How to Evaluate a PPC Agency: 10 Questions to Ask Before Signing

Evaluating a PPC agency is a diagnostic, not a vendor selection. The questions that separate a real practitioner from a pitch deck are the ones that force specifics: who owns the account and the data, how the agency measures non-brand performance, who is actually in the account on a Tuesday morning, what the last 30 days of change-log activity look like, and what the agency would change in your first month. Order matters. Measurement and ownership come first, because every downstream claim depends on them. The ten questions below are designed to surface diagnostic thinking and to expose the gap between trained pitch language and the vocabulary of a working paid media lead.

This guide walks through what makes agency evaluation hard, the ten questions in the order they should be asked, what a diagnostic answer sounds like for each one, and the two-meeting sequence that uses a free audit as the central evaluation tool. The article is built from portfolio pattern recognition across $30M+ in monthly managed ad spend at Opascope. It is also written for CFOs and finance leaders who sign off on agency contracts but do not run the day-to-day relationship, since these ten questions translate directly into the contract terms, ownership clauses, and measurement standards a finance partner can defend. For PE-backed companies and portfolio operators, the same framework applies when evaluating an inherited agency at a portfolio company or qualifying a vendor before a post-close switch, where the cost of an unexamined agency relationship usually shows up as wasted spend before it shows up in the P&L.

Agency evaluation is hard because most agencies train for pitch calls, not diagnostic ones. Every agency website lists the same attributes: senior team, custom strategy, transparent reporting, proactive communication. None of those claims carry information by the time you reach a shortlist of three. The information lives in how an agency answers a specific question with a specific number, not in how its homepage reads.

The asymmetry of the situation works against the buyer. The agency has run dozens of pitch calls and knows the standard questions. The buyer has usually run fewer than five total agency evaluations across an entire career, often after getting burned twice by previous agencies. The result is a conversation where the agency has a script and the buyer does not.

Quick takeaways:

• Every agency website claims the same four attributes. By the shortlist stage, those claims carry no information.
• A pitch answer sounds hedged and abstract: “usually,” “typically,” “depending on the account.”
• A diagnostic answer sounds specific: named people, numeric thresholds, willingness to say what is currently broken in your account.
• The questions below are designed to force a demonstration, not a description.

A senior paid media team should be able to answer every question with a number, a name, or a specific change. Vagueness in any of these answers is a signal in itself.

Ask these ten questions in order. Measurement and ownership come before team. Team comes before tools. Tools come before process. The order matters because the answers cascade: an agency that hedges on account ownership in question one rarely produces a credible answer for the change log in question four.

1. Who owns the account, the pixel, and the conversion data?
2. How do you measure non-brand performance, and what is your floor?
3. Who will be working on my account day to day, and how many other accounts do they manage?
4. Can I see the change log from a comparable account you currently run?
5. What does your free audit actually find, and will you name dollar impact on each line?
6. What is your 12-month client retention rate?
7. How do you think about the parts of my program that are not inside Google Ads?
8. What is the contract length, the cancellation clause, and what happens to the data at offboarding?
9. When was the last time you told a prospect they did not need PPC?
10. What is one specific thing you would change in my account in the first 30 days, and why?

Account ownership is the single largest switching cost in a bad agency relationship. The right setup puts the Google Ads manager account, the pixel, the GA4 property, and the conversion actions on the buyer’s side, with the agency granted access. The wrong setup is any configuration where the agency owns one of those assets and hands you reporting extracted from its environment.

A good answer sounds like: “Everything lives in your Google account. We have access. If you fire us, you change our access and nothing breaks.” A pitch answer hedges with “we manage that for you” or “we have a system for handoff” or “we transfer everything at the end of the engagement.” Historical conversion data and audience lists are non-trivial assets. Losing a 12-month remarketing list is a material loss at scale.

Quick takeaways:

• Insist on owning the MCC, GA4 property, pixel, and conversion actions. The agency gets access, not ownership.
• Server-side tracking and CAPI setups extend the question. If the agency built a GTM server container, find out who pays for hosting after you leave.
• Across portfolio audits, incomplete data handoffs from prior agencies are one of the most common findings during onboarding. Plan for it.

Non-brand performance is the only part of a Google Ads program that is fully attributable to the agency. Branded search would happen without any agency in the seat. An agency that cannot articulate a non-brand performance floor has not thought about where its own value shows up in the account.

A good answer names a numeric floor and describes what happens when the account drops below it. For most ecommerce, a non-brand ROAS floor of 1.5x is a reasonable starting line, with cost per qualified lead benchmarked against target CAC for B2B. A pitch answer talks about ROAS generally or describes “holistic measurement” without naming a threshold. The follow-up that separates the two is direct: “What do you do when non-brand goes below that floor?” The diagnostic answer has a playbook. The pitch answer says it would “look at it.”

Quick takeaways:

• A diagnostic agency volunteers the distinction between platform ROAS and blended ROAS. A pitch agency conflates them.
• Across portfolio audits, attribution is broken or misconfigured in roughly half of accounts before remediation.
• Watch for any answer that defines success without a number. Numbers are how non-brand performance gets defended.

For an example of what reattribution work changes when the measurement layer is fixed, see how cost per paying customer dropped after attribution rebuild at a cloud infrastructure platform.

Ask for named people and an account ratio. The people who pitch the account are almost never the people who run it. An account-to-analyst ratio above 8:1 means the person in your account has fewer than 25 hours a month to spend on it, which is not enough to run a well-managed Google Ads program at scale.

A good answer is specific: named day-to-day account manager, LinkedIn profile, account ratio, named senior escalation path. A pitch answer reaches for “our senior team” or “you will work directly with our specialists” or “we have dedicated experts.” Both are forms of refusing to commit to a person. A published industry analysis of agency staffing flagged analysts managing 100+ accounts at roughly 24 minutes per account per week [1]. That is the floor of what bad ratios produce.

Quick takeaways:

• Ask for the day-to-day analyst by name, not just a role.
• Ask whether the seniority of the person in the account on Tuesday matches the seniority of the person on the pitch call.
• An account-to-analyst ratio under 8:1 is the practical quality line at the senior level.

The Google Ads change log is the single most useful proxy for agency quality [2]. An active account should show 20 or more changes per month across bids, negatives, ad copy, audience rules, and structural edits. A low change-log count in a live account is the clearest indicator of autopilot management.

A good answer shares anonymized screenshots of the change log from an account of similar size and vertical. A pitch answer says “we cannot share client data” and changes the subject. Legitimate confidentiality exists, but a partially redacted change log is almost always shareable. Refusing to show any usually means there is nothing to show. What to look for: variety, not just bid adjustments. Cadence at weekly or better. Structural work like campaign splits and audience updates, not only surface tweaks.

Quick takeaways:

• 20+ monthly edits in an active account is a working floor. Below that, the account is on autopilot.
• The mix matters more than the count. Bid changes alone are not management.
• A blanket refusal to share any change-log evidence is itself a signal.

The free audit is the most useful part of the evaluation, if you know how to grade it. A diagnostic audit surfaces specific findings with a dollar impact per line, names problems the buyer did not already know about, and tells the buyer which findings can be fixed without hiring the agency. A pitch audit lists generic observations, recommends a retainer as the only solution, and avoids numbers.

Grade the audit on three axes. Specificity: does it name your actual ad groups, your actual search terms, your actual conversion actions. Dollar impact: does each finding carry a numeric estimate of wasted spend, missed revenue, or conversion-rate gap. Honesty: does the agency tell you which findings you can act on without hiring them. The willingness to do that last one is the strongest single quality signal in the evaluation.

Quick takeaways:

• A well-run audit should close at a meaningfully higher rate than the industry average, which sits around 35%. The gap between the two numbers is the diagnostic.
• The buyer’s test: hand the audit to a smart friend. Can they act on it without further explanation.
• Generic observations and “book a strategy call” recommendations are the two most reliable tells of a pitch audit.

Client retention is the strongest trailing indicator of agency quality. Ask for 12-month retention as a percentage. Anything under 75% should give you pause. Anything above 90% is exceptional [3]. The number carries more signal than any case study because retention requires sustained performance across an entire client base, not a one-quarter highlight.

A good answer is a specific number with a willingness to talk about churn patterns. A pitch answer reaches for “most of our clients stay multiple years” or refuses to share a number. The follow-up that separates the two: “What is your churn pattern? When do clients leave?” The diagnostic agency answers. The pitch agency does not.

A PPC-only agency that stops at the edges of Google Ads will fix maybe 20% of the problem in most accounts. The biggest dollar leaks in mid-market programs sit outside the ad accounts: landing-page conversion rate, offline conversion imports for B2B, server-side tracking, channel mix, and creative capability. Ask what the agency does about each. Watch for the ones it cannot speak to.

In one DTC brand spending $1.4M a month on ads, roughly 10% of leads were lost to slow landing page load and conversion rate issues before the page was rebuilt. The rebuild cut top-of-funnel costs roughly in half. None of that work happens inside Google Ads. Channel mix is the other open frontier. Meta CPMs increased 30 to 40% year over year from 2024 to 2025 across managed accounts, which means an 18-month-old channel mix decision is almost certainly wrong by default.

Quick takeaways:

• Landing pages: does the agency build and test them, or does it hand you ad copy and hope.
• Offline conversion imports: for B2B, this is the difference between optimizing for form fills and optimizing for pipeline.
• Server-side tracking and CAPI: often described as “taken care of.” Verify with a screenshot of the actual conversion path.
• Channel mix: a working agency has a view on Meta vs. Google vs. emerging channels, not only on the channel it manages.

The contract is where an agency’s confidence in its own work shows up. Month-to-month with 30-day notice is the cleanest arrangement. Long-term lock-ins without performance outs usually mean the agency is selling duration rather than outcomes. Ask specifically about data at offboarding: the MCC stays with you, the pixel stays with you, the conversion setup stays with you.

A good contract spells out account ownership, data portability, and cancellation without friction. A pitch contract buries ownership terms, offers 12-month minimums, or frames cancellation as a negotiation. Historical performance data, audience lists, and custom conversion setups are the handoff items most often lost in bad offboardings. Ask in writing.

Ask this question as written. The answer reveals whether the agency has ever turned away revenue to preserve diagnostic integrity. An agency that cannot name a specific case has incentive alignment problems and will eventually face a conflict between your account performance and its retainer continuity.

A good answer names a specific situation. “We told a B2B company with a 14-month sales cycle and no offline conversion setup to fix their CRM first before spending more on Google.” A pitch answer falls back on “we always recommend what is best for the client” or redirects to case studies. The specific story beats every generic claim about putting clients first.

This is the last question, and it is the most revealing. By the end of an evaluation, the agency should have looked at the account enough to name one specific change and the reasoning behind it. If it cannot, either the audit was shallow or the team is not right for the engagement.

A good answer names a specific change and the mechanism. “Your non-brand campaigns have Advantage+ Audience enabled on retargeting. We would turn that off in week one because it is blending cold and warm audiences in your learning signal.” A pitch answer says “we would start with a full audit to understand the account” or describes a generic onboarding sequence. If the answer is good, it is also a free data point. You can act on it whether you hire the agency or not. For more on what falls apart in the first 30 days, see common first-30-day misses.

Use the ten questions across two meetings with a free audit in between. Meeting one: questions 1, 2, 4, 6, and 8. These force specifics on ownership, measurement, verifiable work, retention, and contract. Meeting two, after the free audit: questions 3, 5, 7, 9, and 10. These use the audit itself as the diagnostic and force the agency to commit to a specific first-30-day view.

Quick takeaways:

• Meeting one is a filter. Any agency that fails on ownership, measurement, or contract is out before the audit conversation begins.
• The free audit happens between meetings. Grade it on specificity and dollar impact per finding.
• Meeting two is a confirmation. By the end, the agency has either shown diagnostic thinking or has not.

What questions should I ask a PPC agency before hiring them?

The ten questions that matter most cover account ownership, non-brand measurement, day-to-day staffing, change-log depth in a live account, free-audit specificity, 12-month retention rate, full-system breadth outside Google Ads, contract and offboarding terms, whether the agency has ever told a prospect not to hire them, and what the agency would change in the first 30 days. Each question has a diagnostic answer and a pitch answer. The point is to separate the two.

How do I know if a PPC agency is actually good?

Three signals carry more weight than any pitch: client retention rate (above 90% is exceptional, under 75% is a warning), change-log activity in a comparable account (20 or more changes per month is normal in active management), and free-audit specificity (findings named in dollars with a willingness to tell you what you can fix without them). If all three are strong, the agency is good. If any one is weak, look at the next candidate.

Who should own the Google Ads account, me or the agency?

The advertiser should own the MCC, the GA4 property, the conversion actions, and the pixel. The agency gets access, not ownership. Any agency that owns one of those assets creates a switching cost the buyer pays at offboarding. Ask in the first meeting. A good agency confirms the setup immediately. A pitch agency hedges.

Should I get a free PPC audit before hiring an agency?

A free audit is the primary evaluation tool when graded properly. A good free audit names specific findings with dollar impact per line, identifies problems the buyer did not already know about, and tells the buyer which findings can be fixed without hiring the agency. A pitch audit uses generic observations and recommends a retainer as the only solution. The audit itself reveals more about agency quality than any case study can.

What is a good client retention rate for a PPC agency?

Twelve-month retention above 90% is exceptional. Above 75% is acceptable. Below 75% is a warning. Below 50% means the service model has a structural problem. Ask for the number directly. An agency that will not share it, or gives generic answers like “most clients stay multiple years,” is signaling something about its own data.

What contract terms should I avoid with a PPC agency?

Avoid multi-year minimums without performance outs, any contract without a clear 30-day cancellation clause, any language that gives the agency ownership of the account or pixel or conversion setup, and any handoff language that does not guarantee data portability at offboarding. Month-to-month with 30-day notice is the cleanest form. Agencies confident in their work do not need long lock-ins.

How can I tell if a PPC agency is managing my account or running it on autopilot?

Check the Google Ads change log [2]. In a live account, 20 or more changes per month is normal. Edits should be varied: negative-keyword additions, bid adjustments, ad copy tests, audience updates, structural edits. A low monthly change count in an active account is the clearest sign of autopilot management. Ask the prospective agency to show the change log from a comparable account before signing.

What is the difference between a PPC agency and a performance marketing agency?

A PPC agency manages paid search and paid social campaigns, usually Google Ads, Microsoft Ads, and the core paid social platforms. A performance marketing agency also covers landing pages, conversion-rate work, offline conversion imports, server-side tracking, creative production, and cross-channel attribution. For most mid-market buyers spending over $50K a month, the performance marketing agency is the better fit because the largest dollar leaks usually sit outside the ad accounts.

What is the difference between platform ROAS and blended ROAS in agency reporting?

Platform ROAS is what Google or Meta reports, calculated using that platform’s own attribution. Blended ROAS is total revenue divided by total ad spend across all channels. Platform ROAS is structurally biased upward because each platform claims credit for conversions it may not have caused. A good agency volunteers the distinction. A pitch agency conflates them. If reporting only shows platform ROAS, real performance is probably smaller than the dashboard suggests.

What is a red flag in a PPC agency’s pitch?

Watch for these: refusal to share account ownership terms, no numeric benchmark for non-brand performance, refusal to show a change log from a live account, free-audit findings without dollar impact, long-term contracts with no performance outs, and generic answers to the “what would you change in the first 30 days” question. Any one is a warning. Two or more means pick another agency.

Should I hire a big PPC agency or a small specialist?

The right question is not size, it is seniority. Ask who will be in the account day to day and how many other accounts that person manages. A small specialist with senior paid media leads usually outperforms a large agency where junior analysts run the execution. An account-to-analyst ratio under 8:1 is the practical quality line. Size is a proxy for seniority at best, and often not even that.

When the evaluation is done, book a PPC audit with the agency that gave the most diagnostic answers. The audit itself becomes the final test.

1. Conveyor Marketing Group, “Top 10 Questions to Ask Before Hiring a B2B PPC Agency”: https://www.conveyormg.com/blog/questions-ask-before-hiring-b2b-ppc-agency
2. Google Ads Help Center, “Change history”: https://support.google.com/google-ads/answer/2998031
3. Bruce Clay, “15 Questions to Ask When Hiring a PPC Agency”: https://www.bruceclay.com/blog/hiring-a-ppc-agency/