Give each media buyer the same account decision to work through. A useful answer connects the evidence to an action, explains the next check and states what would change the decision. That gives you something more concrete to compare than confidence on a sales call.

The example below uses a losing ad inside a campaign with profitable siblings. It lets you assess how a candidate handles margin, incomplete evidence and the next creative test.

Build one scenario and lock your criteria before anyone speaks

Use the same scenario for every candidate so differences in the questions do not drive your assessment.

Set the criteria before the interviews. Look for:

1. Evidence. What in the account supports the call, specifically.
2. The next check. What they would look at next, and when.
3. The reversal condition. What would have to happen for them to admit the call was wrong.

Here is a scenario you can steal. Swap in your own numbers.

An ad has spent $180 over three days with zero purchases. Your breakeven cost per purchase is $45. The ad sits inside a CBO campaign where two other ads are converting profitably. What do you do?

Work out the acquisition allowance before comparing performance decisions. The candidate needs that financial boundary to explain how much further spend is justified.

An answer that needs a follow-up

Suppose the candidate answers:

"I would pause it. The performance is weak."

Pausing may be reasonable, but this answer does not explain the decision. Ask about tracking, conversion lag, the loss limit and the evidence from the rest of the campaign.

The follow-up is not "are you sure?", because they will say yes. Ask them to show the missing reasoning: "Walk me through what you would check before acting, and what you would do after."

The good answer commits to your numbers

A more complete answer might sound like this:

"I would check purchase tracking and conversion lag, then compare the $180 spend with the agreed loss limit. It is four times the $45 breakeven CPA, with no recorded purchases, so pausing is a reasonable provisional decision. I would record why, check whether the same problem affects other ads and decide what to test next. Late purchases or a tracking fault would change how I interpreted the creative. The two profitable siblings are useful context, but their spend does not prove an ad with little delivery would fail."

This answer uses the actual costs, separates the action from the diagnosis and states what could change the conclusion.

Ask how they handle an ad with almost no spend

An ad that spent $180 with no purchases is different from one that barely delivered. Ask the candidate how they would treat the second case.

A useful answer starts with eligibility and the amount of evidence. Replacing an underfunded ad can be a practical choice; a separate test can also be justified when the concept matters. The candidate should explain what that test would answer, its budget and why it is worth running. Minimum spend at ad set level does not guarantee equal delivery among its ads.

Four more red flags

• They guarantee results. Nobody controls the auction. A guaranteed ROAS is a sales device, and the buyers with real track records are the ones who refuse to promise one.
• They dismiss every previous provider without showing the relevant account evidence. Ask for the specific problem, the decision they made and its outcome.
• They ask for admin access when advertiser access covers the job. More on this below.
• Their sample report reads the dashboard back to you. A report should tell you what they decided, why, and what they need from you next. You can read the numbers yourself for free. You are paying for the decisions.

What access should you actually grant?

Meta ad accounts have three permission levels. Admin controls the account itself, including permissions and billing. Advertiser can create and edit ads, which is the actual day-to-day job. Analyst is read-only.

Give a new buyer advertiser access. It covers everything you are hiring them to do: launching, editing, adjusting budgets, pausing, scaling. A candidate who insists on admin before doing any work should name the specific task that requires it. There are legitimate ones, but "it is easier" is not on the list. And if you want a low-risk start, grant analyst access first and pay for a diagnosis of the account before anyone touches anything.

Ask how you will see the changes after they start. Activity history records account changes; a report should add the reasoning and follow-up. Agree on both before the engagement begins.

The copy-paste interview script

"Our breakeven cost per purchase is $45. One ad has spent $180 over three days with zero purchases. It sits in a CBO campaign where two other ads are converting profitably. Walk me through what you would do."

Then, whatever they answer:

"What would you check before acting?"

"One ad in that campaign is getting almost no spend. What do you do with it?"

"What should this account look like in two weeks if your call was right?"

"What would make you reverse it?"

"After you start, how will I see what changed?"

Compare whether candidates check measurement, use your economics, distinguish insufficient delivery from a measured loss and explain the next decision. A guaranteed return or an answer that ignores your loss limit deserves further scrutiny.

Use the interview to compare candidates

Use the answers alongside references, relevant work and the proposed scope. For the commercial comparison, see management pricing and agency versus freelancer. If the responsibilities are still unclear, start with what a media buyer does.