How long does custom software take? Honest timelines.

After “how much does it cost,” this is the question I hear most, and it gets the same useless stock answer from most of the industry: “it depends.” It does depend. But you deserve better calibration than that before you plan a busy season around a new system.

So here are honest ranges, and — more useful — the things that actually move them.

The short version

A small, focused tool — the kind that kills one painful spreadsheet or automates one workflow — is measured in weeks. Not months. If someone quotes months for a single-workflow tool, ask what the extra time is buying you and listen carefully to the answer.

A proper system — quoting, scheduling, a customer portal, something with several user roles and an integration or two — is usually a couple of months from kickoff to your team using it daily. The build itself is often faster than that; the calendar time comes from everything around the build, which is the part nobody warns you about.

Multi-system projects run longer, but if you’re commissioning one of those you’re already having a different conversation than a blog post.

What actually eats the time

Here’s the industry secret: on small projects, typing the code is rarely the bottleneck anymore. AI-assisted development compressed that part dramatically. What still takes real time:

Decisions. Every “should it work like this or like that?” that sits in someone’s inbox for four days adds four days. The projects that finish fast are the ones where somebody on your side can answer questions the same day. That’s it. That’s the biggest lever, and it’s yours, not your developer’s.

Integrations. Your accounting package’s API is somebody else’s software, somebody else’s documentation, and somebody else’s support queue. Most integrations are fine. The occasional one fights for a week. You can’t always know which in advance, and an honest quote says so.

Real-world testing. The first version your team touches will surface things nobody thought to mention — the customer type that doesn’t fit the form, the Tuesday process that’s different from the Friday one. That feedback loop is where good software comes from, and it needs a week or two of people actually using the thing while it can still be shaped cheaply.

Timeline red flags, both directions

Suspiciously fast counts too. “We’ll have it done next week” for a system that touches your invoicing usually means someone plans to bolt a template onto your business and call it custom. I’ve written about the other warning signs — schedule fantasy is on the list in both directions.

The enterprise-flavoured quote — nine months, three phases, a discovery phase that costs more than the tool should — is the opposite failure. That’s a big company’s process billed to a small company’s problem. The buy-vs-build post covers when you should walk away from custom entirely, and an inflated timeline is a good reason to at least re-run that math.

How to keep your own project fast

Pick one workflow, not five — the same advice as scoping the cost, because scope is where cost and schedule are both decided. Name one person who can answer questions quickly. Plan for your team to actually try the early version instead of waiting for “done.” Do those three things and you’ll land at the fast end of every range above.

If you want a real timeline for a real workflow, tell me about it — or read up on how we run custom projects first. Either way you’ll get a date you can plan around, not a shrug.