Agency Pitch Brief Template: Give Every Agency the Same Problem to Solve

Many agency pitches fail as comparisons before the first deck is opened. The buyer sends a prompt along the lines of 'show us how you would grow the brand', four agencies each decide what growth means, what budget to assume and which metric counts, and the buyer then tries to rank four answers to four different questions. An agency pitch brief template fixes this at the source. It states one business problem, one baseline with an observable target and one set of constraints, and it deliberately withholds the tactical answer. The gap between how buyers and agencies see briefs is wide: in the 2021 BetterBriefs survey reported by the IPA, 80% of marketers believed they wrote good briefs while only 10% of creative agencies agreed. What follows sets out the nine fields, the method for the three hardest ones, the constraints annex and the scoring approach that make proposals comparable on thinking rather than on assumptions.

A pitch is a measurement exercise, and most buyers run it without a control. When each agency chooses its own problem, budget and success metric, the result tells you which agency picked the friendliest assumptions, not which one thinks best. GPI's view is that the brief should behave like a well-designed test: fix the inputs, publish the scoring weights in advance, and require every claim in every response to arrive with its baseline, period and stated limitation. Written this way, the brief also protects you from the attribution trap. It asks agencies to explain how they would know a plan worked, which is a harder and more useful question than asking them to promise that it will.

Why open prompts make agency pitches impossible to compare

The variance you create when each agency guesses the problem

An open prompt hands three decisions to each bidder: what the problem is, how much money is in play and what success looks like. Agency A reads 'grow the brand' as an awareness problem with a large media budget and reach as the metric. Agency B reads it as a conversion problem with a modest budget and cost per acquisition as the metric. Agency C treats it as a positioning problem and proposes a research phase before any spend. All three may be competent. None can be ranked against the others, because the responses differ on inputs rather than on capability.

The cost of this shows up at scoring time. The evaluation team ends up scoring the assumption each agency chose rather than the thinking each agency did, and the agency whose assumption happened to match the CMO's private mental model wins on luck. Pitch brief comparability is a property of the brief, not of the evaluators.

The perception gap between marketers and agencies on brief quality (BetterBriefs, 2021)

Buyers tend to be confident their brief is clear. The 2021 BetterBriefs research reported by the IPA found that 80% of marketers thought they wrote good briefs, while only 10% of creative agencies agreed. Treat this as historical perception data from a 2021 survey: it measures how each side felt, not an objective quality score of any brief. The same respondents estimated that around 33% of marketing budget is wasted through poor briefs and misdirected work, a self-reported estimate rather than audited spend. Both figures are useful for one reason: a buyer's confidence in an open prompt is a weak guide to how agencies will actually read it.

What 'the same problem' actually means in a pitch

Comparability requires exactly three fixed inputs, shared identically with every bidder on the same day:

• One problem statement, in commercial terms, with no channel, format or tactic named.
• One baseline with a target metric, a measurement window and a stated data source.
• One constraint set covering budget range, brand and legal limits, and terms each agency must accept before pitching.

Everything outside those three inputs is the agency's answer, and the answer is what you are buying. Before drafting anything else, write the single sentence you would accept as the problem statement from every agency, and check that no channel appears in it. If a tactic has slipped in, you have written part of the answer, and every response will be graded on how well it agrees with you.

Pitch brief, creative brief and RFP: which document does what

Search for 'agency brief template' and most of what returns is a creative production brief: deliverables, tone of voice, asset lists and approval routes. Templates such as those published by monday.com and Demand Metric are built for directing work after a partner is in place. They help a team execute a campaign. They do not help a buyer select the partner. Three documents do three jobs, and mixing them is how a pitch turns into a production quote.

The pitch brief defines the problem to solve

The pitch brief is the strategic prompt. It sits inside or alongside the RFP and contains the problem, the baseline, the customer evidence and the constraints. Its job is to make every agency diagnose the same situation. It should never contain a preferred channel, a deliverable list or a tonal instruction, because each of those pre-empts the strategic work you want to evaluate. Breef frames the agency brief as the document that determines the quality of work returned; in a pitch, the work returned is the diagnosis, so the brief must leave room for it.

The creative brief directs execution after the agency is hired

The creative brief is written with the winning agency once the strategy is agreed. A field such as 'preferred tone of voice: playful, bold' belongs here. In a pitch brief, that same field should be replaced by the customer truth the tone must serve, leaving the tonal answer to the agency. When execution instructions leak into the pitch brief, agencies stop diagnosing and start quoting, and the pitch reveals production capacity instead of strategic capability.

The RFP carries the commercial and governance vehicle

The RFP holds fee structures, itemized scoping, IP and ad account ownership, team continuity clauses and measurement audit rights. GPI's advertising agency RFP template and its agency RFI template cover those in depth. This article summarizes them only as acceptance conditions in the constraints annex.

Distinctions above draw on the framing in Breef and on GPI's separate RFP and RFI guidance. Audit any existing template in your organization, tag each field as 'problem', 'constraint' or 'execution', and move every execution field out of the pitch brief.

The agency pitch brief template, field by field

Copy the nine fields below into a document. Complete fields 1 to 6 internally before contacting any agency, then circulate the finished brief to every bidder on the same day. Each field states what to write, what to withhold and how evaluators will use it. All example text uses placeholders; none of it is a measured result.

The table is the reusable template. The notes below explain the intent of each field; the constraints and high-stakes fields get fuller treatment in later sections.

Field 1: Business problem statement

Write the problem the business needs solved, in one or two sentences, with no channel or tactic named. An illustrative entry with placeholders: 'Repeat purchase rate among first-time buyers acquired in the last 12 months is [baseline]; we need it to reach [target] within [window] without raising acquisition cost above [ceiling].' Nothing in that sentence tells an agency to run retention email or paid social, so each bidder has to diagnose before proposing.

Field 2: Baseline and observable target

Give the current value of the metric, the target, the measurement window and how the metric is measured today, including the data source and its known weaknesses. If your attribution is contested internally, say so. Agencies that model from a shared baseline produce plans you can lay side by side; agencies that must guess the baseline produce plans that differ before a single idea is on the page.

Field 3: Customer truth and market background

Include audience facts you can source: research findings, sales data, service transcripts, category share where you have it. State your current positioning and the competitive context as evidence. Withhold the persona wish list, and mark internal beliefs as beliefs. The field exists so agencies can build on, or argue with, the same body of evidence.

Field 4: What we already know does not work

List approaches you have tested, when, and what the documented outcome was. This stops agencies re-proposing known failures and lets you see which bidders read the brief closely. Withhold your theory of why each approach failed; a strong response will offer its own explanation.

Field 5: Mandatory constraints

List non-negotiables: brand rules that cannot flex, legal or regulatory limits in your category, markets or platforms that are excluded and why. Keep preferences out of this field. A constraint is something a proposal fails for breaching; a preference is something a proposal can argue against.

Field 6: Fixed budget range

State a range and what it covers. The range is a constraint, not a negotiating position, and it is the single input most responsible for scoping variance when it is hidden. The reasoning is in the constraints section; fee models and payment terms belong in the RFP.

Field 7: Response format and evidence requirements

Require a structured response in a fixed order: diagnosis, hypothesis, plan, measurement approach, evidence for every claim, and the named team with roles. Set page limits. Require that any cited result carries its baseline, period, comparison or control, and a stated limitation. This is the field that makes evidence quality comparable across bidders.

Field 8: Evaluation criteria and weighting

Publish the scorecard with weights. Agencies should know before they write that diagnosis and evidence outweigh presentation. Publishing the weights also disciplines your own team, because the scorecard used in the room must match what bidders were told.

Field 9: Timeline and Q&A protocol

Set the dates and one shared Q&A window. Every question and answer is circulated to every bidder. If an answer changes the problem statement, the brief is reissued as an amendment with a revised deadline. The IPA's 2021 BetterBriefs summary reported a wide gap in how marketers and agencies perceive brief quality; a disciplined Q&A protocol is the mechanism that closes that gap during the pitch rather than after appointment.

Writing the high-stakes fields: problem statement, baseline and customer truth

Three fields carry most of the risk: the problem statement, the baseline and the customer truth. These are also the fields buyers feel most confident about, which is exactly the pattern the 2021 BetterBriefs survey reported by the IPA points to, where buyer confidence in briefs ran far ahead of agency agreement. Treat that as 2021 perception data, but let it prompt an extra pass on these three fields.

Turning a symptom into a problem statement

Start from the business decision the campaign must change, not from the metric that is currently annoying. Follow this sequence:

1. Name the commercial consequence if nothing changes, in the terms finance uses.
2. Remove every solution word: channel names, formats, 'campaign', 'content', 'strategy'.
3. Attach the measurable gap between where you are and where you need to be.
4. Ask a colleague outside marketing to read the draft and name the tactic it implies. If they can, remove the implication and redraft.

A weak draft reads 'We need a TikTok strategy to reach Gen Z.' It names the answer and hides the problem. A stronger pattern reads 'Awareness among [defined segment] is [baseline] versus [competitor benchmark from named source]; we need to close that gap within [window].' The rewrite states the gap and its measurement, and leaves channel choice to the agency's diagnosis.

Choosing a baseline agencies cannot argue with

Use the metric your finance or analytics team already reports, not a number assembled for the brief. State the period, the data source and its known limitations. If last-click attribution overstates one channel and everyone internally knows it, write that down. A contested baseline that is honestly described is far more useful than a clean baseline nobody trusts, because agencies can then propose how they would measure around the weakness. A brief that hides the weakness invites plans that inherit it.

Stating customer truth as evidence rather than aspiration

Include only insights that carry a source: research studies, sales data, service transcripts, review analysis. Write the source next to each insight. Where the organization holds a belief without evidence, label it a belief. Agencies then know which claims they may challenge and which they should build on. A customer truth written as aspiration ('our customers value authenticity') is unusable; a customer truth written as evidence ('in [study], [proportion] of lapsed buyers cited [reason]') gives every bidder the same raw material.

How to tell a strong strategic diagnosis from a rewritten brief

Read the diagnosis section of every response against a three-level rubric:

• Echoed: the response restates the brief in slightly different words and moves straight to a channel plan.
• Extended: the response restates the problem in its own language, adds a sharper causal hypothesis and connects it to the baseline.
• Reframed with evidence: the response challenges or refines the problem using the customer evidence supplied or new evidence it cites, and names what it cannot yet know.

The rubric and patterns above are GPI editorial guidance; the perception-gap framing rests on the IPA's 2021 report. A strong response will often disagree with part of your brief. That is a good sign, provided the disagreement arrives with evidence and a statement of what remains unknown.

Fix the constraints every agency must accept before pitching

The constraints annex is a one-page attachment every agency returns before the pitch date, signed as accepted or marked up. Silent omissions are disqualifying; negotiating terms after selection is where a pitch process loses the comparability it worked to build.

Publish the budget range instead of hiding it

Hiding the budget to 'see what they suggest' produces proposals scoped to different financial assumptions, which is the same non-comparability problem in commercial form. A published range normalizes scope and shows what each agency prioritizes inside the same money. The 2021 BetterBriefs respondent estimate reported by the IPA, that roughly a third of marketing budget is wasted through poor briefs and misdirected work, is self-reported and dated, but it illustrates the point: a range is a constraint that stops misdirected scoping, not a negotiating weakness. ID Comms similarly treats clarity of scope and expectations as central to a media agency pitch.

Fix the measurement basis, including how you will judge incrementality claims

State how success will be measured: the baseline metric, the window and the data source. Then require each agency to say what its proposed measurement can and cannot establish. Require every case study cited in the pitch to carry a baseline, a period, a comparison or control, and a stated limitation. Without those four elements, a 'result' is an anecdote, and anecdotes cannot be compared across bidders. Detailed incrementality methods and platform ROAS auditing are covered in GPI's advertising agency RFP template.

Require acceptance or redline of ownership and continuity terms

Summarize the governance terms as acceptance conditions: the buyer owns ad accounts and receives first-party data back at exit; the named senior team in the response is the team that works the account, with written notice and approval for substitutions. Clause drafting stays in the RFP. The pitch brief simply asks each agency to accept or redline before pitching, so the shortlist is not built on terms one bidder later refuses.

Ask how proposals handle consent-constrained measurement

Add one required question: how does the agency evaluate performance when user-level attribution is incomplete because of consent and platform privacy limits? The answer you want describes an approach, its assumptions and its limits, rather than naming a preferred tool.

The constraint list above is GPI editorial guidance informed by ID Comms on pitch practice. To see how documented agency evidence is presented publicly, look at a GPI profile such as DEPT's paid media listing and compare it with the case-study requirements in your brief.

Replace speculative creative with a paid or diagnostic response

What unpaid spec work costs the agency and hides from the buyer

Speculative creative is expensive for the bidder and uninformative for the buyer. Sidekick Accounting offers a hypothetical worked example: a creative director who normally bills 60% of time at £150 per hour and spends 40 unbillable hours on a pitch forgoes about £3,600 in revenue. That is the source's illustration, not an industry average, and the numbers will differ for every agency. The mechanism is what matters. Agencies absorb the cost, recover it across paying clients, and rationally spend pitch hours on what wins pitches, which tends to be polish and speed rather than research.

That is the hidden cost to the buyer. Unresearched creative rewards the agency that can produce attractive work quickly on thin inputs and penalizes the agency that would normally spend the first weeks understanding the customer. You end up selecting for the wrong discipline.

The diagnostic response: what to ask for instead

Make the diagnostic response the default deliverable, structured as Field 7 requires:

1. A diagnosis of the problem in the agency's own words.
2. A strategic hypothesis tied to the baseline.
3. A measurement plan stating what it can and cannot establish.
4. A resourcing plan with named people.
5. A worked example drawn from the agency's own documented past work, with baseline, period, comparison and limitation.

The past example replaces spec creative as the demonstration of craft. It shows how the agency actually works when it has done the research, which is the condition under which you will be paying it. Guidance on briefing performance creative from Launchcodex makes a related point for live work: creative built without inputs tends to be creative built on guesses.

When creative execution is genuinely needed, fund it

Sometimes the decision does depend on seeing creative. In that case, pay for it. Set a stipend, define the scope tightly, and issue a shared research pack so all agencies build from the same inputs; otherwise the funded creative reintroduces the comparability problem. State in the brief who owns the pitch materials, whether paid or unpaid, so neither side is surprised later. Strike any request for finished creative from the brief unless these three conditions are met.

Choosing the brief tier: brand, campaign or tactical pitch

Classify the pitch as brand, campaign or tactical before writing Field 1. The nine fields do not change; the scope of the problem statement, the length of the measurement window and the evidence type required all do.

Brand-level pitch: positioning problem, long baseline

Frame the problem as a perception or preference gap among a defined segment, measured by a tracker or research study you name. Require the agency to state how it would measure movement over a window long enough for perception to shift, and what leading indicators it would watch in the meantime. Evidence in responses should be positioning and research work with tracked outcomes over comparable windows.

Campaign-level pitch: a bounded commercial objective

Frame a bounded objective with a fixed budget range and a defined window. Require a hypothesis that connects the plan to the baseline, and evidence from past campaigns that carries baseline, period and comparison. Many pitches sit at this tier, and it is where a leaked channel preference does the most damage.

Tactical pitch: a channel or performance problem with a known baseline

Here naming the channel is legitimate, because the channel is the given rather than the answer. A tactical problem statement may read 'paid search cost per qualified lead is [baseline] against a target of [target] within [window]'. The same wording in a brand-tier brief would wrongly prescribe the solution. Evidence should be measured performance work in that channel with stated limitations.

Adjusting scorecard weights by agency discipline

The framework applies to media buying, creative and full-service agencies alike. Shift weighting toward measurement rigor for media agencies, strategic diagnosis for creative agencies and integration across disciplines for full-service bidders. The ID Comms view of media pitches as a structured process rather than a beauty contest is consistent with weighting substance ahead of presentation regardless of discipline.

Tier definitions above are GPI editorial guidance; the pitch-process framing draws on ID Comms.

How to evaluate responses against the unified brief

Build the scorecard from the weights you published in Field 8, so the evaluation matches what agencies were told. Have each evaluator score independently before any group discussion, then reconcile differences field by field rather than on overall impression.

Score diagnosis, evidence and measurement before presentation

Score the diagnosis with the echoed, extended and reframed rubric from the high-stakes section. Score evidence by whether each cited result carries baseline, period, comparison and limitation, mirroring the case-study requirement in the brief. Score measurement plans on whether they name a business decision the measurement will inform, not on the sophistication of the tools listed. Presentation is scored last and lightly. There is a commercial reason to expect polish everywhere: OpenAsset's 2024 compilation reports that advertising, media and telecom firms derive around 46% of revenue from RFPs. That figure needs a firm caveat. OpenAsset attributes it to a third-party report (WebinarCare), no primary methodology, sample or observation period is available to check, and it is a 2024 data point rather than a current benchmark. Treat it strictly as directional context for how much agencies have riding on the response itself, and note that the link from revenue dependence to heavy investment in presentation is this article's inference, not something the compilation tests. The practical implication still holds at a directional level: response quality will be high across the board, so polish cannot separate bidders and evaluators must score substance first.

Weights above are illustrative placeholders; set your own in Field 8 and publish them. The RFP revenue figure is from OpenAsset, a 2024 compilation that attributes it to a third-party report; no primary methodology is available, so it is directional context only.

Read every proposal against the same baseline

Because Field 2 fixed the baseline, every plan should start from the same number. If a proposal restates the baseline differently, ask why before scoring; it may have found an error in your data, which is valuable, or it may have quietly chosen a friendlier starting point, which is a warning.

Document what each agency could not establish

For every proposal, record what it did not or could not prove: claims without comparisons, measurement plans without limits, team commitments without terms. The shortlist decision should rest on documented gaps as well as documented strengths. This mirrors how GPI's methodology weighs proof of measurable outcomes and operating maturity when assessing agencies for its directory; the same standard of documented evidence that a public profile is held to is a reasonable standard for a pitch response.

Use GPI's evidence standards to check what your pitch brief surfaces

One problem, one baseline, one constraint set, and no prescribed tactic. Hold to those four and the proposals you receive will differ on thinking and evidence rather than on assumptions, which is the only kind of difference a scorecard can rank.

Match the brief's evidence requirements to how GPI documents agencies

The evidence standard in Field 7, that every claim carries its baseline, period, comparison and limitation, is the same standard GPI applies when it documents agencies: proof of measurable outcomes and operating maturity, assessed on what can be shown rather than asserted. Before issuing the brief, compare its evidence requirements with the criteria GPI uses to assess agencies in its methodology, and tighten any field that would accept an unsupported claim.

Use directory profiles to see what documented proof looks like before you shortlist

Where a bidder has a public GPI profile, you can cross-check how its documented capabilities and evidence are presented against what it claims in the pitch. DEPT's paid media listing shows how documented proof and capabilities appear on a live profile, and the ACE Agency listing is a further example of the same format. GPI has not run or scored any pitch; the profiles are a reference for how documented evidence reads, not a verdict on any response.

Finalize the nine fields internally, attach the constraints annex, issue the brief to every bidder on the same day, and score against the weights you published.

FAQ

How many agencies should receive the same pitch brief?

Enough to give you a genuine comparison and few enough that you can read every response against the rubric properly. In practice that usually means a shortlist rather than an open call, screened beforehand with an RFI. Every agency on the list receives the identical brief, data pack and constraints annex on the same day, with the same Q&A window. Adding a bidder late without the full pack breaks comparability for the whole set.

What if an agency asks a question that changes the problem statement?

Treat it as an amendment, not a clarification. Any answer that alters Field 1, Field 2 or the budget range is reissued to all bidders as a numbered brief amendment with a revised deadline. For example, if one agency asks whether the range includes working media, the answer and the clarified range go to every bidder the same day. No agency should scope on private information, and the question itself is useful evidence of how closely that bidder read the brief.

Can we run a pitch without disclosing the current baseline?

You can, but you recreate the problem the brief exists to solve: each agency will assume a different starting point and the plans will not be comparable. If the data is commercially sensitive, share it under NDA rather than omit it. If the baseline is genuinely unknown, say so in Field 2 and make establishing it part of the problem, so every agency is asked to propose how it would measure from zero.

How do we brief incumbents and challengers on identical terms?

Give challengers the same data pack the incumbent already holds, including the documented failures in Field 4 and the honest limitations of your current measurement. Then apply the same evidence standard to the incumbent's case studies as to everyone else's. An incumbent that cannot show baseline, period and comparison for its own work on your account has told you something important about its operating maturity.

Should the brief include our own hypothesis about what is wrong?

Only if you label it as a hypothesis in Field 3 and invite agencies to challenge it. An unlabelled hypothesis reads as an instruction and pulls every response toward it, which flattens the very differences in diagnosis you want to score. A labelled one gives you a clean signal: agencies that accept it uncritically, agencies that extend it, and agencies that reframe it with evidence.

What happens when an agency redlines a constraint rather than accepting it?

Treat redlines as evaluation data on operating maturity, not as automatic disqualification, unless the constraint is marked non-negotiable in Field 5. A reasoned redline on team substitution notice, for instance, may reveal a more honest resourcing model than a silent acceptance. Silence is the disqualifier: an agency that returns no annex at all has not engaged with the terms, and negotiating them after selection is where comparability is lost.