Choose a marketing agency by deciding what you are buying before you talk to anyone, then shortlist on proof that matches that specific need. Everything else — the pitch deck, the client logos, the awards wall — is noise until those two things are settled.
Most agency regret follows a predictable pattern. The buyer compares three agencies on reputation, picks the most impressive pitch, and discovers six months later that the delivery team is junior, the strategy is generic, and the original brief was never properly understood. This guide is built around preventing that sequence, not finding "the best agency" in the abstract.
TenetDecide what you are buying
Before you open a single proposal, answer one question: are you buying judgement, execution, or both?
Judgement means strategic diagnosis — someone who can look at your market, your constraints, and your current performance and tell you what the real lever is. Execution means reliable production: campaigns built, ads managed, content shipped, media bought, data reported. These are genuinely separable capabilities, and most agencies price them together even when they are only strong at one.
If you already know what to do and need a team to do it consistently, you are buying execution. If you are not sure what to do, or your current approach is not working and you cannot identify why, you are buying judgement. If you need both and have no internal marketing lead, you are effectively hiring a fractional marketing department.
Knowing which of these applies shapes every question you ask from here. A consultancy with senior strategists but thin production capacity is a strong buy for problem one and a poor buy for problem two. A well-run performance agency might be the opposite. Defining the job before you shop stops you from scoring agencies on the wrong criteria.
The questions that have diagnostic power
Generic question lists ask things like "what's your process?" Any agency answers that smoothly. The questions below are harder to rehearse because they force an agency to reveal how it thinks and where it breaks down.
For each one, the answer matters less than whether the agency can give a specific, honest one.
"What do you think is the real constraint on our growth right now?"
A strong agency asks clarifying questions before answering, then names something precise — conversion rate, offer-market fit, traffic quality, attribution gaps, sales follow-up. A weak one skips straight to the services it sells: "You need better SEO" or "Your social presence is underdeveloped." The question tests listening and diagnostic rigour simultaneously.
"Which channel would you prioritise first, and why would you de-prioritise the others?"
The second half is what matters. Any agency can recommend a channel. Fewer can explain what they would not do and why. A strong answer names trade-offs: this channel takes six months to compound, that one requires a creative volume you probably cannot sustain, the third makes sense only once conversion rates improve. A weak answer recommends everything.
"Walk me through your first 30, 60, and 90 days. What happens before you spend a dollar of our budget?"
Good agencies run discovery before they run campaigns. The first 30 days should be heavy on audit, access, baseline measurement, and brief refinement. If the answer jumps to deliverables and ad launches without a diagnostic phase, the plan is a template, not a response to your situation.
"How will you define success, and what metrics will you not use?"
The second clause is the diagnostic part. A good agency names business outcomes, qualified pipeline, conversion rate, CAC, revenue contribution; and can explain why impressions and follower counts will not appear in the headline report. An agency that defends vanity metrics when pressed is usually optimising for the appearance of activity rather than the outcome you care about.
"What did a recent campaign that underperformed look like, and what did you change?"
This is frequently more revealing than a success case. Agencies that can discuss failure with specificity — here is what we believed, here is what the data showed, here is what we changed — demonstrate the kind of analytical honesty that sustains long engagements. Agencies that only have wins to show are either very selective with their portfolio or not being straight with you.
"Who will be in the room for our weekly meetings, three months after signing?"
This is the most important question on this list, and it belongs in its own section.
The pitch team versus the delivery team
The single most common cause of agency disappointment is buying the senior team that pitches you and receiving a junior team that executes. It happens because agency business development is built around the most impressive people, and those people have limited capacity for ongoing account work.
This is the switch that causes most agency regret — and the one question that catches it before you sign.
who will be in the room for our weekly meetings, three months after signing?
This is not always bad, senior strategy plus competent junior execution can work well; but it must be disclosed and it must be priced accordingly. The problem is when it is hidden.
Ask these questions directly:
- "Who will own our account day to day?"
- "Who owns strategy, and how many hours per week will they spend on us?"
- "Which of the people in this meeting will attend our monthly review?"
- "What happens to our account if our lead strategist leaves?"
A strong agency names roles and people, clarifies how much senior time the account receives, and describes an escalation path if something goes wrong. A weak one answers with "our team" or pivots to describing the agency's collective expertise rather than the individuals on your account.
Protect yourself in the contract. Request a schedule naming key roles. Include a clause requiring written notification if the account lead or senior strategist changes, and the right to re-negotiate if the replacement does not meet the seniority band agreed. These are reasonable terms. An agency that refuses them is telling you something important about how they manage accounts.
TenetHow to check references properly
Standard reference checks are almost useless because agencies provide the happiest clients. To get real information, you need to ask for a different kind of reference.
Ask for a client who left.
A churned client can tell you what happens when the agency underdelivers, how honest they are under pressure, whether the senior team stays involved when results are disappointing, and whether the handoff process is clean. These are the conditions that matter most, and you will not learn about them from a client who is currently paying and reasonably satisfied.
If an agency refuses to provide a single former client contact, treat that refusal as its own answer.
When you speak to any reference — current or former — avoid the soft questions ("were they easy to work with?") and use these instead:
- "Did the team you were pitched match the team that ran the account?"
- "How did they respond when results were behind target?"
- "Were there any surprises in scope, billing, or staffing?"
- "If you left, what was the reason, and how did they handle the transition?"
You are listening for consistency, specificity, and candour. References who struggle to give concrete examples, or who only praise without nuance, are often less informative than references who name one genuine problem and explain how it was resolved.
Marketing agency red flags
Some warning signs are serious enough to end the process before a proposal is reviewed.
Five things in a proposal that should end the conversation before you review it — hover each to see why.
any one of these deserves a direct follow-up question. several together are grounds to end the conversation.
Guaranteed rankings or guaranteed leads. No agency controls search algorithms, auction dynamics, or buyer behaviour. Guarantees of specific outcomes are either dishonest or structured in a way that makes them meaningless in practice.
A proposal with no discovery. If an agency quotes a monthly fee and sends a scope of work after a 30-minute call, the proposal is generic. A genuine response to your situation requires understanding your customers, your current data, your internal capacity, and your constraints. If they skip that, the plan is a template.
Vague deliverables. "Monthly marketing support," "content strategy," and "ongoing optimisation" are not deliverables. A strong proposal names outputs, cadence, channels, and how each item connects to the agreed outcome. Vague deliverables make comparison impossible and protect the agency, not the client.
Ownership of your accounts and data. You must own your ad accounts, analytics properties, CRM data, domain, and creative assets. If an agency insists on holding these or makes access conditional, walk away. Losing access to your own historical data is a serious operational risk, and the threat of it is sometimes used as a tool to prevent churn.
Overconfidence about channels they happen to sell. An agency that recommends the same channel mix to every client regardless of context is selling inventory, not strategy.
Before you sign
The contract is the last point at which you can protect yourself before performance becomes the only measure. Negotiate these terms before signing.
Four terms in the contract do most of the protecting — here's what each one locks down.
the contract is the last point at which you can protect yourself before performance becomes the only measure.
Notice period. A 30-to-60-day rolling notice after an initial committed term is reasonable. Contracts requiring six months' notice to exit should prompt careful scrutiny of what the agency fears losing.
Account and data ownership. Make this explicit in writing. You retain full access to every account, file, and dataset regardless of how the relationship ends.
Named staff. List the account lead, strategy lead, and channel specialists by name or role. Include a clause requiring your approval before a material change to the delivery team.
First-90-days outcome. Define what the initial phase produces: a completed audit, baseline metrics, a prioritised roadmap, and the measurement framework for everything that follows. This is useful for both parties. If the agency resists putting it in writing, that hesitation matters.
For small businesses in particular, the contract term and exit flexibility deserve close attention — a 12-month lock-in with no performance review clause is a significant exposure when budgets are tight and conditions change quickly.
The shortcut through this whole process
Every question in this article exists because an agency is hard to see inside. Tenet Operator removes the problem rather than helping you navigate it: one named person accountable for shipping, the work visible in your own account as it happens, and no rotating delivery team to lose halfway through. If the diligence is starting to feel disproportionate, it is worth a look before you shortlist.
See how Tenet Operator works →
TenetFAQ
What questions should I ask a marketing agency before hiring?
Prioritise questions with diagnostic power over generic ones. Ask what they think the real constraint on your growth is, which channel they would de-prioritise and why, who will work on the account day to day, what a recent underperforming campaign looked like, and how they will define success without vanity metrics. The quality and specificity of the answers separates strong agencies from polished pitches.
What are red flags when choosing a marketing agency?
Guaranteed results, proposals built without real discovery, vague deliverables, agency ownership of your accounts and data, and an inability or refusal to name the delivery team. Any one of these deserves a direct follow-up question. Several together are grounds to end the conversation.
How long should a marketing agency contract be?
Long enough to produce meaningful results in your chosen channel, typically three to six months minimum; with a rolling notice period after that. Longer terms can be justified for complex work, but only when paired with defined milestones, a performance review mechanism, and a fair exit clause.
Should I choose a specialist or full-service agency?
If you have a clear, bounded problem in one channel, a specialist usually outperforms a generalist. If you need coordinated execution across multiple channels and lack internal marketing leadership, a strong full-service agency is often more efficient than managing several specialists separately. The answer follows directly from what you decided you are buying.
How do I know if a marketing agency is doing a good job?
Measure progress against the business outcomes agreed before the engagement started — pipeline, revenue, conversion rate, CAC, or retention depending on what was scoped. Healthy relationships also show honest reporting when results disappoint, clear next steps after each review, and a delivery team that can explain what is working, what is not, and what they are changing.
