Before a single word goes into the document, five things need deciding. Get these right and the designer's first round of work will land close to what you actually need; skip them and the first round becomes a guessing exercise that costs you a revision cycle to correct.

Purpose

Say what the design has to do. "A website that lets a first-time visitor book a consultation without phoning" gives a designer something to solve. "Something modern and clean" gives them nothing to check their work against, so they will guess, and the guess is rarely the one you had in mind.

Audience

Name who the design is actually for, in terms specific enough to be useful: a parent choosing a nursery, a procurement officer comparing three suppliers, a donor deciding whether to give again. A designer making choices for "everyone" is really making them for no one, and the work tends to show it.

Deliverables

List exactly what you expect to receive: a logo file, a full visual identity, a set of website pages, a component library for a product interface. These are different jobs with different scope, and conflating them is a common source of disappointment on both sides. If you are not sure which one you need, brand identity, visual identity and a logo are three different purchases and worth reading before you write the brief.

Constraints

Tell the designer what is fixed and cannot move: an existing colour used across the organisation, a platform the site has to run on, a deadline tied to an event, a font already licensed for other material. A constraint stated upfront saves a designer from producing good work that then has to be unpicked because it collided with something you already knew about.

Ownership

Decide, before work starts, whether you expect to own the source files (the original working files a designer produces, as opposed to a flattened image or PDF) and the full copyright once the job is paid for. In UK law, under the Copyright, Designs and Patents Act 1988, copyright in commissioned design work does not automatically pass to the client; it needs an explicit IP assignment (a clause transferring copyright to you) written into the agreement. State what you expect here, and check the designer's terms cover it.

Budget as a ceiling

A budget ceiling is worth stating if you have one, because it tells the designer what scope is realistic and saves both sides drafting a proposal that never had a chance of being accepted. It is a constraint: the brief is not the place to negotiate price, and a designer will scope and quote the work in writing once they understand what you actually need.

Settle these five before you write the rest of the brief. The buyer guides cover related decisions, including how design work is priced and what you should expect to own once it is delivered.

Feedback on a design round shapes what happens next more than the brief did, because it tells the designer what to change and why. A reaction states how the work made you feel: "I don't like the colour" or "it feels too corporate." A direction states what should change and against what standard: "the colour needs to work in greyscale for print, and this shade doesn't" or "the tone should read as approachable rather than corporate, because our customers are booking a personal service, not a bank." The designer can act on a direction. A reaction leaves them guessing which of several possible changes you meant, and the next round often guesses wrong.

A reaction

"This doesn't feel right" is a reaction. It's true, and it costs the designer nothing to hear, but it gives them nothing to do. The next version might change the colour, the typeface, the layout or all three, and there's a reasonable chance none of them address what actually bothered you.

A direction

"This doesn't feel right, because the typeface reads as playful and we need to look established" is a direction. It names the problem and the standard the work is being measured against. The designer now has one variable to change, and a second round built on this feedback is more likely to be the last one you need.

Turning a reaction into a direction before you send it

Most feedback starts as a reaction. That's normal, and the useful habit is to ask what's underneath it before you write the email. If you don't know why something feels wrong, say that, so the designer isn't chasing an explanation that isn't real. "We're not sure why yet, but can you show us two alternative directions on this one element" gets further than a confident reason that turns out not to be the one that mattered.