A lead-count guarantee is a different product, not a worse one
If you've looked for marketing help in the last year you'll have seen these. "Fifteen to thirty plus leads, in under thirty days, or you don't pay." "Add forty to three hundred thousand in new monthly revenue within a hundred and twenty days, or we work for free." "If you don't make back at least eight times your investment, you don't pay."
Those are real ads running in Australia right now. I'm not naming anybody, partly because it isn't the point and partly because you can go and read them yourself in the Meta Ad Library in about four minutes, which is a better use of your time than taking my word for it.
I want to be genuinely fair about these, because the easy version of this article is a sneer and that helps nobody. Some of these offers are honoured. Some of the people running them are good at their jobs. And for a particular kind of business, buying one is a perfectly sensible decision. What I'd like to do instead is explain what you're actually buying, because it's a different product from the one we sell, and the difference isn't quality. It's what the thing is pointed at.
A guarantee has to be written against something the seller controls
That sentence is the whole article, so it's worth sitting with.
Nobody can guarantee you booked jobs. Whether a job gets booked depends on your quote, your price, whether you rang back, how you came across on the phone, what the customer's budget turned out to be, and whether their partner liked you. The agency controls none of that. If they guaranteed it, they'd be guaranteeing your performance, and they'd go broke on the first client who doesn't answer the phone.
So the guarantee gets written against the last thing in the chain the agency does control. That's the number of enquiries. Sometimes it's the number of form fills, sometimes calls over a certain length, but it's always a count, because a count is the only thing they can produce without your cooperation.
That's not a trick. It's arithmetic. If you want a guarantee, it has to attach to a count, and once it attaches to a count, everything downstream of it changes.
What happens next is entirely predictable
Put yourself on their side of it for a second. You've promised thirty leads in thirty days and you don't get paid otherwise. What do you do?
You make the enquiry as cheap as possible to give. You run the campaign objective that produces the most responses per dollar, which usually means a form that prefills the person's details, or a message thread, because both take one tap. Both have a specific failure mode and both have been written up here already. 
A Messenger conversation isn't a lead and The lead form settings that decide whether your leads are real You widen the targeting, because narrow targeting costs more per response. You write the ad to be appealing rather than qualifying, because a question in the copy is a person who doesn't click.
Every one of those decisions is correct if the number is the goal. Every one of them also selects for the least committed person available, because the least committed person is by definition the cheapest one to collect.
And that's how a business ends up telling me that all its leads are price shoppers. Not because the agency was bad. Because the agency was asked for a count and delivered a count, and a count is what it optimised into existence.
I've watched a business pay three different companies at once, each briefed on volume, each producing volume, and the owner's honest summary was that none of it turned into work. The brief was wrong, not the three companies. The longer version of that argument, and why the website is the thing deciding which jobs arrive, is here. The job you didn't get asked to quote
The conditions are the part nobody publishes
Here's the practical bit, and it's the finding I'd want if I were you.
I went through about twenty of these ads and the landing pages behind them. The guarantee is always in the headline. The conditions are on none of the pages. Not hidden in small print at the bottom, not on a terms page, not linked. Simply absent.
That doesn't mean the conditions are unfair. It means you can't read them before you get on a call, which is a choice somebody made about the order in which you find things out.
So ask, in writing, before you sign. Four questions.
What counts as a lead? A form fill from someone who typed a fake number is a lead by most definitions. Ask whether a lead has to be contactable, and what happens to the ones that aren't.
What do I get if you miss? This is the one that matters most and the one people forget to ask. Most of these make good in more service, not in money back. More months of work is a different thing from a refund, particularly if the reason you're leaving is that the work didn't suit you.
What do I have to do for the guarantee to hold? There's nearly always an obligation on your side: a minimum spend, a response time, using their tracking number, not pausing the campaign. All reasonable. All worth knowing before rather than after.
Can I see the last three months of results for a business like mine? Not a testimonial. The actual numbers, with the client's name removed if they'd rather.
Worth knowing what standard those headline promises are held to, as well. The ACCC's guidance on false or misleading claims applies to a guarantee in an ad the same way it applies to anything else a business says to win work, and asking what a promise actually commits to is exactly the check the law assumes a buyer will make.
If those come back clearly and fast, you're probably dealing with someone straight, and the offer might be right for you. If they come back as "let's jump on a call", you've learned something for free.
When you should buy one
I said this was a different product rather than a worse one, so let me make the case for it properly.
If your work is genuinely volume work, if every job is worth roughly the same and the difference between your best and worst customer is small, if you can absorb a lot of enquiries without it wrecking your week, and if you're mostly short of people knowing you exist, then a count is a reasonable thing to buy. A guarantee on top of it is a reasonable way to limit your downside. I'd take that deal over a vague retainer with no promises attached, and I'd say so to your face.
That describes a real set of businesses. Mobile services with a fixed call-out. Anything where the job is standard and the margin comes from doing a lot of them.
What it doesn't describe is a business where one job can be worth thirty of another. In that business the count is close to meaningless, and optimising for it actively works against you, because the enquiries that are cheapest to collect are the ones least likely to be the job you actually want. Deliberately making the enquiry harder to give is one of the few reliable fixes. The jobs you don't want are the ones you're asking for
Why we won't offer you one
Not because we're above it. Because we'd have to break our own argument to do it.
We tell clients that the count isn't the point, that the question is which jobs arrive rather than how many. If I then guaranteed you a number, I'd have every reason in the world to send you the cheapest enquiries I could find, and you'd have no way to tell the difference until you'd wasted three months ringing people who don't remember filling anything in. I'd be pointed at the thing I just told you not to care about.
The same logic is why we charge a flat fee rather than a percentage of your ad spend, which is the same problem wearing a different hat. If your agency is paid a percentage of your ad spend, read this
What we'll promise instead is the boring version: we'll tell you what we're doing and why, you'll be able to see what happened, and if the honest answer is that you shouldn't be advertising yet, you'll get that answer rather than a campaign. The whole method is published, including the parts that would let you do it without us. The Foundational Marketing Blueprint
That's a worse thing to put in a headline. It's the one I can actually keep.