How to Choose a DTC Growth Agency: What to Ask, What to Ignore, and the Red Flags That Cost Brands Millions

There's an uncomfortable truth about agency evaluations: the brands that get burned most badly are usually the ones who did the most due diligence. They got the case studies, checked the references, sat through three rounds of strategy presentations, and still ended up in a 12-month retainer with an agency that couldn't move the needle on their actual business.

The failure mode isn't lack of diligence — it's diligence aimed at the wrong things. Brands evaluate decks and pitches when they should be evaluating systems and people. They ask about results when they should be asking about process. They get impressed by client logos when they should be interrogating team structure.

Having been on both sides of this — as an agency founder who has pitched hundreds of brands and as someone who has watched what actually drives outcomes — here's the framework I'd use if I were hiring an agency tomorrow.

The Scope Match Problem

The single most common reason agency relationships fail has nothing to do with the agency's capabilities. It's a scope mismatch: the brand hired an agency that was genuinely excellent at solving a different size problem.

An agency built for brands doing $2M–$10M in revenue operates completely differently from one built for brands at $30M–$100M+. Their team structure is different. Their benchmarks are different. Their risk tolerance is different. Their creative testing cadence is different. The problems they're optimized to solve don't overlap as much as the pitch would suggest.

The brands in the $2M range need agencies who can build the machine from scratch — establish channel structure, build initial creative learnings, find the first scalable audiences. The brands at $50M need agencies who can optimize a machine that's already running — tighten efficiency at scale, manage creative fatigue systematically, find growth at higher spend levels without CAC blowout. These are fundamentally different skill sets, and agencies that are genuinely good at one are rarely excellent at the other.

What the Pitch Process Actually Tells You

Every agency pitch is designed to maximize your confidence in hiring them. That's not cynicism — it's just how pitches work. The information presented is curated, the team showing up is the best-case team, and the strategy presented is what sounds most compelling in a conference room setting, not necessarily what would actually happen.

This means you have to deliberately extract information the pitch isn't designed to give you. Here's what to look for:

How They Diagnose, Not What They Prescribe

Ask the agency to walk you through how they'd approach the first 30 days of your account. Listen for questions, not answers. An agency with real diagnostic chops will ask about your unit economics, your current creative library, your attribution setup, your seasonality, your offer architecture, and your retention data before making any prescriptions. An agency pitching a template will jump straight to "we'd restructure your campaigns like this and test these audience segments."

The prescription is irrelevant before the diagnosis. Confidence in the prescription without the diagnosis is a red flag — it means they're pitching what worked last time, not what's right for your specific situation.

Who Actually Works Your Account

One of the most consistent complaints in agency relationships is the bait-and-switch: senior strategist closes the deal, junior coordinator manages the account. This isn't always malicious — agencies often structure senior leadership in a client acquisition role — but it does mean you need to meet the actual team before signing.

Ask directly: who is the day-to-day account manager for accounts at my spend level, what's their experience, can I meet them before we sign? If the answer involves any amount of hedging — "we're still finalizing team assignments" or "our senior team oversees all accounts" without specifying who runs day-to-day — treat that as a yellow flag.

The Questions That Actually Matter

After 10+ years of observing what separates productive agency relationships from expensive mistakes, these are the questions I'd ask in every evaluation:

1. What does your creative testing process look like at my spend level?

This question reveals more than any case study. A disciplined answer will describe: how many creative variants they test per week, what they define as a statistically meaningful result versus noise, how they structure tests to isolate variables, and what their pipeline looks like from brief to launch to decision. A vague answer — "we test continuously and optimize based on what's working" — means they don't have a system, they have an intention.

2. What happens when creative stops working?

Creative fatigue is inevitable at scale. The question is whether the agency has a process for anticipating it and replacing winners before performance craters, or whether they react after CAC starts climbing. Push them on the specifics: at what frequency does fatigue typically hit for brands at my spend level, how do you signal it to clients, what's the brief-to-launch timeline for replacement creative, who owns that process on both sides?

3. How do you think about incrementality?

This is the sophistication test. An agency optimizing for platform-reported ROAS is often optimizing for attribution credit, not actual incremental revenue. Ask how they distinguish between ads that drove revenue and ads that got credit for revenue that was going to happen anyway. If they can't articulate a view on incrementality — even a simple holdout test approach — they're flying on dashboard metrics, which at scale will consistently mislead budget decisions.

4. Can you show me a case study from a brand at my exact spend level with specific numbers?

Not similar spend level. Exact. The results from a brand at $30K/month tell you almost nothing about performance at $200K/month because the problems, the audience saturation dynamics, and the creative velocity requirements are categorically different. Press for specifics: baseline CAC when they started, CAC at peak scale, spend level when results were achieved, timeline, and what actually drove the improvement. If they can't or won't provide specifics, the case study is decoration, not evidence.

5. How do you handle underperformance?

Every agency presents the upside scenario. Force them to describe the downside one. What do they do if CAC trends up three weeks in a row? Who escalates, what's the escalation protocol, what decisions get made at what threshold, and who makes the call to pivot strategy versus execute harder on the current plan? The answer tells you whether they have accountability structures built in or whether they manage-by-optimism until a client threatens to leave.

The Red Flags Worth Walking Away From

There are patterns that consistently predict poor outcomes. Most brands see them during the evaluation and rationalize past them. I'd suggest not doing that.

Vague Case Studies

"We 3x'd their ROAS." This tells you nothing actionable. What was the baseline ROAS? What spend level? Over what time period? Was it Q4 seasonality or a sustainable structural improvement? What happened to CAC? Did the brand continue with the agency after the result?

Case studies that lack specifics aren't hiding excellence — they're hiding the context that would let you accurately evaluate the result. Ask for the actual numbers, the actual timeline, and whether you can speak to the client. If either of the last two items is declined, discount the case study entirely.

Percentage-of-Spend Pricing

The incentive structure of a percentage-of-spend model is straightforward: the more you spend, the more the agency earns. This creates an inherent conflict with your interest, which is to spend efficiently — to grow spend only when the marginal dollar is producing acceptable returns. An agency on a percentage model profits when you scale, regardless of what scaling does to your unit economics.

This doesn't mean percentage-of-spend agencies are bad — it means the incentive structure creates pressure you should be aware of. The cleanest alignment comes from flat-retainer models with performance benchmarks both sides agree to upfront.

30-Day Guarantees on Cold Accounts

Any agency promising meaningful results within 30 days on a cold account with fresh creative is overpromising. Cold accounts take 4–8 weeks minimum to generate the signal volume needed to make real optimization decisions. Creative testing takes longer. Trust the agency that gives you an honest learning-period timeline over the one that promises results before the data exists to support them.

The Evaluation Scorecard

After the formal evaluation is complete, use this scorecard before making a final decision. None of these dimensions alone should be disqualifying — but patterns of weakness across multiple dimensions should be.

• Scope match: Is the median client at my spend level? Do their benchmarks feel calibrated to my size?
• Diagnosis quality: Did they ask more than they prescribed in the discovery process?
• Team clarity: Do I know exactly who runs my account day-to-day, and have I talked to them?
• Creative system: Can they describe their testing and refresh process with enough specificity that I'd know if it was happening?
• Case study depth: Do they have at least one case study at my spend level with verifiable specific numbers?
• Incentive alignment: Is the fee structure aligned with efficiency, not just scale?
• Exit flexibility: Can I exit within 60 days without penalty if the relationship isn't working?

An agency that scores well across all seven isn't guaranteed to be the right fit — culture, communication style, and domain expertise in your specific category still matter. But an agency that struggles on more than two of these dimensions is showing you structural problems that will surface in the relationship regardless of how good the pitch was.

What Good Actually Looks Like

The best agency relationships share a set of structural characteristics that are worth knowing going in, because they help you recognize the right fit when you see it.

First, the agency treats your unit economics as the north star — not platform ROAS, not CTR benchmarks, not impressions. Contribution margin per order, CAC payback period, LTV trajectory. Every strategy recommendation is grounded in how it affects those numbers, not how it improves dashboard metrics.

Second, there is a clear creative feedback loop. The agency has a system for briefing new creative, testing at volume, identifying winners, and systematically killing losers before they drag performance. This loop runs continuously, not reactively. You should be able to see it in operation, not just hear it described in the pitch.

Third, the communication is direct when things aren't working. The best agency relationships involve an account team that tells you when a strategy isn't performing before you notice it on your own — not one that waits for you to flag declining numbers and then explains why the platform algorithm changed. Agencies that own their outcomes are worth significantly more than agencies that attribute everything to external factors.

Finally, the relationship evolves. Twelve months in, the strategy should be materially different from month one — not because it failed, but because you've generated enough signal to make better decisions. An agency still running the same structure and creative approach after a year isn't learning from the account. The best agencies are always building on what they've found, compounding creative intelligence, and continuously stress-testing what they believe is working.

The agency that can show you that kind of systematic learning — not in a deck, but in the actual account history — is the one worth paying a premium for.

---