A Google Ads agency RFP is the difference between a $300k 12-month commitment that delivers and one that disappoints. The 30 questions below cover the six categories that actually predict agency quality: account ownership, fee transparency, strategy methodology, tracking, operations, and exit. Each question has clear right and wrong answers. Run the framework objectively and the agency that wins your business is the one that will actually serve it well.
This guide is for advertisers preparing an RFP, evaluating proposals, or doing due diligence on an incumbent. We cover all 30 questions, the scoring framework, immediate disqualifiers, and reference-check protocols. Before issuing the RFP, baseline your current account performance with our free 5-axis Google Ads audit — knowing your starting point makes agency proposals concrete instead of abstract.
Updated 2026-05-09 with current AI tooling, Performance Max, sGTM, and Consent Mode v2 expectations.
- 30 questions across 6 categories: ownership, pricing, strategy, tracking, operations, exit.
- Score 0-3 per question, maximum 90 points. Pass at 70+, disqualify below 50.
- 5 immediate disqualifiers: refused account access, no itemized fees, no named senior, no written exit, no PMax brand exclusion expertise.
- 3 reference checks minimum per shortlisted agency, including 1 churned client.
- Paid 90-day trial preferred to long-term contract; quality agencies welcome trials.
Why RFPs matter in 2026
Three forces make 2026 the year RFPs matter more:
1. Agency quality variance widened. AI tooling, Performance Max, and privacy shifts have produced a wider gap between top agencies and median agencies. The 2022 method of picking on price or referral routinely produces costly mismatches.
2. Switching costs rose. A bad agency choice means 6-9 months minimum (90 days to discover, 90 days to switch). The cost of a wrong choice on a $50k/month account is $300k-$500k in lost revenue. Diligence is cheaper than recovery.
3. Standard contracts are inadequate. Off-the-shelf agency contracts often favor the agency on critical points (account ownership, exit, asset transfer). RFP responses surface the bargaining points before signing.
The 30-question framework gives you the structured leverage to evaluate agencies on the dimensions that matter — not just on price or chemistry.
How to use the 30 questions
Step 1 — Issue the RFP to 5-7 candidate agencies. Allow 10-14 days for response. Required format: written response, no marketing fluff, specific answers to each question.
Step 2 — Score independently. Two reviewers score each agency separately, then reconcile. Avoid groupthink in scoring; differences in independent scores often reveal the most important questions.
Step 3 — Shortlist top 3. Above 70 points and no disqualifying answers. Schedule 90-minute presentations with the proposed senior strategist (not just sales).
Step 4 — Reference check the shortlist. Three references each, including one churned client. Document specific questions per reference.
Step 5 — Final decision. Combine RFP score + reference scores + presentation impression. Negotiate contract with the winner using gaps identified during the process.
Questions 1-6: Account ownership & access
1. Will I own my Google Ads account, or will you? Acceptable: "You own. We have admin access via MCC link." Disqualify: "We own; you have user access only."
2. Will I have admin access on day one of the engagement? Acceptable: "Yes, immediately upon contract signing." Disqualify: "After 90 days" or "viewer access only".
3. Who owns the sGTM container, GA4 property, and BigQuery dataset? Acceptable: "You own all of these." Acceptable conditional: "We host on your Google Cloud account." Disqualify: "We own; transferred to you at termination."
4. Will I have access to all third-party tools used on my account (Optmyzr, SEMrush, etc.)? Acceptable: "Yes, view-only or admin as appropriate." Mediocre: "View-only summary reports." Avoid: "We don't share tool access."
5. Can I export all my historical data at any time? Acceptable: "Yes, on demand. Standard data exports include all reports, custom audiences, search terms history, and creative inventory." Disqualify: "Available at termination only" or "Custom data exports billed separately."
6. What is your MCC architecture for managing my account? Acceptable: clear description with sample. Disqualify: vague answer or refusal to describe.
For broader account architecture context, see our multi-account MCC strategy guide.
Questions 7-12: Pricing & fee transparency
7. Provide an itemized fee structure separating: management retainer, tracking setup, creative production, platform fees, and any other charges. Acceptable: clear table with specific dollar amounts per line item. Disqualify: lump sum without breakdown.
8. What is the management fee model — flat retainer, % of spend, or hybrid? Acceptable: clear answer with rationale for the chosen model at our spend tier. Mediocre: "It depends" without specifics.
9. What's included in the management retainer, and what's billed separately? Acceptable: bullet list. Disqualify: vague "comprehensive management."
10. What is your fee floor (minimum monthly fee regardless of spend)? Acceptable: specific dollar floor (e.g. $2,500/month minimum). Acceptable: "No floor on accounts above $X spend." Concerning: "No floor" with low % of spend (under-resourcing risk).
11. How do you handle scope changes that exceed standard retainer? Acceptable: clear hourly or project rate; written approval before work. Disqualify: post-hoc invoicing without approval.
12. What's your typical fee evolution over a 24-month engagement? Acceptable: tied to spend growth or scope expansion. Disqualify: arbitrary annual increases without justification.
For comprehensive fee benchmarks, see our agency cost guide 2026.
Questions 13-18: Strategy & methodology
13. Walk me through your approach to Performance Max for an account in our vertical. Acceptable: specific answer with brand exclusion, asset group structure, audience signals, and incrementality measurement. Disqualify: generic "we follow Google's recommendations."
14. How do you decide between Performance Max and Standard Search for a given goal? Acceptable: framework based on vertical, spend tier, attribution clarity, and creative inventory. Disqualify: "We use Performance Max for everything" or "We avoid Performance Max."
15. Show me 3 search-terms exclusion lists you've built recently. Acceptable: anonymized examples with rationale. Disqualify: refusal or generic placeholders.
16. What's your approach to RSA testing and how often do you refresh creative? Acceptable: specific cadence (typically monthly for active accounts) with measurement framework. Disqualify: "When Google flags low-quality assets" (reactive, not proactive).
17. How do you measure incrementality, and how often? Acceptable: specific methods (geographic hold-out, time-based, brand exclusion) with cadence (quarterly minimum). Disqualify: "We use Google's reporting" (Google's reporting is not incrementality).
18. What AI tooling does your team use, and what tasks does it handle? Acceptable: specific tools (LLMs for ad copy, Python scripts for management, custom dashboards) with examples. Disqualify: "We use Google's automated bidding."
For detailed PMax methodology, see our complete PMax guide 2026.
Questions 19-24: Tracking & reporting
19. Will you deploy server-side GTM, and what's the timeline and cost? Acceptable: clear plan with timeline (typically 4-8 weeks) and cost ($5k-$25k). Disqualify: "Not necessary" for accounts over $50k/month spend.
20. How will you implement Enhanced Conversions for Web and Leads? Acceptable: specific plan with hashed identifier strategy and offline upload pipeline. Disqualify: "We'll look into it later."
21. How will you connect my CRM to Google Ads for offline conversion uploads? Acceptable: specific tooling (native API, Zapier, custom script) with daily upload cadence. Disqualify: "Manual upload monthly."
22. What's your reporting cadence and structure? Acceptable: weekly tactical summary, monthly executive report, quarterly business review with named senior. Disqualify: "Dashboards only" without human review.
23. How will you measure our true ROAS — last-click, branded-excluded, or incremental? Acceptable: all three views with monthly reporting. Disqualify: last-click only.
24. How will you handle Consent Mode v2 for our EU traffic? Acceptable: specific plan with CMP integration and modeled conversion verification. Disqualify: "Not in scope" if account has EU traffic.
For tracking foundation guidance, see our server-side tracking guide.
Questions 25-30: Operations & exit
25. Who will be my dedicated senior strategist, and how much time will they spend on my account weekly? Acceptable: named individual with bio, with specific weekly hours (typically 6-15 hours for accounts $50k-$200k/month spend). Disqualify: "Team-based, no single owner" or "We'll assign post-signing."
26. What's your typical client retention rate, and what's your turnover by senior strategist? Acceptable: specific numbers with context (e.g. "85% retention; senior turnover 15%/year"). Disqualify: "We don't track that."
27. What is your standard contract term, and what's the exit notice period? Acceptable: 6-12 month term with 30-60 day exit notice. Disqualify: 18+ month lock-in or 90+ day exit.
28. What is your asset transfer protocol at termination? Acceptable: specific list of exports with timeline (typically 14 days post-notice). Disqualify: vague "we'll cooperate."
29. Do you have any non-compete or non-solicitation clauses I should know about? Acceptable: limited or none. Concerning: broad NDAs preventing knowledge transfer to next agency.
30. Provide 3 client references at similar spend tier — including at least one client who has terminated with you. Acceptable: 3 references with contact info. Disqualify: refusal to provide churned-client reference.
For switching protocol if the agency engagement doesn't work out, see our switching playbook.
Scoring framework
Pass thresholds:
- 75-90 points: top-tier agency, proceed to references
- 60-74 points: viable, proceed with caution
- 50-59 points: weak, only consider if pricing is highly favorable
- Below 50: disqualify
Mandatory minimum scores per category:
- Account ownership: 12+/18 (any 0 disqualifies)
- Tracking: 12+/18 (any 0 disqualifies)
- Other categories: no individual question must score 0
To estimate financial impact of agency quality differences, run our wasted ad spend calculator and verify Quality Score health post-onboarding with our Quality Score checker.
Red flag answers (instant disqualification)
Five answers that warrant immediate disqualification regardless of other strengths:
1. "We own your Google Ads account." Account ownership lock-in is the single biggest 2026 contractual risk. Walk away.
2. "Our fee is all-inclusive — we don't itemize." Bundling concealment usually means hidden costs surface post-signing.
3. "Your senior strategist will be assigned after contract signing." Quality agencies name the strategist in the proposal. Anonymous staffing means whoever is available when you sign.
4. "We don't offer written exit clauses." No exit clause = unlimited lock-in. Reject.
5. "We use Google's recommendations for Performance Max." A 2026 agency that doesn't have specific PMax expertise (brand exclusion, asset group surgery, hold-out testing) is a 2022 agency. PMax mistakes cost more than fee savings.
Agencies that respond with "we'll figure that out together once we know your account" are dodging the question. Quality agencies have opinions and frameworks that they apply to the client's situation. Vague flexibility is usually low expertise. The right answer combines methodological clarity with client-specific adaptation, not pure ad hoc.
This Google Ads agency RFP framework is updated quarterly by SteerAds. Last update: 2026-05-09. The 30-question structure is based on 2025-2026 panel data from over 100 agency selection processes across USA, UK, FR/DE, and GCC markets.
For complementary reading, see our agency cost benchmarks 2026, our agency audit framework, and our agency switching playbook. To baseline your account before issuing RFP, request our free 5-axis Google Ads audit, model spend exposure with our wasted ad spend calculator, and for enterprise multi-account RFPs reach out via our contact form.
Sources
Official sources consulted for this guide:
FAQ
What questions should I ask a Google Ads agency before hiring?
The 30 essential questions cover six categories: account ownership and MCC access (6 questions), pricing and fee transparency (6), strategy and methodology (6), tracking and reporting (6), and operations and exit (6). Each question has objectively right and wrong answers; agencies that dodge questions or give vague answers in any category should be disqualified. Quality agencies welcome detailed RFPs because they differentiate themselves; agencies that resist suggest they have something to hide. Use the structured 30-question RFP for any agency engagement above $30k/month spend.
How long should the RFP process take?
A standard RFP process takes 21-35 days end-to-end: 5 days to issue and clarify questions, 10-14 days for agency responses, 5 days for shortlist and second-round questions, 3-5 days for final presentations, 2-3 days for reference checks and decision. Compressed timelines under 14 days produce poor agency responses (rushed, generic) and don't allow for meaningful reference checks. Extended timelines over 60 days lose quality agencies who decline due to opportunity cost.
Should every agency engagement use a formal RFP?
No — RFPs make sense above $30k/month spend or for multi-account/multi-region engagements. Below $30k/month, a structured discovery call with the same 30 questions covered conversationally is sufficient. Above $200k/month, the RFP often expands to 50+ questions with formal scoring committees and multiple presentation rounds. The threshold isn't arbitrary; it correlates with the dollar value of getting the decision right.
What are immediate disqualifying answers in an RFP?
Five answers should disqualify an agency immediately: (1) refusal to grant client admin access to the Google Ads account; (2) refusal to provide an itemized fee structure separating management, tracking, creative, and platform fees; (3) inability to name the senior strategist who will own the account; (4) refusal to commit to written exit clauses with 60-day notice; (5) inability to demonstrate Performance Max brand exclusion expertise. Each indicates fundamental gaps that won't improve once contracted.
How do I score 30 RFP responses objectively?
Score each question 0-3 points: 0 for missing/refused, 1 for vague, 2 for clear, 3 for excellent. Maximum 90 points across 30 questions. Quality agencies score 70+; consider only agencies above 60. Below 50 = disqualify. Weighting is built into the framework: account ownership and tracking questions are 'must-pass' (any 0 score disqualifies); pricing and operations are scored normally; strategy and methodology are scored with senior-strategist input. Document scores per question for transparency.
Should the RFP include a paid trial period?
Yes, ideally a 90-day trial at standard contracted rate with clear exit at trial end. Quality 2026 agencies offer paid trials because: (1) it lets the client assess fit without long-term commitment; (2) it forces the agency to deliver in the first 90 days, when 80% of long-term value is established; (3) it filters out agencies who can only sell on long contracts. Avoid 'free trials' — they signal under-staffing and rarely deliver real work. Standard fee for 90 days, with documented success criteria.
How important are reference checks vs the RFP responses?
References are usually more predictive than RFP responses. References should: (1) be at similar spend tier and vertical; (2) be 6-18 months into the engagement (long enough to know, recent enough to be relevant); (3) include at least one client who has churned from the agency. References from current happy clients are easy; agencies that won't share past or churned clients are signaling weak stories. Always validate at least 3 references per shortlisted agency.
What questions reveal the agency's AI tooling sophistication?
Five questions reveal AI sophistication: (1) Which AI tools does your team use daily — and for what tasks? (2) How do you use LLMs for ad copy generation, and how do you ensure quality vs hallucination? (3) Do you use script automation or Python for account management — and at what scale? (4) How is your team trained on prompt engineering? (5) Show me one example of an AI-driven optimization deployed in the last 90 days. Agencies stuck on 2022 tooling will struggle to answer these concretely. Quality 2026 agencies have explicit AI workflows.
Should I ask agencies to show their MCC structure?
Yes, with anonymization. A quality agency can describe their MCC architecture: how clients are organized, how reporting is structured across accounts, how they handle billing and budget control. They should be able to walk through a sample (anonymized) account showing the level of granularity in campaigns, ad groups, audiences, and reporting. Agencies who can't or won't show structure suggest they manage all clients identically (which is a bad sign) or that their structure is genuinely poor.