How do you evaluate a marketing agency before you sign?
Agency churn peaks at agencies that sell tactics, not systems. Use this checklist before you sign—then apply if you want fixed-fee infrastructure.
Agency churn is highest at exactly the agencies most service businesses hire. The questions that actually predict whether one will work out—before you sign.
Most operators evaluate a marketing agency the way they'd evaluate a vendor: case studies, a pitch deck, a reference call or two, a gut check on whether the people seem sharp. All reasonable. None of it is what actually predicts whether the relationship survives past year one.
This guide is for $1M–$10M service operators who already spend on marketing and want a practical checklist—not another vanity deck. The businesses that get burned usually got burned by a structure problem, not a talent problem. Knowing what to check before you sign is the cheapest insurance available.
Key Takeaways
- Evaluate systems, not tactics. Channel specialists and pitch decks fail when nobody owns the full revenue path.
- Ask for revenue metrics first. Qualified leads, cost per lead, and attribution beat impressions and engagement.
- Score strategy, measurement, ops, accountability, and commercial model—five criteria, not gut feel.
- Fixed-fee vs hours matters. Pricing should reward outcomes and clarity, not padded time.
- Proof before pitch. Ask for case patterns and exit honesty before you believe the deck.
- If you already have an agency, score them the same way—churn predictors show up in reporting long before the breakup.
- Structure precedes scale. If they won't share how the system compounds, keep shopping.
What “good” looks like for a service business
A good marketing partner for a service business does not “run ads” or “do SEO” in isolation. They install and operate a revenue system: demand that brings qualified volume, authority content that compounds trust, and conversion paths that turn inquiries into booked work—measured against revenue, not vanity dashboards.
That is different from a channel vendor. Channel vendors optimize the metric they own. You pay three invoices, get three stories, and still cannot answer whether marketing is buying pipeline or noise. Good looks like one accountable owner for the outcome, shared definitions of a qualified lead, and reporting that connects spend → lead → close → revenue.
If your business runs on calls, estimates, and booked jobs, “good” also means ops discipline: response speed and creative built for operators—not awards decks. Structure precedes scale. More spend on a weak structure amplifies volatility.
The evaluation framework (five criteria you can score)
Use this as a scorecard before you sign—or against your current partner. Rate each criterion 1–5. Anything under 15 total is a structural risk, not a “give them another quarter” problem.
1. Strategy: systems over channel menus
Ask whether they lead with a documented plan for how demand, authority, and conversion work together—or with a menu of tactics. A systems partner names the revenue constraint and how the next 90 days address it. A tactics partner starts with keywords and creatives and never names the business problem.
2. Measurement: metrics that predict revenue
Insist on marketing metrics that predict revenue—qualified leads, cost per lead, close rate visibility, and revenue influence—not impressions, reach, or engagement as the headline. If revenue metrics are “available on request,” they are not the operating system. They are theater.
Ask how they handle attribution for service businesses in the real world: multi-touch journeys, phone calls, form spam, and offline closes. Perfect multi-touch models are rare. Honest, operator-usable attribution is not optional.
3. Creative and ops: work that survives contact with the calendar
Evaluate whether creative, landing paths, and follow-up assume a busy ops team. Ads that dump unqualified leads into a slow inbox are not wins. Ask who owns lead routing, response SLAs, and the loop when sales says “these leads are junk.” Good agencies treat that as fuel; weak ones treat it as an excuse.
4. Accountability: one owner when revenue stalls
In a fragmented setup, every vendor owns their own metric and nobody owns revenue. The same failure can exist inside one agency if paid, SEO, and content never share a number. Ask: if revenue stalls, who is accountable, and how do they diagnose why? Straight answers show up in process and case patterns—not kickoff enthusiasm.
That is why one agency vs multiple vendors is an evaluation question, not a brand preference. Fragmentation is a structural choice with predictable churn.
5. Commercial model: what the fee rewards
Hourly billing rewards time spent. Percent-of-spend models can reward budget inflation. Fixed-fee, scoped pricing removes the incentive to pad hours—but only if reporting and scope are honest. Ask what happens to price when the business grows or scope changes. No answer usually means change orders are coming.
Score these five. Write the scores down. Pitch chemistry is not a sixth criterion.
Questions to ask before you sign
Skip the case-study montage for a moment and ask these in writing. The answers predict the relationship better than the deck.
What do you report on, and does it tie to revenue?
If the answer leans on impressions, reach, engagement rate, or “brand awareness,” that is an agency reporting on what is easy to show. Ask whether qualified leads, cost per lead, and revenue attribution are in the standard report—or whether you would have to ask for them separately, every month, forever.
Who owns the number that matters?
Ask directly: if revenue stalls, who is accountable across channels, and how do they find out why? If ownership fragments by channel with no roll-up to pipeline, you are buying activity, not outcome.
What is the pricing model, and what does it reward?
Ask whether the model rewards effort (hours) or clarity of outcome (scoped system work). Ask what happens when scope expands. Ask whether ad spend is marked up. Ambiguity here is not sophistication—it is risk.
What happens if it does not work?
Every pitch describes the upside. Ask about exit process, notice period, and—honestly—how many clients they lost in the last year and why. Specific, non-deflecting answers beat polished retention narratives.
How do you know if my current marketing agency is any good?
If they cannot help you evaluate your current setup without immediately pitching replacement, treat that as a tell. A serious partner can diagnose structure first. Tools and scorecards that force clarity without a sales call are a positive signal; pressure to “hop on a discovery call” before sharing how they work is not.
Red flags that predict churn
Industry churn research puts the risk in sharper focus than anecdotes. Small agencies (10 or fewer employees) run roughly 32% annual churn. Project-based engagements churn at 42% versus 18% for ongoing retainer relationships. Single-channel specialists run hottest: PPC-only shops see roughly 49% annual churn, while full-service agencies churn the least of any service type at roughly 25%. (Focus Digital, Average Marketing Agency Churn: 2026 Report.)
Clients do not usually leave because the work was bad. They leave because they could not tell whether it was working. Research into why agencies get fired points less at incompetence and more at misaligned goals, reporting the client did not trust, and a strategy nobody outside the agency could explain in plain language. (The DSM Group, Why Marketing Agencies Get Fired.)
Operational red flags that show up before the breakup:
- Reporting decks full of impressions with no line to revenue or qualified leads
- Vague ownership when a channel underperforms
- Open-ended hourly billing with no natural efficiency incentive
- Strategy that cannot be restated by the operator in one paragraph
- “We’ll figure measurement out after launch”
- Refusal to discuss exits, notice, or recent churn with specifics
Clients don't usually leave because the work was bad. They leave because they couldn't tell whether it was working.
How to score your current agency (if you already have one)
You do not need a new pitch to know whether your current partner is working. Run the same five-criterion framework against last quarter’s reality.
Ask yourself:
- Can you point to qualified leads, CPL, and revenue influence in the last report without digging?
- If pipeline dipped, did someone own the diagnosis across channels—or did each vendor defend their silo?
- Are you paying for a compounding system or a stack of disconnected deliverables you will replace piece by piece?
- Has the commercial model stayed clear as scope shifted—or has it drifted into change orders and unexplained hours?
- Could you explain the strategy to a peer operator in under two minutes?
If you cannot answer those cleanly, the problem is already visible. “How to tell if your marketing agency is working” is not a vibe check—it is whether revenue accountability exists in the operating rhythm.
An agency relationship is also a unit-economics problem: fee versus durable output over the full engagement, not first-quarter excitement. The same discipline behind CAC and LTV applies to agency spend itself: are you buying infrastructure that compounds, or renting activity?
Before you renew or replace anyone: take the free Revenue System Scorecard (about four minutes, no sales call) to see where structure is actually constraining you—or review how we price fixed-fee engagements if you want commercial clarity without a pitch theater.
Fixed-fee vs retainer / percent of spend (commercial clarity)
Commercial model is not a footnote. It is an incentive system.
| Model | What it tends to reward | What to watch |
|---|---|---|
| Open-ended hourly | Time spent | Padding, unclear scope, surprise invoices |
| Percent of ad spend | Larger budgets | Spend inflation without efficiency pressure |
| Vague “retainer” | Availability theater | Deliverables that never map to revenue |
| Fixed-fee, scoped system | Clarity and completion of agreed work | Still requires honest reporting and scope discipline |
Fixed-fee is not magic. Paired with vanity reporting, it is still a bad buy. Paired with revenue metrics and a documented system, it removes the structural conflict where the agency profits from inefficiency.
If you want the transparent version of how Prime packages this—Growth Blueprint first, then managed Growth Engine, ad spend never marked up—see fixed-fee pricing. Same price today as tomorrow; no discovery-call pricing theater.
Process and proof: evaluate the work, not the promises
Proof should show system patterns, not isolated stunts. Look for rescue-then-relaunch, infrastructure-from-zero, and controlled expansion—and ask what was measured the same way across them.
Browse the marketing case studies index for how operators in different verticals installed demand, authority, and conversion as one system. For a durable channel build in junk/hauling, see the Junk Control case study. One study is a pattern sample; the index lets you compare fit without over-indexing on a single vertical.
Every engagement we run starts with a documented Growth Blueprint so you can evaluate strategy before execution. We report against revenue and qualified leads—not impressions. Accountability has to be built into the structure before you can trust the results.
Decide with a clear next step
If you are evaluating a first agency or a third replacement, do not start with chemistry. Start with structure: score the five criteria, demand revenue metrics in the standard report, and refuse commercial ambiguity.
When you want a structured second opinion—not another vendor deck—apply. Enrollment is selective (four new clients per month). If you are earlier in the decision, the Scorecard is the faster diagnostic.
Ready to replace guesswork with structure? Apply in about five minutes—hear back within 48 hours. Or take the Revenue System Scorecard first if you want a clear read on constraints before you commit.