Agency RFI Template: What to Ask Before Building Your Shortlist

An agency RFI template should narrow the long list using documented evidence before anyone writes a proposal. Strategy questions and sample plans can turn that screen into an unpaid RFP. The ANA and 4As reported agreement among surveyed members in 2025 that RFIs needed simplifying, although they published no statistical breakdown. This guide provides marketing agency RFI questions, a field-by-field request for information template and a scoring process covering measurement, team integration and commercial fit. If you are still defining what kind of partner you need, GPI's guide to choosing a growth marketing agency covers the selection criteria this template turns into questions.

GPI's view is that an RFI is a measurement instrument for a hiring decision, and it should be built the way good measurement is built: start from the decision, ask only what changes it, and treat every answer as a claim with evidence, a method and limitations behind it. That is why this template spends its scarce space on how an agency has measured incrementality, how its creative and media people work together, and how it structures fees, rather than on awards or client logos. An agency reporting platform-attributed returns as proof of impact has told you something useful, but not what it thinks. The RFI is where you find that out cheaply, before anyone writes a proposal.

RFI vs RFP: Keep the Screening Tool a Screening Tool

What an RFI is for and what it is not

An RFI gathers vendor capabilities; an RFP requests a detailed proposal for defined work. Federal procurement makes that boundary explicit: Arphie's summary explains that RFI responses are information and will not result in a contract award. Private advertisers can state a comparable boundary for their screening stage.

The distinctions in this table follow the Precoro definitions and the federal information-not-proposals framing described by Arphie; the RFQ row is included only for completeness, since agency services are rarely specified tightly enough to quote as a commodity.

Why agency RFIs balloon into RFPs

RFIs expand when stakeholders keep adding requests for rates, category strategy and sample plans. The ANA/4As standardized template, published in July 2025, addresses that pressure with a common, shorter document. Construction offers a limited analogy: its RFI clarifies existing documents. That process concerns a signed contract, whereas agency RFIs precede hiring; the shared lesson is to keep the request within its stated purpose.

What happens if you skip the RFI stage

Unqualified agencies enter a longer, costlier process. Without a screen, you evaluate full proposals from firms that would have failed a budget floor, a channel fit test or a conflict check in ten minutes. Reviewers spend hours on documents that cannot advance, and agencies commit unpaid strategic work to a bid they could never win, which erodes their willingness to participate the next time you run a review. An RFI costs a fraction of either.

Write one sentence at the top of your RFI stating that responses are information for shortlisting, that no strategy or pricing quotes are requested, and that a separate RFP will follow for shortlisted agencies.

Start With the Shortlist Decision, Then Write the Questions

Name the decisions the RFI must support

Build agency screening questions around the decisions. Arphie's RFI guidance puts it plainly: define the decision before writing anything, then remove any question that will not influence it. For an agency search, the shortlist decisions usually reduce to five:

1. Confirm budget fit: can the agency work within your spend and fee range?
2. Check channel and market fit against recent work.
3. Assess measurement maturity: methods used and limitations disclosed.
4. Review conflicts and ownership before advancing candidates.
5. Confirm the availability and current workload of the proposed team.

Each decision needs one or two fields. Budget fit needs a confirmed range and a fee-structure description. Channel fit needs a table of relevant work. Measurement maturity needs a description of methods run. Conflicts need a client list by category and an ownership statement. Availability needs named people and current load. Nothing else is required to shortlist.

Remove every question that does not change a decision

Response quality tracks request quality. Deep.stream's sourcing guidance makes the point that the format of the request determines the quality of what comes back, and padded requests reliably produce padded, uncomparable answers. The reverse also holds: specific, detailed questions make it easier for vendors to respond and give you a fuller picture than open prompts such as 'describe your approach.' Test each draft question against the five decisions. If you cannot name the decision it informs, cut it, however interesting the answer might be.

Set the response format before the questions

Fix the shape of the answer before you finalize the wording. Asana's template guidance notes that identical questions to every vendor allow side-by-side comparison and cut reviewer time spent on formatting. Apply three format rules: a word cap per field, yes/no answers that must be followed by evidence, and structured tables for case studies with fixed columns. These make reviewing a matter of reading down columns rather than across decks.

Draft a two-column sheet: left column lists the shortlist decisions, right column lists the fields that inform each. Delete any field with an empty left cell.

The Agency RFI Template, Field by Field

Copy the fields below into your document. Each field gives the purpose, the exact prompt, the format constraint and the decision it serves. Two public documents are useful starting points: the ANA/4As 2025 standardized RFI template and the Smartsheet creative agency RFI template. This version adds fields on measurement and creative-media integration that generic templates leave thin.

Part A: What you give the agency

Agencies cannot answer relevance questions without knowing your decision, scope and budget range, and the clarity of your request drives the quality of their response. Complete every Part A field before sending. Keep the whole of Part A to one page.

A1. Business context and the decision you are making

Purpose: let agencies self-select out early.

Prompt text: Describe your company, category, primary markets and the business decision this review supports (for example, replacing an incumbent, consolidating channels or entering a new market). Name the outcome you expect the agency to influence and the timeframe.

Format: 150 words maximum. Decision served: channel and market fit.

A2. Scope, channels and markets in play

Purpose: define what is being screened.

Prompt text: List the channels, markets and services in scope, those explicitly out of scope, and any existing partners who will remain in place.

Format: bullet list, noting which items are confirmed and which are under consideration. Decision served: channel fit and conflicts.

A3. Budget range, timeline and the review process

Purpose: keep the RFI on the screening side of the boundary. State a range or minimum and ask for confirmation rather than a quote, consistent with the ANA/4As screening-tool position.

Prompt text: Our anticipated annual working media range is X to Y and expected agency fees fall within Z. Confirm whether you work at this level. The RFI closes on [date]; shortlisted agencies will receive an RFP on [date] with a decision by [date].

Format: stated range plus dated timeline. Decision served: budget fit.

A4. Response format, length limits and deadline

Purpose: guarantee comparability.

Prompt text: Answer each numbered field in order, within the word cap stated, using the tables provided where indicated. Do not attach credentials decks or case study PDFs; unrequested material will not be scored. Submit by [date and time zone]. Questions may be sent to [contact] until [date]; all answers will be shared with every participant.

Format: rules, not a question. Decision served: all of them, by making responses comparable.

Part B: What you ask the agency

Every agency receives identical questions in identical order. Hold to the word caps you set, since the caps are what make side-by-side comparison possible and keep reviewer time on evaluation rather than reformatting.

B1. Firm profile, ownership and conflicts

Prompt text: State your legal entity, ownership structure and parent company (if any), offices, headcount and year founded. List current clients in our category and adjacent categories, and any conflict policy that applies.

Format: 150 words plus a client list table (client, category, service, start year). Decision served: conflicts and ownership.

B2. Relevant experience, documented case studies and references

Prompt text: Provide up to three engagements relevant to our scope. For each: period, client objective, the metrics used and how each was defined, the result, your exact role versus the client and other partners, whether the client has approved disclosure, and a named reference willing to speak at RFP stage.

Format: one structured table per case study, 200 words each, unranked. Decision served: channel fit, measurement maturity, and which claims need verification later.

B3. Proposed team and capacity

Prompt text: Name the people who would lead this account, their roles, years at the firm and current number of accounts. State whether they are committed to this engagement if you are appointed, and what happens if they leave.

Format: table (name, role, tenure, current accounts) plus 100 words. Decision served: team availability.

B4. Channel capabilities and tooling

Prompt text: For each channel in scope, state years of active practice, current spend under management as a range, platform certifications held by named team members, and the tools you use for planning, buying and reporting, identifying which are proprietary.

Format: table by channel plus 150 words of commentary. Decision served: channel fit.

B5. Measurement approach and reporting

Prompt text: Describe measurement methods you have actually run in the last three years (for example, holdout tests, geo experiments, marketing mix modeling, platform lift studies), on which channels, and what each could not establish. Describe your standard reporting cadence and what it includes.

Format: 250 words plus a methods table (method, channel, when, limitation). Dashboards are not methods and should not be listed as such. Decision served: measurement maturity.

B6. Creative and media integration

Prompt text: Describe how creative and media teams work together on a live account, including who attends which meetings and how performance data reaches creative decisions. Give one example where a media finding changed a creative choice, or a creative finding changed a media plan, and how the result was measured.

Format: 200 words. Decision served: fit for integrated scope; it also produces a claim to verify with the reference in B2.

B7. Commercial model and minimum engagement

Prompt text: List the fee structures you have used in the last two years (retainer, percentage of media, performance-linked, project), your minimum annual engagement, and what is typically excluded from fees such as production, tools or data costs. Do not provide a quote.

Format: 150 words as a structured list. Decision served: budget fit.

B8. Compliance, data handling and brand safety

Prompt text: Describe your data protection practices for client and customer data, relevant certifications, your brand safety controls for programmatic and creator buys, and any regulatory issues or disputes in the last three years.

Format: 200 words; yes/no with evidence for certifications. Decision served: risk screen before RFP.

The word caps in this table are working recommendations rather than measured standards. They follow the principle that detailed, specific questions yield more complete answers and Asana's point that identical structures make side-by-side review practical. Adjust the caps if you must; keep the structure.

High-Stakes Fields: How to Read the Answers on Measurement, Integration and Money

Four fields do most of the work in separating shortlists: B5, B6, B7 and the verification columns in B2. For each, decide in advance what a disqualifying answer looks like, then score responses against that standard rather than against each other. This section describes answer patterns; it makes no claim about what agencies typically write, because no such evidence exists in the research behind this guide.

Measurement and incrementality: what a credible answer contains

A credible answer names measurement methods, channels, limitations and decisions changed. Ask about holdouts, geo tests, marketing mix models and platform lift studies the agency has actually run. Attributed returns do not establish causation. A paused channel, budget reallocation or creative rotation shows how analysis informed action, applying the template's decision-first logic.

These patterns are interpretive guidance derived from the template logic and from GPI's practice of assessing agency claims on documented evidence, stated methodology and limitations. They are not drawn from a survey of agency responses, and the ANA/4As position that the RFI is a screen means a credible answer should prompt further questions at RFP stage rather than settle the matter.

Creative-media integration: evidence of working practice, not org charts

Read B6 for working practices: joint reviews, shared reporting and a dated example of delivery data changing a creative decision. Separate departments can claim integration without demonstrating it. Check the example with the B2 reference. If a word cap excluded the measured outcome, request it as a follow-up before judging the answer.

Commercial model: pricing structure without a quote

Ask for fee structures, minimum engagement and exclusions to establish budget fit. A quote requires a scoped assignment and crosses the screening boundary the ANA and 4As sought to restore. Clarify an unsolicited quote with the agency. Refusal to describe its commercial structure leaves the budget-fit decision unsupported.

Verifying claims: case studies, metric definitions and references

As federal RFI practice illustrates, responses provide information for evaluation. B2 records period, metric definition, agency role, disclosure status and reference contact so claims can be checked. Missing fields limit verifiability. The RFI identifies claims for follow-up; the RFP, reference calls and due diligence perform that verification.

Tailoring the Template by Agency Type

Parts A and B stay constant across searches. Only B4 (capabilities) and B5 (measurement) gain discipline-specific prompts, so one master document serves every review. The same rule applies as elsewhere: add a prompt only when it informs a shortlist decision, and phrase it specifically enough to answer in a table.

Paid media and performance agencies

• Add to B4: platform certifications held by named team members, account structure philosophy, and how budgets are paced across the month.
• Add to B5: which incrementality methods have been run per platform, and how the agency reconciles platform-reported and incremental results.

Creator and influencer agencies

• Add to B4: how talent is sourced and vetted, how contracts and usage rights are handled, and how disclosure compliance is managed across markets.
• Add to B5: how creator content performance is measured beyond reach and engagement, including any paid amplification testing.

Amazon and retail media agencies

• Add to B4: catalog and retail readiness dependencies the agency expects the client to own, and named team experience with the advertising console and the DSP.
• Add to B5: how the agency separates ad-driven from organic sales, and what it reports when that separation is uncertain.

For retail media searches, the GPI guide on how to choose an Amazon ads agency lists the console, DSP and catalog questions worth adding to field B4.

Creative and production agencies

• Add to B4: the strategy-to-production workflow, testing cadence, and production capacity by format.
• Add to B5: how creative performance data reaches the people making the next round of work, with an example.

The flags in this table are prompts for follow-up questions at RFP stage rather than disqualifiers. They are built from GPI's directory categories and the template's decision logic, not from data about how any agency type typically answers.

Choose the discipline you are hiring for, append its prompts to B4 and B5, and leave the rest of the template unchanged.

Issuing, Managing and Scoring RFI Responses

Step 1: Build the long list and confirm basic fit

1. Gather candidates from directories, referrals and category research, then run a public check on each before sending anything: ownership, category focus, visible client conflicts and whether the firm appears to operate at your budget level. Ten minutes per agency here removes firms that would otherwise consume hours of reviewer time. Keep the long list large enough that the screen has something to remove and small enough that every response will be read.

Step 2: Issue identical documents with a Q&A window

1. Send the same document, on the same day, with the same deadline, to every agency. Open a Q&A window of fixed length and publish every question and answer to all participants, since comparability depends on every agency answering the same questions with the same information. Answering one agency privately, even on a small point, gives it an advantage the rubric cannot see.

Step 3: Give a realistic response window

1. Set the window from the response format's workload. In Responsive's 2025 cross-industry research, more than 75% of respondents reported pressure for faster, more personalized answers. Only 29% reported effective self-service knowledge access, while 88% of high-growth teams had some knowledge centralized. These are response-team figures across industries, not agency benchmarks. A fixed format and realistic deadline give respondents room to substantiate answers.

Step 4: Score against a pre-agreed rubric

1. Finalize weights and score definitions before the RFI goes out, and circulate them to reviewers so the criteria cannot move after responses arrive. Weight the shortlist decisions from the second section, score each field from 0 to 3, and require a reviewer note per score. The weights below are hypothetical, shown to illustrate the structure; set your own to reflect the decision at hand.

Weights in this table are illustrative and hypothetical. The scoring approach extends the side-by-side comparison logic of standard RFI templates and the comparison-sheet format Smartsheet offers, adding weights and evidence definitions so that two reviewers can reach similar numbers independently.

Step 5: Calibrate, shortlist and close the loop

1. Have two reviewers score independently, then reconcile differences on the high-stakes fields first, since those carry the most weight and the most interpretation. Record why each agency advanced or did not in a sentence or two; this is the document you will need when a stakeholder asks why their preferred firm is absent. Tell every non-shortlisted agency the outcome and, where you can, the deciding factor. Agencies remember how a review was run, and future participation depends on it.

How GPI's Evaluation Criteria Fit Into Your RFI

A good agency RFI template is short, decision-led and evidence-focused, and it leaves strategy and quotes to the RFP. That is what ANA and 4As members asked for when they called for the RFI to return to a screening role, and it is what decision-first question design produces when applied to agency selection.

GPI assesses agencies on documented evidence, stated methodology and acknowledged limitations. The high-stakes fields in this template ask agencies for the same three things: what they did, how they measured it and what the measurement could not show. Applying one lens to RFI responses and to directory listings means your shortlist and your long-list research rest on the same standard. The directory categories and buyer guides can supply long-list candidates and capability questions; they cannot promise how any agency will perform for you.

The RFI verifies nothing on its own. It identifies which claims need verification, and the RFP, reference calls and due diligence carry that work. To check your RFI against the evidence categories GPI uses to assess agencies, read the Growth Partner Index methodology.

Frequently Asked Questions

How many agencies should receive an agency RFI?

Enough that the screen removes some, and few enough that every response is read closely. None of the sources reviewed for this guide sets a numerical benchmark. Start from the long list your public check produced, and if it exceeds what two reviewers can score in the time available, tighten the public check rather than skim the responses.

Should the RFI ask for pricing at all?

Ask for fee structures, minimum engagement and exclusions rather than a quote. That keeps the document on the screening side of the line procurement guidance draws between RFI and RFP, and it still answers the budget fit question. A quote requires scoped work and belongs in the RFP.

Can we combine the RFI and RFP into one document to save time?

You can, but you trade a short screen for a long proposal from every agency on the long list, including those that would have failed basic fit. The ANA and 4As surveyed members on both sides and found agreement that the RFI had drifted toward the RFP and needed simplifying. Combining the documents formalizes that drift and shifts unpaid work onto agencies that cannot win.

How long should agencies get to respond to an RFI?

Estimate the working days needed to complete the fields, tables and evidence requests. The captured sources provide no agency-specific response-time benchmark. Responsive's 2025 cross-industry report found more than 75% of response teams under pressure for faster, more personalized answers; account for that workload when setting the deadline.

What if an agency refuses to share case study metrics at RFI stage?

Treat it as a follow-up flag rather than a disqualifier. Client confidentiality is a legitimate reason, so ask for anonymized or ranged figures with the metric definition, period and the agency's role intact. An agency that cannot supply even that for any engagement is showing you how it documents its work.

How do we handle incumbent agencies in an RFI?

Send the identical document, apply the identical deadline and score on the same rubric. Record where the incumbent has an information advantage, such as knowledge of your budget or data, and raise it in the calibration discussion so reviewers can judge whether a strong answer reflects capability or access. Treating the incumbent differently in either direction undermines the shortlist you are trying to defend.