The worst software briefs are the most detailed ones. Forty pages of buttons, screens and dropdown menus, describing a solution the author designed in their head, usually a faithful copy of the current process with its problems lovingly preserved.
The best briefs do something braver: they describe the problem so well that a good developer can earn their fee solving it. Here's how to write one of those, and it's shorter than you'd think.
One paragraph: what's happening operationally that made you pick up the phone? Quotes taking days. Double-keyed orders. A spreadsheet nobody trusts. The reason for the project is the north star every later decision navigates by, so put it first and make it honest.
Walk through the current workflow as it actually runs: what triggers it, who touches it, where the information lives, where it waits. Include the exceptions and the workarounds, because they're where the real rules hide. If you've done the exercise from mapping a workflow, this section writes itself. Then describe the people: who uses this daily, who manages, who needs to see but never touch. Roles now save permission redesigns later.
"Staff must know when a booking changes" is a requirement. "A popup notification" is a solution, and possibly a bad one. State the need and leave the how open, and you'll get a developer's best thinking instead of an estimate for your first guess. It's genuinely fine, useful even, to write "we don't know how this should work, but it must achieve X".
Priorities: what must the first release do, what's later, what's out entirely? Data: what exists today, in what state, and does it need migrating? Integrations: which tools must this talk to, and which are being retired? A brief that says "quoting is in scope, invoicing stays in the accounts package" has just saved everyone a month of assumption-untangling.
Exceptions (what should happen when things go wrong), evidence (what needs an audit trail) and support (who looks after this after launch). Briefs that mention these three read as written by someone who's run an operation, and they get noticeably better proposals back.
"Quotes out same-day" and "no order typed twice" are success measures. Invented precision ("37% efficiency gain") is decoration. Two or three honest measures give the project a finish line, and give you a fair test to hold everyone to, us included.
A good brief isn't a specification. It's the opening move of a conversation, and projects like Jones International's systems started as exactly this: a clear problem, honestly described. Write yours, keep it under five pages, and send it to a developer whose response tells you how they think. If you'd like a second pair of eyes before it goes anywhere, that's a service we're happy to provide with no obligation attached.