How to write a project brief that gets better proposals

We receive about two hundred briefs a year. Maybe fifteen of them are good. The rest fall into two species: the one-paragraph email ("need new website, modern, budget TBC") and the 60-page procurement RFP that specifies the solution in such detail that the only job left is typing. Both produce the same outcome — proposals that guess, prices padded for the guessing, and a selection process that rewards whichever agency guessed most flatteringly rather than thought most clearly.

A good brief is not long. It's decisive. It tells a studio what problem you're solving, what you already know, what constraints are real, and how you'll choose. Everything else is annex. Here's what we wish every brief contained — written from the receiving end, so you get better proposals from whoever you send it to, including us.

Section one: the problem, with evidence

The single highest-leverage sentence in any brief is a specific problem statement: who hurts, how you know, and what it costs. Compare:

"We need to redesign our website." (a task)

"Trial signups from organic search have fallen 30% year-on-year while traffic grew 15%. Exit surveys say buyers can't tell which plan fits them. We lose them at the pricing page." (a problem)

The second brief lets an agency bring its actual expertise — maybe the fix is information architecture, maybe it's positioning, maybe it's performance. The first brief has already decided, and every proposal you receive will be priced furniture rearrangement.

Attach the evidence you have: analytics exports, churn interviews, the support-ticket themes, that internal deck where someone nailed it. Briefs that share evidence get proposals that reference the evidence. Briefs that hide it (sometimes from competitive caution, usually from disorganisation) get proposals built from the agency's imagination, which is free and therefore worthless.

If you can't yet state the problem with evidence, say so — and make finding the problem the engagement. That's a legitimate shape; it's what a discovery sprint exists for, and it costs a fraction of building the wrong thing confidently.

Section two: what you've already decided (and what you haven't)

Buyers worry that constraints limit creativity. Studios experience the opposite: unexplained constraints produce work that gets killed in week six. The brief should sort every known decision into three buckets:

Fixed. "We must stay on Salesforce." "The rebrand launches at the AGM on 14 October, immovable." "The board has approved the new name; this project applies it." Fixed means fixed — an agency that argues with a genuinely fixed constraint is telling you something useful about itself.
Negotiable. "We'd prefer headless but will hear arguments." "We think we need iOS and Android; if the research says web-first, convince us." Naming something negotiable is an invitation to expertise — and it's how you find out which agencies have any.
Unknown. "We don't know how our content team will actually publish." Unknowns in the brief become discovery work with a price attached, instead of change requests with resentment attached.

The sorting is the work. A brief with honestly sorted constraints will outperform a brief with three times the detail and none of the honesty, because it tells the studio where thinking is welcome and where it will be punished. Our own approach page describes how we scope around exactly these buckets.

Section three: budget, honestly

The most common brief failure: "budget: please propose." Buyers withhold the number to avoid being anchored; the actual effect is that every agency writes a different project. One proposes an $18k templated sprint, another an $180k platform rebuild, and you compare documents that share no assumptions — so you compare personalities and fonts instead.

You don't have to publish your exact ceiling. But give a band — "we've approved $80–120k" or "this is a sub-$40k problem for us this year" — and say what's behind the band: capital budget vs operational, whether it includes content and internal time, whether phase two exists. Then ask agencies to say what they'd do at the top and bottom of the band. That single question produces the most informative section of any proposal, because it forces trade-offs into the open.

A comfortable studio will also tell you when your band is wrong — too low for the stated ambition, or (rarely, memorably) too high. Both are the answer you needed. More on reading the resulting quotes in how studios price.

Section four: who's deciding, and how

Two silent killers of good selection processes: an unnamed decision-maker, and undefined evaluation criteria. Fix both in the brief.

Name the decision forum: "The shortlist is chosen by our head of product and marketing director; the CEO signs off; procurement runs compliance only." This matters because agencies staff proposals to the audience — a brief that will be decided by practitioners deserves (and receives) a different proposal than one decided by a committee scoring matrices.

Then publish your evaluation criteria with weights, even rough ones: team seniority 30%, relevant evidence 25%, approach to our specific problem 30%, price 15%. Yes, agencies will write to the rubric — that's the point. You get proposals addressing what you actually care about instead of a tour of each studio's favourite trophy. If you're tempted to hide the criteria so agencies "show their true colours", what you'll learn is who has the best guesser, and you will marry the guesser's sales director while the delivery team remains a mystery. (On that particular trap, see choosing an agency: signals over showreels.)

Section five: the timeline with a pulse

"ASAP" is not a timeline; it's a mood. A usable timeline has three dates: when you'll pick, when you can genuinely start (when content, access and internal people will actually be available — not when the project plan wishes), and the deadline that's real, with its reason. "We need the pricing page live before September renewals — that's 60% of annual revenue re-papering" arms an agency to phase the work around your value, not their convenience.

Also state your own capacity honestly. The briefs that produce the smoothest projects name the internal humans: "Sarah (marketing lead) can give 4 hours/week; engineering can pair one day a week; legal turnaround is a fortnight." Agencies price client-side drag whether you mention it or not; naming it gets you a plan that works with reality and usually a lower number.

Section six: how to respond

End with mechanics: the date, the format (ask for a short written proposal — 6–10 pages forces choices — plus a call, rather than a free-form deck), the contact for questions and the deadline for them, and whether you'll share anonymised questions and answers with all participants (you should; it raises every proposal and costs you one email).

One more mechanic that pays for itself: state that the proposal should name the actual team, with who-does-what, and that the people presented are the people delivered. This single sentence filters more reliably than any scoring matrix we've seen.

The template, compressed

Steal this outline — it fits on four pages:

The problem (with evidence and what it costs)
What we've decided: fixed / negotiable / unknown
Budget band and what's behind it
Who decides and evaluation criteria with weights
Timeline: selection date, real start date, the deadline and its reason
Your capacity: named people, hours, review turnaround
How to respond: format, dates, question process, the named-team requirement

Everything else — brand guidelines, tech stack documentation, the analytics dump, past research — goes in an annex link. The brief is a decision document, not a data room; its job is to make a hundred small, expensive decisions for fifty recipients at once. Write it like the landing page of your project: one reader, one job, no decoration that costs comprehension.

Key takeaways

State the problem with evidence and cost; never pre-write the solution unless solving it yourself is truly off the table.
Sort every constraint into fixed / negotiable / unknown — the sorting is where expertise gets invited in.
Give a budget band and ask what changes at the top and bottom of it.
Publish the decision forum and weighted evaluation criteria; proposals written to your rubric are proposals you can compare.
Replace "ASAP" with the real deadline and its reason, and name your internal capacity.
Four pages of decisions beats sixty pages of specification. Annex the rest.

FAQ

Won't sharing evidence (analytics, churn data) with multiple agencies leak competitively?

Share it under NDA or in reduced resolution — trends, ratios and funnel shapes without raw customer data. Agencies don't need your customer list; they need the shape of the problem. A brief stripped of all evidence signals either distrust or disorganisation, and strong studios quietly deprioritise both.

We're procurement-run and can't share a budget band. Any workaround?

Share the shape instead of the number: "this is funded as a single-quarter operational project" versus "this is a multi-year capital program". Even a rough magnitude ("not seven figures") aligns proposals enough to compare. If you truly can share nothing, expect to receive incomparable proposals and budget extra time to normalise them — the opacity isn't free, you're just paying it later.

Should we tell agencies who else is pitching?

The number and type, yes ("three studios, all product-focused independents") — it keeps everyone honest about fit. Names, no; it turns the process into shadowboxing. Incidentally, agencies that demand to know are telling you how they'll handle your confidential information later.

How long should we give agencies to respond?

Two weeks for a written proposal, questions due at the end of week one. Shorter and you select for agencies with bench time (not the compliment it sounds like); longer and the energetic ones produce the same proposal with more adjectives.

Do we even need a brief for a small project — say, a landing page?

You need the decisions, not the document. For small work, the compressed template above fits in one page: problem, budget band, date, decider, capacity. Five good sentences beat five days of discovery calls. When you're ready, our contact form is deliberately built around this exact structure — so even the small briefs arrive decision-ready.