When Should You Switch Ad Agencies? The 2026 Timing Guide
September 18, 2026
To hire a Facebook ads expert for an online course or coaching business, judge them on cash collected and Meta's rules, never on cheap leads or screenshots. The right person reads your funnel the way your bank account reads it. They write ads that sell without the income promises Meta won't run. And they do the work themselves instead of handing it to someone you never met.
You've probably lived some version of this already. The webinar fills up and the applications come in. The calendar looks great on Monday morning. By Friday your closers are talking about terrible show rates and virtually zero closes from the calls that did happen. Meanwhile the report from whoever runs your ads says cost per booked call went down, so on paper it was a good week.
Or it's the other version, where every agency you talk to pitches the same plan. Run a low-ticket offer up front so the ads pay for themselves, then make it all back on the high-ticket program. Except the back end never shows up consistently enough for the math to work.
I hear both stories all the time from course and coaching founders. That's why the real question is not who's good at Facebook ads. It's who can be held to the number your business actually runs on.
That person is one experienced operator who owns everything from the ad to the sale. A traditional agency can work when you need a lot of hands across a lot of channels. But the person touching your account is usually managing 10+ other businesses that week. A coaching funnel punishes split attention fast.
There are two very different funnels hiding under the word "course." They do not share a scoreboard.
A direct-purchase funnel (a low-ticket course, a challenge, a membership sold at checkout) gets judged like ecommerce. You look at cost per purchase and ROAS. You look at average order value once order bumps and upsells land. Then check refunds a few weeks later.
A booked-call funnel (a high-ticket program sold on the phone) gets judged on cost per booked call, show rate, close rate and cash collected. The real verdict lags behind the ad by however long your sales cycle runs. So the early weeks get read on leading numbers and the banked numbers come later.
A generalist usually optimizes for the cheapest booked call because that's what Ads Manager hands them. That's how you get a full calendar and a flat bank account. A booked call that never shows is not a lead.
Someone who knows this business asks for your show rate and close rate before touching a campaign. They'll want call outcomes tracked next to the ad data so you can see which ads produce buyers. Form fills alone tell you almost nothing.
The second difference is the rules. A lot of the hooks that feel natural in coaching are exactly what Meta's Advertising Standards are written to stop.
As of October 2026, Meta's policy on Privacy Violations and Personal Attributes says ads "must not contain content that asserts or implies personal attributes." The list includes physical or mental health and "vulnerable financial status." Meta's own examples of ads it won't run include "Are you bankrupt? Check out our services." and "Depression getting you down? Get help now." That's basically a money coach's hook and a mindset coach's hook, word for word.
The fix is not to get timid. Talk about the outcome and the method instead of telling the reader what's wrong with them. Meta's page says plain "you/your" language is fine as long as it isn't tied to one of those attributes.
Start with the money rules, because they decide which ads are even allowed to run.
Meta's Prohibited Commercial Practices ad policy says ads must comply with the matching Community Standard on Prohibited Commercial Practices. That standard names get-rich-quick schemes. It bans "offers of opportunities of unrealistic financial reward for unclear or minimal effort." If you sell a business, sales or money course, your ads live or die by that sentence.
Trading and investing educators should read that standard's investment section too. It bars offers that reference past returns "to create an expectation of similar future results."
It is not only Meta either. On October 26, 2021 the FTC sent a Notice of Penalty Offenses to more than 1,100 companies, including investment and business coaching businesses. The FTC's announcement says the cases behind it cover "representations that participants will make a profit, or that represented profits are typical."
Your students' wins are not your ad's promise. The person you hire has to know the difference between telling one real student's story and implying every buyer gets that result.
So look for someone who can write a hook for your offer on the call without a single income claim in it. Ask them to do it right there. A good one gets excited about that problem because they've solved it before.
The next thing to check is scope. In this industry the ad is maybe a third of the job. The registration page, the VSL, the application questions and the follow-up that gets people to show up are the rest. If the person only touches campaigns, you've hired someone to fill a leaky bucket faster.
Then ask for proof in your funnel type. They should know cost per booked call, show rate and close rate from memory, or cost per purchase and refund rate if you sell at checkout. ROAS screenshots from a store that sells socks do not tell you anything about your program.
The last one is the one almost nobody checks. How did this person get vetted in the first place? Most of what comes up when you search this question is courses on running Facebook ads yourself and gig listings. None of it tells you how to test the person.
When we built our own filter it ended up as 79 points across 5 stages. One stage hands the candidate a real anonymized ad account to audit and explain to a non-technical owner. The 2026 UM Vetting Benchmark shows fewer than 1% of marketer applicants pass it. You do not need 79 points to hire well. But you should make anyone you're serious about do real work on a real account before you sign anything.
The biggest red flag in 2026 is still a pitch built on cost per lead or cost per booked call that never mentions show rate, close rate or cash. That's someone optimizing for the number they control. It is also the fastest way to end up with closers who hate the leads.
Here are the others I'd walk away from:
There's one more that's specific to this business. Your refund policy and your payment plans show up in your ad account. Meta's Community Standard says it watches user feedback about "failure to honor stated refund or return policies." It also watches for "enrolling customers in recurring payment plans or subscription services without clear disclosure of terms."
When that feedback keeps coming in despite warnings, Meta says it can go as far as "restricting or disabling the responsible accounts or entities." A payment plan nobody explained is an ad account problem. The right hire asks how you sell and deliver, not just how you advertise.
Expect the first month to look quiet. A good operator spends it lining up call outcomes or purchase data with the ads. They cut what's wasting money and figure out what already works so they don't break it.
The second month is when the creative work shows up. Expect new angles, new hooks and usually a rebuilt webinar or VSL page, because in coaching the offer and the page sell as much as the ad. The third month is about pushing budget into whatever proved itself.
On a booked-call funnel you'll finally have enough finished sales cycles to judge it on cash by then. So grade them on cash collected, not calls booked. Grade the early weeks on leading numbers without pretending they're the final answer.
Here's how that played out for two clients. Beyond Driven is a coaching business whose Meta ads were run by a Unicorn Marketer. Julian Bryant wrote on Trustpilot that they were "hitting a ceiling with our paid acquisition at around $180K–$200K in monthly ad spend" and later "achieved our first $1M month."
Dr. Z's Functional Medicine Academy sells high-ticket practitioner education on Meta. It saw a 316% increase in ROAS, a 77% reduction in CPA and an 18% increase in AOV. That's what happened for them, not a promise about what happens for you.
The other thing you should feel within 90 days is speed. Tyler Finkelstein, founder of Self-Talk Plus, an app in the education space, put it this way: "your expert works within your company, your Slack channels. We move faster and get better results than we ever did with traditional agencies."
Most people in this space charge a monthly retainer. Some add a percentage of ad spend and a few offer a revenue share. The structure matters less than what it buys. Revenue share does deserve a closer look in this industry, though.
A pure revenue share sounds perfectly aligned until you work out the details. What counts as revenue on a payment plan? Who eats the refunds? Do they get paid on sales your closers made from your own email list? If you go that route, write those three answers into the agreement before anything runs.
We charge a retainer too and what it buys is attention. Each Unicorn Marketer runs 3 to 5 accounts at a time, while the person running ads at a typical agency is managing 10+. You also approve the specific person before they spend a dollar of your money.
The expensive version of this decision is not the fee anyway. It's three months of ad spend going to the wrong person plus a launch window you don't get back.
If you want more on the money side, I broke down what an ad agency actually costs and how many accounts one person should really be running. There's also a full list of questions to ask before you hire anyone.
Apply for a Unicorn Assessment. We dig into your ad account, your webinar or VSL funnel, your offer and your creative. Then we build custom creatives and funnel concepts for your brand so you see the thinking before any engagement conversation happens.
If it makes sense to keep going, we present two Unicorn Marketers we think fit your business. You grill them as hard as you want, including having them produce sample ads. Then you approve the one you want before any work begins. Within the first 60 days you can switch or walk away for any reason.
Most charge a monthly retainer, some add a percentage of ad spend and a few offer revenue share. The fee matters less than what it buys, so ask how many other accounts the person is running and whether they own the landing pages and follow-up too. On revenue share, agree upfront how payment plans, refunds and sales from your own email list get counted.
Yes, but carefully. Meta's Community Standard on Prohibited Commercial Practices bans "offers of opportunities of unrealistic financial reward for unclear or minimal effort." Its personal attributes policy also blocks ads that imply someone's financial situation. You can sell the skill, the method and real student stories, as long as the ad doesn't promise or imply typical earnings.
Cost per booked call, show rate, close rate and cash collected, with cash mattering most. Cost per lead alone hides whether the calls ever happened. Because a sales cycle can run for weeks, expect early reports on leading numbers and later reports on closed revenue. Ask them to label clearly which is which in every report.
An agency works if you need many people across many channels at once. One experienced person usually works better for a coaching funnel because the ad, the webinar or VSL page, the application and the follow-up are one connected system. Whatever you choose, ask how many other accounts the person doing your work is running this week.
Give it about 90 days. The first month is setup and cleanup, the second is new creative and pages and the third is scaling what proved itself. On a booked-call funnel you won't know your real close rate until a few sales cycles finish, so judge the early weeks on show rate and cost per booked call.
A course can teach you how Ads Manager works, which is useful even if you hire. The hard parts in this industry are writing hooks Meta will approve, rebuilding the webinar or VSL page and connecting call outcomes to the ads. If you'd rather spend that time coaching your students, hiring someone who's done it before is usually the better trade.
%20(1).webp)