Agency Meta Ads Teardown: How an Ecommerce Agency Booked 243 Sales Calls
Want my team to build this for you? Work with Advergent. Rather build it yourself? Join the coaching waitlist.
An agency that sells to ecommerce brands made $289,000 from $59,100 in Meta ad spend between May and August 2025, a 4.89x return. It booked 243 sales calls at $243.21 each by running its own client acquisition as a high-ticket funnel: ads written to one founder profile at a time, a VSL page with an application, and a pre-call sequence.
I ran that lead generation as the agency's fractional CMO, so this teardown is from the inside. The agency stays anonymous, and the dashboards are in the ecommerce agency case study. This post adds the reasoning at each stage and what I'd watch next. It's the second teardown in the series, after the B2B SaaS Meta ads teardown.
Why is it so hard for an agency to get clients with Meta ads?
Because the buyer has heard every agency pitch before. Anyone can start an agency this afternoon, so an ecommerce founder scrolling past your ad has already been pitched "we'll scale your store" by plenty of other agencies. If your message reads like theirs, the founder has no reason to give you 30 minutes.
Ecommerce founders get pitched by agencies constantly, so a generic ad gets ignored and the ones who do click arrive skeptical. That's market sophistication: the market has heard every claim, so a plain promise no longer lands and only a specific mechanism cuts through. Your agency's own acquisition is a different job from the ads you run for clients, because your buyer already knows what an agency pitch sounds like.
Roughly 80% of whether paid acquisition works is decided by the offer, before you open Ads Manager. Take "full-service marketing agency, we work with anyone and do everything." There's no specific problem, person or method in it, and that shape typically lands between $20,000 and $100,000 a month at 80-hour weeks.
Compare it with the example I use in how to build a high-ticket offer that sells on cold traffic: an AI-enabled CRO agency for ecommerce brands doing $1M+. The problem is checkout page conversion, the person is a $1M+ D2C brand, and the method is specific enough that nobody can price-compare it. That's an illustration, not this client's offer, but it's the standard I hold an agency's ads to.
What did the agency's Meta ads funnel look like, stage by stage?
Five stages: Meta ads written to one ecommerce founder profile at a time, a VSL landing page with an application form, a booking page and confirmation page, a pre-call sequence with a setter call within 5 minutes of booking, then the sales call. Each stage gave a skeptical founder one less reason to leave.
| Stage | What we built | Why it matters for this buyer | What the numbers show |
|---|---|---|---|
| Ads | Meta ads written to one founder profile at a time | A founder has to recognize their own situation before they stop scrolling | $59,100 in spend, May to August 2025 |
| Landing page | VSL page plus an application form | Shows how the agency thinks, then lets founders qualify themselves | 243 booked calls |
| Booking | Booking page and confirmation page | Puts a qualified founder on the calendar while intent is highest | $243.21 per booked call |
| Pre-call | Email and SMS reminders, a setter call within 5 minutes, retargeting content | The founder arrives trusting the agency, not comparing it to the last pitch | Not split out in the case study numbers |
| Sales call | The agency's own sales team | Closes a founder who has already done most of the deciding | $289,000 in revenue, 4.89x |
The ad's only job is to make the right founder stop and the wrong one keep scrolling, which is the cheapest qualification there is, because you don't pay for the people it removes. The VSL does the convincing, and the application decides who gets to see a calendar.
What the table can't show is that one person owned the first four rows. As fractional CMO, I owned the whole path from ad to booked call and was judged on revenue, not cost per lead. When the ads, the page and the pre-call sequence belong to different people, each gets optimized for its own number, and nobody owns whether the calls turn into clients.
How do you write Meta ads that ecommerce founders don't ignore?
Write each ad to one founder profile, name a specific problem they already know they have, and explain why the agencies they've tried didn't fix it. If your explanation doesn't make them think "that's why the last two agencies didn't work for me," you have a brand name, not a mechanism.
In this account the ads went to one ecommerce founder profile at a time, not one broad message for every store owner. On a smaller budget I'd stick to three or four avatars with the highest chance of buying, and build the creative around each one's pain points, desires and the proof that would land with them.
Write the mechanism as a fill-in-the-blank: most founders trying to fix this problem focus on the cheap, visible variable, while the expensive, invisible one stays broken. That same paragraph becomes your ad script, your sales pitch and your landing page copy, which is why it's worth writing before you build the VSL around it.
What proof does an agency need before a founder will book a call?
Proof from businesses that look like the founder reading it. Ten to fifteen case studies of people in the prospect's position do more than famous logos, and on an agency's own funnel that proof belongs in the VSL, on the confirmation page and in the retargeting, so the founder keeps seeing it until the call.
One cold email campaign got roughly a 1% reply rate. The only change was a P.S. listing the most relevant case study for that specific company, and the reply rate roughly tripled to about 3% off the same list. Agency buyers respond to the same thing, because the question in a founder's head is whether you've done it for a business like theirs.
That's the job the VSL and the retargeting content did in this account: show how the agency thinks before asking for a call. A VSL for most offers runs 6 to 8 minutes, but in a market that has seen the same offer over and over again, we go down to 3 to 4 minutes and spend that time on standing apart from competitors and building authority. The length rules are in how long a high-ticket VSL should be.
The pre-call sequence carries the proof the rest of the way. A setter called within 5 minutes of every booking, email and SMS reminders kept the call in front of the founder, and retargeting kept the agency's content in their feed. A founder who keeps seeing how the agency thinks turns up already trusting it, so the call goes on their business, not on the last agency that let them down.
Is $243 per booked call good for an agency?
It was here, because the average booked call brought in about $1,190 in revenue against $243.21 to book it, a 4.89x return over four months. A cost per booked call on its own tells you nothing. What makes it good or bad is what a booked call becomes once your sales team has had the conversation.
$243.21 was the cheapest booked call in any of my case studies. The brand-new agency offer paid $271.91, the PR firm about $268 and the high-ticket B2B SaaS $383.09. But the cheapest call didn't bring in the most per call: the brand-new agency offer made about $1,300 per booked call, and the SaaS added recurring revenue that comes back every month. Rank them on cost per booked call alone and you get the order wrong.
So work it out from your own numbers. At my benchmark of a 30% close rate and a $7,000 average deal, a qualified call is worth about $2,100. My own ideal client is worth at least $15,000 in lifetime value, which is why I'd rather pay $300 for a call with someone capable of buying than $100 for a call my closer disqualifies in 4 minutes. The method is in how to tell whether your cost per booked call is actually bad.
What would I watch next in this agency's account?
Stage tracking first. The numbers in this case study stop at spend, booked calls and revenue, so I'd want show rate, call quality and close rate tracked weekly against my floors of 80%, 80% and 30%. Then I'd raise the budget only as fast as the sales team could take the extra calls.
Revenue per booked call blends the show rate, call quality, close rate and deal size into one number. If it slips from about $1,190, the dashboard can't tell you whether fewer founders turned up, more of them were the wrong fit, or the closers stopped converting. Each has a different fix, and only one lives in the ad account. On the brand-new agency offer, tracking each stage showed call quality at 73.2% against my 80% floor, so the next improvement was making the application filter harder, not touching the ads.
Second, closer capacity. At 4.89x over four months the agency had room to keep raising the budget, as long as its team could take the extra calls. I plan around roughly 80 calls per closer per cycle. Past that, booked founders wait days for a slot, nobody calls them within minutes, and the show rate falls for reasons that have nothing to do with the ads. The math is in how many sales calls one closer can handle.
Third, the pace. Raise budgets 10 to 30% every one to three days on the ad sets that have earned it, judge weekly, and keep new angles coming: once an angle wins, 80% of that week's new ads go on it and 20% go on fresh tests. The checks are in scaling Meta ad spend without lead quality collapsing.
What can your agency copy from this teardown?
The structure, not the numbers. Narrow your offer to one problem for one kind of founder, write the ads to that founder, put a VSL and an application in front of the calendar, call every booking within 5 minutes, and judge the account on cost per booked call and revenue, never on cost per lead.
- Write the offer as one sentence first. A specific problem, for a specific person, solved in a specific way. If it doesn't fit in one sentence, your ads will read like every other agency's.
- Say who it's not for in the ad. It's the cheapest filter you have, because it works before the click.
- Put an application before the calendar. Application-first books about 20% of leads against roughly 60% for a bare calendar, but the people who book show up at around 80% and close at about a third. The questions are in the application questions that filter without killing volume.
- Build the pre-call sequence before you scale. A setter call within 3 to 5 minutes of booking, reminders and retargeting, or you're paying for more no-shows, faster.
- Give a new account at least $1,400 over 14 days before you judge it, then review weekly.
If you'd rather have me and my team build this for your agency, apply here and I'll give you my honest opinion on whether we can help. If you want to build it yourself with guidance, the coaching waitlist is open. Either way, start with the offer, because every ad you write for a skeptical founder is only as specific as the sentence behind it. The workbook below walks you through it.
FAQ
Do Meta ads work for marketing agencies?
Yes, when the agency sells a high-ticket service through sales calls and its message is specific enough to stand out. This agency made $289,000 from $59,100 in Meta ad spend between May and August 2025. A generic "we'll scale your store" ad gets ignored by founders who are pitched constantly.
How much should an agency pay per booked call?
It depends on what a booked call is worth to you. This agency paid $243.21 and brought in about $1,190 per booked call. Work out your own close rate and client value before you set a target.
Should agency ads send people to a calendar or an application?
An application, in my accounts. A bare calendar books around 60% of leads and fills your closers' week with people who can't buy. An application books about 20%, and the people who get through show up and close far more often.
What does a fractional CMO do for an agency's lead generation?
In this account I owned the agency's own Meta lead generation from the ad to the booked call, as a part-time marketing lead inside the business rather than an outside vendor. That meant I was judged on revenue, not on cost per lead.