How to Choose a Marketing Agency Without Getting Burned Again

40%. That is the share of businesses that industry research reported by Ad Age found were planning to switch marketing agencies. Nearly half cited delivery issues as the primary reason they ended their last agency relationship. Gartner's CMO research on agency partnerships reaches the same conclusion from a different angle. Agency churn is consistently driven by a delivery gap, not a pricing gap.

Most of those relationships started with a pitch that looked credible. A polished deck. Case studies. A senior person in the room. The problem is not that agencies are universally dishonest. It is that the standard process for choosing a marketing agency rewards people who are good at pitching, not people who are good at delivering. What follows is the evaluation framework to run before signing anything, the questions polished agencies cannot prepare for, and how Leapyn answers its own criteria.

What to Look For When Choosing a Marketing Agency

Six criteria separate agencies worth hiring from agencies worth avoiding.

1. Own Your Assets (Non-Negotiable). Admin access means full control, not view-only. Before any contract is signed, the agency should transfer admin-level access to your Google Ads account, Meta Business Manager, GA4 property, website CMS backend, and all creative assets to your team. An agency that owns your ad accounts is an agency that holds your business hostage at the end of the relationship. Any resistance to this before work begins is itself the evaluation result.
2. Insist on Transparent Reporting. The metrics worth tracking are cost per qualified lead, cost per sales conversation, and pipeline-connected outcomes. Impressions, reach, and follower counts are not reporting. They are a distraction from whether the marketing is producing revenue. A transparent agency provides direct access to the underlying platform data, not a proprietary dashboard that sits between you and the numbers. Ask what the reporting will contain before you sign, not after you are three months in.
3. Vetting the Team. The person who runs the pitch is rarely the person who runs the account. Ask for the full names and LinkedIn profiles of the specific people who will manage your campaigns, the account-to-team-member ratio, and whether the strategy lead in this conversation is the same person as the day-to-day account manager. Small-roster agencies have a structural advantage here because there are fewer people to shuffle between accounts. A large agency with a junior team structure cannot answer these questions honestly.
4. Demand Data, Not Just Creativity. An agency that understands how to produce commercial results will ask about your customer acquisition cost targets, gross margins, average deal value, and close rate before recommending a single channel or tactic. An agency that does not understand this will ask about brand, aesthetic preferences, and which competitors you admire. One useful diagnostic is to ask what data they would need from you to recommend a strategy. The answer tells you immediately which kind of agency you are talking to.
5. Verify with Real References. Website testimonials are selected by the agency. Real references are current clients you choose from a published client list, not the two contacts in the case studies deck. Ask references what the agency got wrong in the first 90 days and how they handled it, whether they would hire the agency again knowing what they know now, and whether the day-to-day contact is still the same person from the initial call. References who reveal how the agency handles problems are more useful than references who confirm the agency is good.
6. Understand Their Failed Campaign Process. Every agency has had campaigns that underperformed. The ones worth hiring know exactly which campaign it was, which metric revealed the problem, what was changed, and what happened as a result. If the answer pivots to success stories, blames the platform, or claims nothing has ever failed, end the meeting. The ability to describe a specific failure with specific detail is the clearest indicator of agency maturity available in a one-hour conversation.

Red Flags That Reveal a Marketing Agency Is Not Worth Hiring

Four patterns consistently appear in agency relationships that disappoint.

1. Guaranteeing Quick Results. No marketing channel produces guaranteed outcomes on a fixed timeline. An agency promising number-one rankings in 30 days or a specific lead volume in the first month is either describing tactics that produce short-term numbers at long-term cost, or setting an expectation they know they will miss in order to close the contract. When you see a guarantee, ask which tactic produces this result and what happens contractually if the guarantee is not met.
2. Charging Percentage of Ad Spend. This fee structure creates a situation where the agency earns more when your ad budget is higher, regardless of whether the higher budget is producing better results. An agency paid as a percentage of spend has a structural incentive to resist reducing your budget even when the data says to reduce it. Ask for a fixed management fee instead, or a performance-indexed fee tied to qualified leads or pipeline outcomes rather than to how much you are spending.
3. Opaque Communication and Reporting. If you have to send three emails to find out what is happening with your campaigns, that is an information asymmetry problem. If the agency explains performance using proprietary terminology that makes comparison to platform data impossible, that is a transparency problem. Both are diagnosable before signing. Ask them to describe the exact format and frequency of reporting before you commit. Vague answers to this question are informative.
4. Too Polished or Too Cheap. A pitch deck that could belong to any client, case studies without named contacts and specific numbers, and testimonials from companies you cannot independently verify are signs of recycled materials without genuine client evidence. A price point so low that it cannot support senior talent spending meaningful time on your account means junior staff are doing the work regardless of what the proposal says. The due diligence question is the same in both cases. Who exactly is doing the work?

How to Start Right With a New Marketing Agency

Starting with a defined 90-day pilot rather than a full long-term commitment is the structural protection the standard selection process consistently fails to provide. The pilot works only if three conditions are defined before day one.

A defined deliverables list before day one. The pilot should begin with a written document specifying exactly what will be produced in 90 days. That includes which campaigns, which content pieces, and which reports at which frequency. Any scope change during the pilot should require written agreement from both parties. An agency that resists defining deliverables upfront has already decided it will not be held to them.

A measurable checkpoint at day 30. Not a final result, but a signal review. Are the initial campaigns structured correctly? Is the tracking infrastructure in place? Are communications happening at the promised cadence? If the checkpoint reveals a significant problem, name it directly, get a written remediation plan with a 14-day correction window, and use what happens next as the actual evaluation. How an agency responds to a problem at day 30 is more informative than how it responded to a question in the pitch.

A clear off-ramp condition. Before the pilot starts, define what would cause you to end it at 90 days and what the operational exit looks like. Asset access, campaign pausing, and data handover procedures should all be documented before work begins, not negotiated under pressure at the end of a relationship that has either succeeded or failed. An agency confident in its work has no reason to resist documenting this.

The Questions to Ask a Marketing Agency That a Polished Pitch Cannot Prepare For

The questions in any agency's prepared presentation are the ones they rehearsed. These are the ones they did not.

"Who specifically will work on my account day-to-day after we sign?" A credible answer includes full names, LinkedIn profiles, and the account-to-team-member ratio. A deflective answer is "our team" or "we staff appropriately to client needs."

"Can I have admin access to all my ad accounts and analytics from day one?" A credible answer is yes, with a specific process for how it is set up before work begins. A red flag is "we manage those centrally" or "we will set that up once we are underway."

"What is your fee structure and does any part of it change if my ad spend goes up or down?" This is the diagnostic for percentage-of-spend pricing. A credible answer is a fixed fee clearly unrelated to budget size. A red flag is "we charge a management percentage on top of media spend."

"Tell me about a campaign that significantly underperformed and what you did about it." A credible answer names a specific campaign, a specific metric, a specific intervention, and an honest outcome. A deflective answer pivots to successes or frames everything as a learning without specifics.

"What data would you need from me before recommending a strategy?" A credible answer asks about customer acquisition cost targets, gross margins, average deal value, current close rate, and sales cycle length. A deflective answer asks about brand and aesthetic preferences.

"Can I speak to three current clients I choose from your published client list, rather than three you select for me?" Most credible agencies will agree. Resistance to this request is informative.

"What happens to my campaigns, assets, and data if we stop working together?" A credible answer is full handover, all access retained by the client, and a documented exit procedure. A deflective answer involves vague language about transition support or extended notice requirements before accessing assets the client owns.

"What would make this engagement unsuccessful, and what do you do when that happens?" A credible answer defines failure specifically, names an escalation process, and describes the conditions under which the agency would recommend stopping. A deflective answer is "we do not let engagements fail" or an immediate pivot to confidence statements.

The Marketing Rules That Reveal Whether an Agency Knows Its Business

Four framework rules appear frequently in marketing research and buyer education. Each is useful as an agency evaluation diagnostic because a knowledgeable agency should be able to apply them to a specific client situation, not just recite them.

The 7-11-4 Rule. A buyer typically needs about 7 hours of content engagement, 11 touchpoints, and encounters across 4 different platforms or locations before purchase intent becomes strong enough to act on. This framework is grounded in Google's research on buyer trust and purchase behaviour. Ask a prospective agency whether they have applied this to a client's content and channel strategy. An agency that plans around individual campaign conversions without accounting for the full trust-building journey will consistently underestimate the time and touchpoints required to convert a cold audience.

The 3-3-3 Rule. In a sales and marketing context, the 3-3-3 principle suggests reviewing three pieces of relevant context about a prospect, spending three minutes on preparation, and making contact within three business days of a buying signal. Applied as an agency evaluation question, it reveals whether the agency integrates sales signals with marketing activity or treats marketing and sales as entirely separate programs. Ask how they handle the moment when a content-engaged prospect shows purchase intent.

The 3-7-27 Rule of Branding. The rule holds that a brand makes an initial impression in 3 seconds, is remembered after roughly 7 interactions, and builds genuine trust after around 27 interactions. The specific numbers are a heuristic rather than a rigorously sourced empirical finding, but the underlying principle is well-supported by research on repeated exposure and brand recognition. Ask a prospective agency how they plan for the trust-building timeline across your specific brand. An agency that plans only for the first impression is not planning for the relationship that produces a commercial outcome.

The 1% Rule. On most platforms, roughly 1% of users create content, 9% engage with it, and 90% consume it without visible interaction. Applied as a content strategy diagnostic, this rule reveals whether the agency understands the difference between activating internal experts and producing brand content on company accounts. Ask how they think about developing thought leadership from people inside the business. An agency that defaults to managing your company social accounts rather than activating your internal experts is running a broadcast publishing model.

How Leapyn Answers Its Own Evaluation Framework

Applying the same framework to ourselves is the only honest way to close a post like this. If the criteria above are worth asking another agency, they are worth asking Leapyn.