How to Write a Market Research Brief or RFP (2026)

Last updated: 23 September 2026

Quick Answer: Write a market research brief or RFP around the decision, the audience and the results you need, and let vendors propose the method. US federal buyers work to the same rule: FAR 37.602 asks for required results, not how the work is done. Then add incidence, markets, deliverables, timeline, budget range and scoring weights.

The US State Department's contracting handbook states the principle plainly: a definitive work statement "is likely to produce definitive proposals," and it cuts the time needed to evaluate them. The same holds for a consumer study.

A brief says what the study must achieve; an RFP adds the rules of the competition, meaning the deadline, response format and scoring. Briefs that return unusable quotes usually prescribe a method and omit the facts no vendor can price without.

What Does a Research Brief Need Before a Vendor Can Price It?

Four facts, each stated as a number or a name rather than an adjective. Everything else in a brief improves the proposal; these four make a price possible at all.

• The decision and its deadline.
• Who qualifies, and how common they are.
• The markets and languages.
• What the vendor hands over at the end.

OMB's Standards and Guidelines for Statistical Surveys set a similar floor: a written plan naming "the decisions the survey is designed to inform" and the precision required, such as the size of difference to detect. "Compare two concepts" can be fielded with 60 people or 600, and only the size of difference to detect tells a vendor which.

How thin a brief can be depends on the delivery model. On a managed research service such as Alchemic, briefs range from one line to fully specified and researchers build the discussion guide from the brief in under an hour. No vendor can supply the decision or the audience, though.

The Research Brief Template, Section by Section

Under FAR 37.602, a federal statement of objectives must carry at least purpose, scope, period and place of performance, background, performance objectives and operating constraints, and the bidder proposes how to do the work. The template maps those elements onto a consumer study and adds the lines a market research vendor prices from. Items 1 to 12 make a brief; add item 13 to issue it as an RFP.

Copyable Research Brief and RFP Template

1. Background: the brand, the category, what prompted the study, and earlier research.
2. Decision: what will change, who decides, by when, and the options on the table.
3. Required results: the questions to answer and the precision needed, for example "detect a 10-point preference gap between two concepts."
4. Decision rule: the result that moves the decision each way.
5. Audience and incidence: who qualifies, screening criteria, the estimated share who qualify and its source, and quotas.
6. Markets and languages: countries, the languages respondents actually speak, and whether markets must be compared.
7. Reach requirement: segments that must appear in the completed sample, and how coverage will be checked.
8. Sample source: vendor recruitment, your customer list or both, and whether your list carries consent to contact.
9. Stimulus: concepts, ads, packs or prototypes, in the format respondents will see.
10. Method: constraints only, plus an invitation to propose the design.
11. Deliverables and timeline: the readout format stakeholders use, the data you keep, and the readout date.
12. Budget and constraints: a budget range, data protection and consent rules, and data ownership.
13. Proposal instructions, RFP only: response deadline, page limit, weighted evaluation factors, and whether non-price factors outweigh price.

What Will Vendors Ask if the Brief Is Silent?

Whatever the quote depends on. If the answer is missing, each bidder assumes one, and the quotes stop being comparable. Answer these in the brief, or share every answer with all bidders.

• How many people do we screen to find one who qualifies? Incidence can move a quote more than any other line, as the breakdown of market research costs and pricing models shows.
• Is the sample ours to find or yours to supply? A customer list changes recruitment cost and raises consent questions.
• Which languages, and who reads the transcripts? Translating after fieldwork is not interviewing in the respondent's language.
• Is there an incumbent, and can we see the last wave? A tracker needs continuity with earlier questions and samples.

These run opposite to the questions to ask a research vendor about respondent reach. Those test the vendor; these test the brief.

How Do You Score Proposals on Reach, Not Features?

Publish the weights before proposals arrive, and weight who ends up in the data above what the dashboard shows. FAR 15.304 makes federal buyers state every evaluation factor and its relative importance, including whether non-price factors combined are significantly more important than, approximately equal to, or significantly less important than price.

Rows are ordered by suggested weight, highest first. Adjust the weights to the decision.

Reach takes the top weight because a missed segment fails silently. The OMB guidelines call coverage above 95 percent, overall and in each major group, desirable, and ask for a bias evaluation below 85 percent. A commercial proposal can at least name the segments its channels miss and say whether it offers a probability or non-probability sample.

Channels decide who can take part, since people who ignore survey links never reach a browser study's screener. Alchemic runs interviews natively inside WhatsApp with no link and no app, places AI phone interviews to any number in supported regions, and publishes 57+ languages including Hindi, Tamil and Telugu, with managed fieldwork or bring your own across 14 markets including the USA and the UK. For US adults who already answer email surveys, a browser panel may reach everyone the decision needs and should score just as well.

Price belongs in the score without overriding it. Defense rules tell contracting officers to avoid, where practicable, lowest-price-technically-acceptable selection for knowledge-based professional services, and to use it only when exceeding the minimum adds little or no value.

Which Brief Mistakes Inflate the Quote?

Four recur, and each makes bidders price risk or unneeded work.

1. Prescribing method and sample size. Order 400 online completes and you pay for 400, even if 150 would answer the question. FAR 37.602 steers federal buyers away from specifying how the work is done or the hours provided.
2. Adjectives instead of standards. "Robust" and "representative" mean something different to each bidder. The State Department handbook warns that words such as "adequate" tend to decrease, not increase, the contractor's obligation. Use a quota, a coverage check or a precision target.
3. Deliverables nobody reads. Contractors price in every required report, the same handbook notes. Ask for the readout stakeholders use and the raw data.
4. Several studies in one brief. Every extra objective lengthens the screener and the interview. Split unrelated questions into separate briefs or waves.

When Should You Skip the Formal RFP?

When bidding would cost more than the study or break something it depends on. A brief is always worth writing; a competitive RFP is not.

• Small or urgent studies. Bidder Q&A windows and scoring can outlast the fieldwork. On Alchemic's AI-moderated interviews, a 200-interview qualitative study runs brief to live dashboard in 3 days, 5 to 7 for complex work, so a direct brief to one or two vendors is faster.
• Trackers with an incumbent. Switching mid-series can break the trend. Re-bid at a natural break, with a bridge wave.
• A decision still being argued. No template fixes a brief whose decision is unsettled. Settle it first, or run a small exploratory study to frame it.

Scoring has a limit too: reach promises stay claims until a paid pilot or first wave reports completes by segment.