Off-the-shelf software is almost always the cheapest choice. If it fits. For accounting, email and calendars you should not have anything built, because that is a solved problem and you would only be running behind everyone else’s development.
Custom gets interesting exactly where you differ from the rest. That is the place a package pushes you into a detour: an extra spreadsheet on the side, a field called “notes” that everyone fills in differently, or three people retyping the same thing every morning because system A and system B do not talk. That detour has a price that never appears on an invoice, and it runs on every year.
What we usually build
Almost everything we build falls into three kinds. A portal for your customers, where they see for themselves what they would otherwise call about. An internal screen for your own people: plan, enter, check, sign off. And a tool that does one specific calculation or assessment that currently needs one colleague’s knowledge.
They have one thing in common: they sit on top of the data you already have. We do not create a second truth. If your customer records are in your CRM, that is where they stay and the application reads along. That is dull, and dull is the point, because an application that keeps its own copy of the customer data is, a year later, the reason nobody knows which address is right.
What you get in the first weeks
Not a design on paper you wait half a year for. Usually there is something that genuinely works within one or two weeks: one screen, with your own data in it, that you can click through. It is not finished, and that is the idea. Something you can use tells you more in ten minutes than three meetings about how it ought to be.
After that it grows piece by piece, in the order that saves you the most. What works stays, and what turns out to be wrong gets thrown away while throwing it away is still cheap.