Paid acquisition · Hiring September 2026 9 min

Eight questions to ask before you hire anyone to run your ads (me included)

Almost every one of these conversations opens with price, and price predicts almost nothing about how it will go. These eight questions predict quite a lot — and the first one can stop you losing years of advertising history.

I am on the receiving end of these conversations, so treat this as self-interested if you like. But I would rather compete on answers than on being the cheapest quote in an inbox, and every question below is one I would want asked of me.

1. Whose name is the Google Ads account in?

Start here. It is the least glamorous question on the list and the one that can cost you the most.

If your campaigns run inside the agency's own account, then everything that account has learned belongs to them: the conversion history, the optimisation signal, the years of accumulated data that make a mature account cheaper to run than a new one. The day you leave, you do not take it with you. You start from zero, with a new account that knows nothing, and your costs go back up to where they were before you ever hired anyone.

The account should be in your name, on your billing, and the person running it should have management access to it. That is the normal arrangement, not a special favour.

If someone hesitates here, you have learned what you needed to know before signing anything.

2. How do you charge — flat fee or percentage of spend?

Both models exist and both can be honest. But they point the incentives in different directions and you should know which direction yours points.

On a percentage, the person managing your money earns more when you spend more, regardless of whether the extra spend produced anything. That conflict does not make anyone dishonest — it just means that when the recommendation is "increase the budget", you cannot fully separate the advice from the interest.

A flat fee removes that. My own arrangement is flat and does not change with the budget, so if I tell you to spend more it is not because it pays me more. Whatever you choose, make sure you can state it in one sentence.

3. What counts as a conversion for you?

This is where a great many reports quietly become fiction.

If a conversion means somebody clicked a button, or reached a thank-you page, or opened WhatsApp, then a report full of conversions can sit alongside a month with no new customers. Those are events, not outcomes. They are worth tracking, but they are not the thing you are buying.

Ask what a conversion means in your account specifically, and whether it can be traced through to work you actually invoiced. If nobody can tell you what happened after the form, then nobody knows whether the advertising works — including the person sending you the report.

4. What are you not going to do?

The answer to this tells you more than any case study.

Anyone who says yes to everything — search, social, SEO, content, email, the lot — is either inexperienced or about to spread your budget so thinly that nothing gets real attention. Somebody who tells you plainly that they do not do organic social, or do not touch a channel you asked about, is showing you they know their edge and have enough work to turn things down.

I do not do organic social media management. I say so early because a client who wants that will be unhappy with me in month two, and neither of us needs that.

5. Can I look at the account whenever I want?

The answer should be an immediate yes, with access handed over without a conversation about it.

You are not going to log in every day, and that is not the point. The point is what it means if the answer is anything other than yes. Reporting that can only be seen through a monthly PDF is reporting that can be shaped before you see it.

6. How long before we see something, and what happens if we do not?

The first half of that question gets asked constantly. The second half almost never does, and it is the more revealing one.

Any campaign's opening numbers are its worst. What matters is whether the person running it has a plan for the case where it is still bad in eight weeks — and whether that plan is specific. "We will optimise" is not a plan. "We will rewrite the ads, add negatives from the search terms report, and if the cost per booked job is still above your ceiling we will tell you the channel does not work for you" is a plan.

Ask whether they have ever told a client to stop advertising. The answer tells you whether you are talking to an advisor or a vendor.

7. What budget makes sense in my particular case?

Notice the framing. Not "what is your minimum" — what makes sense for you.

A good answer works backwards from your numbers: what a customer is worth to you, how many you want, roughly what a click costs in your market, and therefore what budget could plausibly produce that. It will come with uncertainty attached, because it should.

An answer that is the same number for every client is a price list, not an analysis.

8. What do you need from me for this to work?

If the answer is "nothing, we handle everything", be careful. That is what people want to hear and it is almost never true.

Advertising that produces enquiries also produces the obligation to answer them, quickly, by someone competent. Most campaigns that underperform are not badly built — they are well built and feeding a business that takes two days to reply. Somebody has to tell you that up front.

A real answer names what you owe: fast replies, access to your numbers, and someone on your side who records whether the work actually happened.

What to do with the answers

You are not grading a test, and there is no combination of perfect answers. You are listening for two things.

First, whether the person is comfortable being specific. Vagueness under direct questioning does not improve after you sign.

Second, whether anything they said constrains them. Someone who has told you what they will not do, what the account ownership is, and what happens if it fails has handed you things to hold them to. Someone who agreed with everything has handed you nothing.

And if you want to run these at me, I would rather have that conversation than a quote comparison. It is a better use of both our time.

Start with the numbers, not the pitch

Before proposing anything I look at where your money is going and what it produces — your real cost per booked job, where the funnel leaks, what to fix first. Audit · $200, credited in full if we work together.

Related reading

📉 Paid acquisition Your cost per lead is lying to you — here is the number that isn't ⏱ 7 min 🦷 Paid acquisition What a US dental patient actually costs you — and how to find your ceiling ⏱ 8 min 🌎 Paid acquisition Google Ads language targeting does not do what you think ⏱ 7 min
See all articles →