Choosing a digital marketing agency is one of the most consequential decisions a business owner or marketing manager makes. The right questions are your main tool for separating agencies that deliver measurable results from ones that sell a polished pitch. This guide covers what to probe, from verified case studies and reporting to AI use and contract terms, so you can evaluate any agency with confidence.

Which results and case studies to ask for

Past performance is the clearest signal of future delivery. Ask every agency for case studies with specific, measurable outcomes, not a wall of client logos. A logo wall tells you nothing about what the agency actually achieved.

Request results from clients in a similar industry or at a comparable business size. A campaign that drove 312% more qualified leads for a dealership means far more than a vague claim about growing brand awareness. Numbers with context reveal whether the agency understands your customer and buying cycle.

Watch for these red flags:

• Reluctance to share real numbers. Agencies with strong results share them freely.
• Results without attribution. Ask exactly which channel, tactic or campaign produced the outcome.
• Only long-term clients cited. Ask for a recent example from the past 12 months.
• No mention of what did not work. Honest agencies discuss failures and what they learned.

Pro tip: ask the agency to walk you through one campaign from brief to result. The depth of the answer reveals how well they understand their own work.

How they measure success and report

Reporting quality is a direct measure of accountability. Ask for a sample report before you sign anything. An agency that cannot show you what a real client dashboard looks like is not ready to be accountable to you.

Effective reports blend two types of metric. Leading indicators include creative test results, SEO ranking movements and click-through rates. Lagging indicators include pipeline value, cost per acquisition and revenue. A good cadence is weekly or biweekly, not quarterly. Quarterly reporting means three months heading the wrong way before anyone notices.

Ask these directly:

• Who is my named contact for explaining the report data?
• Do I keep direct access to my Google Ads, Meta and analytics accounts?
• How quickly will you flag a campaign that is underperforming?
• What does reporting look like in the first 30 days?

Pro tip: ask whether the agency uses incrementality testing to validate true campaign impact. Google brought the cost of this method down from about $100,000 to $5,000, putting it within reach of most advertisers. It confirms whether your spend is driving new customers rather than claiming credit for sales that would have happened anyway.

Operational and team questions

The person who sells you the account is rarely the person who runs it. Transparency about who does the actual work is one of the most overlooked areas in agency evaluation. Ask directly: who manages my account day to day, and what is their experience level?

Ask these in sequence:

1. Who is my day-to-day account manager? Get a name.
2. What is their experience in my industry? Junior staff running senior budgets is a common problem.
3. What happens if my account manager leaves? No clear answer means real continuity risk.
4. How involved are the founders or senior strategists? If the agency cannot deliver without its founders in the room, your account is exposed.
5. Who owns the ad accounts and creative assets? You should own everything. Any agency that holds account ownership as leverage is not acting in your interest.

Pro tip: ask to meet the team member who will handle your account before you sign. If the agency redirects you to a salesperson, treat that as a warning sign.

How to evaluate their use of AI

AI is part of every credible agency's workflow now, but the gap between genuine integration and surface-level name-dropping is wide. Ask which specific AI tools the agency uses for each deliverable, whether SEO, content, paid media analytics or creative testing. "We use AI" is not an answer.

The right follow-ups reveal real capability:

• Which AI models do you use for content, and how do you review outputs before publishing?
• How do you govern AI outputs to catch errors or factual mistakes?
• Does AI change your pricing or turnaround times?
• How do you use AI in analytics or audience segmentation?

Saying "we use ChatGPT" without describing oversight, governance or workflow is a red flag. Structured AI use means defined processes for checking outputs, not just prompting a tool and publishing the result.

Agencies that use AI well show real productivity gains in delivery. Agencies that use it poorly produce generic content and miss errors. The difference shows up in the quality of their case studies and how clearly they explain their process.

Contract, budget and collaboration questions

Contract clarity prevents the most common agency disputes. Fees and structure should be fully transparent before you commit. Ask for a line-item breakdown of what is included and what costs extra.

Cover these:

• Monthly budget expectations. Ask what a realistic minimum looks like to hit your goals. Agencies that dodge this are not being honest.
• Timeline to results. SEO takes longer than paid search. Ask for realistic milestones, not promises.
• Your responsibilities. Content approvals, brand assets and timely feedback all affect delivery.
• How they stay current. Ask how they track Google Ads policy changes, Meta updates and shifts in search.
• References from current clients. Ask to speak with a client of 12 months or more. A client who stayed tells you more than a glowing testimonial.

The ANA and 4A's pitching principles confirm that strong partnerships start with detailed briefs and early agreement on KPIs and leadership availability. The questions you ask before signing set the tone for the whole relationship.

What I have learned from evaluating agencies

The most revealing moment in any agency evaluation is not the pitch. It is the silence after a hard question.

I have sat across from agencies that could not name the person who would manage the account. I have seen dashboards that looked impressive but carried no lagging indicators, so there was no way to tie ad spend to revenue. And I have watched owners sign without asking who owns the ad accounts, only to find later that the agency kept access as leverage.

These questions are not designed to trip agencies up. They find the ones that welcome scrutiny. A capable agency answers every one without hesitation: real reports, a named account manager, and AI workflows explained with specifics. One area most owners skip entirely is incrementality testing. Asking whether an agency uses controlled experiments to measure true impact separates the ones who understand modern measurement from the ones still leaning on last-click attribution.

Frequently asked questions