Every founder eventually asks the same question: agency, freelancer, or in-house hire? Whichever way you go, you're eventually vetting a media buyer — and most interview processes for this role are terrible, because most founders don't yet know what "good" looks like. Here's the framework I'd want a client to use on me.
What to ask in the interview
- "Walk me through what you check before touching a bid on a new account." The right answer starts with tracking and measurement, not strategy. If they jump straight to "I'd test some creative," that's a flag.
- "Tell me about a campaign you killed, not one you scaled." Everyone has a scaling story. Fewer can talk honestly about cutting a loser before it bled the budget. That instinct matters more than any single win.
- "How do you know if a conversion number is real?" They should mention reconciling platform numbers against your actual backend/CRM data, not just trusting the ads dashboard.
- "What's your relationship with Smart Bidding / automation?" You want someone who feeds the algorithm good data and sets guardrails — not someone fighting the platform with manual bids, and not someone who's fully hands-off and just "lets the AI do it."
- "What would you need from me in week one?" Good answer: tracking access, historical data, clarity on unit economics. Bad answer: just budget and login.
Portfolio and proof red flags
- Only screenshots of ROAS, never screenshots of spend or volume. A 10x ROAS on $200 total spend proves nothing.
- No mention of tracking setup anywhere. If server-side tracking, Enhanced Conversions, or GA4 configuration never come up unprompted, measurement probably isn't a priority for them — and it should be the first priority.
- Case studies with no timeframe. "Grew revenue 340%" over what period, from what base? Vague timeframes usually mean an inconvenient one.
- Can't explain why a specific account underperformed. Ask about a failure. Someone who's only ever run good accounts either hasn't run many, or isn't being honest.
The strongest signal isn't the win they show you — it's how precisely they can explain the loss they don't want to show you.
Structuring a trial period
- 2-4 weeks, one channel, clear success metric agreed upfront — not "grow the business," but a specific CPA or ROAS target tied to your real unit economics.
- Give them tracking access on day one. If they don't ask for it, that's diagnostic on its own.
- Ask for a mid-trial audit summary, not just a final report. You want to see how they think, not just the number at the end.
- Budget the trial to actually produce signal — a $200 test budget can't prove or disprove anything. See the budget framework for the minimum that generates real data.
Agency vs. freelancer vs. in-house
- Agency: broader bench, less individual attention, good for multi-channel scale once budgets are large.
- Freelancer/consultant: direct access to the person actually in the account — no account-manager telephone game. Best fit for founders who want to understand the "why," not just receive a monthly PDF.
- In-house: makes sense once spend and complexity justify a full-time seat — usually somewhere past $15-20K/month in ad spend with multiple active channels.
Whichever path, the checklist doesn't change: tracking-first thinking, honest talk about failures, and a trial structured to actually produce evidence. The media buyer who passes that bar is worth paying for. The one who can't is expensive no matter the price.