The short answer
- A call or VSL funnel gives every paid click one chance. In a typical example, about 6 of every 1,000 clicks become clients and the rest are gone.
- Follower ads keep the lead in an audience you own, so a “no” in week one can still buy in month four.
- They usually stall around month three for two reasons: Meta gets no signal about which followers are qualified, and human follow-up breaks down past about 30 active leads.
- The fix is to send qualified leads back to Meta as conversion events through the Conversions API (CAPI), and to let software run every follow-up.
Part oneWhat’s wrong with a standard call funnel?
Every lead gets one shot. If they don’t book in that first session, the money you spent on them is gone.
Take the standard build: cold click, landing page, VSL, calendar, show-up, close. It’s easy to assemble and easy to sell, so almost every ad agency offers a version of it. The pitch shows you the front end and the close. It rarely shows what happens to everyone in between.
The 994 who drop out bounced off a page or half-watched a video. The only way to reach them again is retargeting, which usually costs more than the first click did.
That’s why lead costs feel brutal in these funnels. You pay cold-traffic prices for a single swing. Webinars run the same maths with a time slot attached: a missed slot is a dead lead you already paid for. Every one of these funnels is hit or miss by design, because the lead has exactly one window to say yes.
Part twoWhy are follower ads different?
A follower ad buys the lead, not the moment. Instead of renting one session of a stranger’s attention, you move them into an audience you own.
A follow is the cheapest action you can buy on Meta. Once someone follows you, they see your content every day, they can get a DM, and they warm up on your schedule at no extra cost per touch.
That’s why follower ads can return more than any other funnel a coaching business runs. A lead who says no in week one can still buy in month four, because they never left. You sell high-ticket to people who already trust you, from traffic that cost a fraction of a call-funnel click.
So if the model is better, why do so many people running it watch it die around month three? There are two leaks, and most people running these ads only know about one of them.
Leak oneWhy does lead quality drop on Instagram follower ads?
Because Meta never learns which followers became qualified leads, so it optimises for the cheapest follow it can find, even as your cost per follower keeps falling.
A follower or profile-visit campaign optimises for the follow itself. Nothing about lead quality flows back to Meta, so it never finds out which followers had the budget and which were scrolling on a lunch break.
In month one this doesn’t matter. The algorithm is still exploring, so you get a broad mix and lead quality feels great. That’s the month everyone screenshots. Then you scale, and Meta does the only thing it can with no feedback: it finds you the cheapest possible follow. Every week of spend drifts the targeting further from your ideal client, because your ideal client is never the cheapest person on the platform to reach.
At this stage, nothing is wrong with your creatives or your offer. The pixel got worse on schedule because nobody fed it anything.
The fix is to close the loop with Meta’s Conversions API (CAPI). When a DM conversation qualifies, that lead is sent to Meta as a conversion event, for example a “qualified lead” event, with the details Meta needs to match it to the person. Meta can then optimise your campaigns toward qualified-lead events and build lookalike audiences from people who actually qualified. Instead of hunting the cheapest follow, it starts finding people who look like your best leads, so your ads get sharper the longer they run instead of decaying.
You can set this up yourself if your DM process records which conversations qualify. It’s also the core of what Comuni AI does: it qualifies every conversation as it happens and sends only the qualified ones to Meta.
Leak twoWhy does follow-up fall apart past 30 leads?
Because the volume outgrows what one person can do, even a good setter with the best intentions.
Run the numbers on a modest pipeline: 40 active conversations, each needing six to eight touches before it books or dies. That’s roughly 260 scheduled follow-ups a week, on top of every new lead landing in the inbox at 2am.
What actually happens is triage. The setter works the easy conversations, replies hours late to the rest, and lets every “maybe next month” quietly expire. On a high-ticket offer, each of those is a sale you already paid Meta to start.
“I personally don’t have the bandwidth to engage in 500 to 1,000 messages a day, but the tech does.”
Ashley Conrad, fitness coach
Software doesn’t triage. A well-built AI setter replies to every new follower within minutes, runs every scheduled follow-up no matter how big the pipeline gets, and books the conversations worth booking. Thirty leads a day or three hundred, it behaves the same way. If you keep human setters, the minimum fix is a follow-up schedule they can’t skip and someone checking it weekly.
A warningCan an AI DM setter get my Instagram restricted?
It can if it’s a cheap wrapper. Instagram restricts accounts that break its automation rules, and a restricted account takes your whole audience down with it.
Many “AI setter” tools are ManyChat flows or thin wrappers bolted onto a language model. Before you connect any tool to your account, check that it sends messages through Meta’s official Instagram Messaging API and paces replies like a person would. If your account gets restricted, everything in this article stops working the same day.
For reference, Comuni AI runs on Meta’s official Business Partner infrastructure with direct API access, and none of our clients has had an account restricted from running it.
