprise scale are now worth building for one planning team.

What has not changed is ownership. Anything you build, you keep. It needs somebody who understands it, and it needs to survive the person who wrote it.

So the question moved. It used to be whether you could afford to build. Now it is whether you should.

02 / 04

Buy it

Buy the systems of record. Accounting, payroll, tax. These carry compliance obligations that change without warning and there is no advantage in owning that burden.

Buy where the vendor's scale is the actual product. Carrier networks, payment processing, fraud screening. You are buying reach and data volume you cannot recreate.

Buy where the thing changes constantly for reasons outside your control. Tax rules, customs classifications, marketplace APIs.

And buy the boring commodity where everybody works the same way and you have no opinion about the process.

03 / 04

Build it

Build where the process is specific to how you operate. If every vendor demo ends with you changing your process to fit their software, you are about to pay for the privilege of working worse.

Build the layer that sits on your own data. Reporting, planning tools, internal applications, anything that reads from your warehouse and helps somebody decide. This is where custom fits best and where SaaS fits worst.

Build when the licence scales with something you are trying to grow. Per order, per user, per product code. That cost curve gets uncomfortable at exactly the moment things are going well.

Build when you have been told twice that something is on the roadmap.

One useful signal. If a spreadsheet already exists in the business to work around a tool you pay for, somebody has already written the specification.

04 / 04

The answer is usually both

Very few decisions are one or the other. The common shape is a bought system of record with things built on top of it.

Keep the ERP. Build the planning tool that reads from it. Keep the ecommerce platform. Build the routing logic that decides where each order ships from.

That way you are not maintaining anything with a compliance surface, and you are not paying a vendor to make you work in a way that does not suit you.

Before deciding either way, ask who owns it in three years and what happens if that person leaves.

If the answer to that is uncomfortable for something you built, build it differently. Put it in tools your own team can read and change, document it properly, and train more than one person on it.