Paid + Creative + Attribution

I told them to stop buying the cheap leads.
Then I proved it.

A sports-tech SaaS client sells AI film breakdown to high school football coaches. Two weeks into the engagement I found that the leads the account was buying most of almost never became demos. I built one campaign the other way to prove the point.

$0
Cost per booked demo,
the campaign I built
0
Booked demos per week
from 10% of the ad budget
$0
Cost per demo, the model
I recommended against
0
Demos from a CRM reactivation
campaign, $0 media spend
The call I made

Two ways to buy a demo. Same product, same audience, same nine weeks.

The account was optimizing for Facebook instant forms because they are cheap. My audit found they almost never turn into demos. So I built one campaign that ignored instant forms entirely and sent every click to the page that already converted. Then I let the numbers settle it.

Still running
Instant form campaigns
$175
per booked demo
Ad spend$11,190
Booked demos64
Share of budget90%

Summer Sprint, Hormozi, and the lead magnet campaigns. Not my builds. They optimize for the form fill, not the demo.

5.8x
My build
Website only, coach only
$30
per booked demo
Ad spend$1,176
Booked demos39
Share of budget10%

One campaign. No instant forms. Every click sent to the demo page. Ten percent of the budget, thirty eight percent of the paid demos.

Week 2 of the audit I had already found it: instant form leads converted to demos at 1.1 percent. Website leads converted at 9.9 percent. The account spent the next three months proving me right with eleven thousand dollars.
The ad that did it

One ad. $30 a booked demo. The ad platform never counted a single one.

I positioned the product against the category leader every coach already pays for. Naming the competitor names the problem. Meta was measuring form fills. I was measuring booked demos, so I tracked it in the CRM instead. That is the only reason we know what this ad did.

$0 per booked demo
May 16 – Jul 13
Flagship campaign, pre-audit$477
Summer Sprint$177
Hormozi, inherited$91
My Hudl ad$30
Competitive Fear
"This coach is watching your film."
Demos it touched
49
Apr 22 to Jul 13
Where it was the only paid ad
32
Nothing else to credit

A top of funnel ad is built to be the first touch, not the last click, so the platform will never show you what it is worth. I read the CRM. Thirty two coaches saw this ad and no other paid ad before they booked.

The creative

I replaced his text ads with a cast.

The account was running walls of black and white text to the whole country. Everything on the right side of this page is mine: written, generated and cut with AI. The Film Room is a mockumentary series with recurring characters, so it runs as a season instead of a one off, and a coach who saw episode one recognizes episode four.

His adsStatic text / instant form
Previous text ad Previous lead magnet ad
Link CTR0.94% Cost / click$1.59 Cost / page view$4.21
My adsAI video / coach only
The Film Room, Ep. 1
Mockumentary / recurring cast
The Film Room, Ep. 4
Same characters, four episodes in
Link CTR2.12% Cost / click$0.65 Cost / page view$0.91
Written, generated and cut with AI

Same product, same audience. The AI video put more coaches on the landing page than the campaign it replaced, on 27 percent of the budget.

The library

The rest of my season.

Vs. Hudl
The $30 per demo ad
Hudl race, second cut
Fully AI generated
Scouting report, 2 hours not 10
Fully AI generated

Scripts, casting, edit, and every placement cut. One person, one week, using AI as the production crew.

Week 1 to 2 / The audit

Six numbers. Six decisions.

I spent the first fourteen days reading four months of CRM and Meta data instead of spending money. Every number below turned into a call I made.

Found
What it meant
What I did
1.1%
Instant form to demoThe lead type the account bought most of. Website leads converted at 9.9%.
Built the new campaign with instant forms turned off entirely. Website conversion only.
39.4%
Async demo page to demoThe highest converting asset in the funnel, and almost no paid traffic reached it.
Pointed every paid click at it. One page, one ask, no detours.
$477
Cost per demo, flagship campaignAlso the single largest line of spend in the account.
Cut its budget and used the money to fund the campaign that came in at $30.
72%
Duplicate rate on the lead magnetReported at $1.51 a lead. The real cost of a new lead was $5.69.
Stopped the account from buying volume it already owned.
0
Demos booked by the email nurture2,294 of 2,520 opportunities were stuck in the first two stages.
Rebuilt the sequences and launched a reactivation campaign against the dead list.
58%
Booked in the same sessionIf a coach left without booking, the odds of getting him back collapsed.
Wrote every ad to disqualify the wrong reader before the click, so the right one arrived ready.
My campaign, week by week

One campaign. Ten percent of the budget. It kept climbing.

Booked demos from the campaign I built, from the day it went live. The first three weeks are the Meta learning phase. It has been the cheapest demo source in the account ever since, and it is still growing.

Learning phase Live

Booked demos per week from my top of funnel campaign. From 3 a week at launch to 7 a week now, on $20 a day.

$0Media spend
The free one

A CRM reactivation campaign. Sixteen demos, zero media spend.

A reactivation campaign against the dead leads already sitting in the database. Trigger link email and SMS, scored and routed in GoHighLevel. No new audience, no new budget, no media cost.

616 of them clicked. Sixteen booked a demo. It ran once, in May, and has not been run since. That is the easiest revenue left on the table.

Build log

What shipped, and when.

Every date pulled from the CRM event log, not from memory.

APR 1

Engagement starts. Nothing goes live.

Full audit of Meta and GoHighLevel. Four months of data, 18 workflows, the pipeline, and the calendar.

APR 15

The website-only campaign launches

No instant forms. Coach-only targeting. Four AI-generated video ads and a UTM convention on every one, so the CRM could finally tell which creative produced which demo.

APR 22

The Hudl ad goes live

The one Meta would report zero results for, and the cheapest source of demos in the account.

MAY 4

Learning phase ends. Demos jump and hold.

Weekly booked demos move from single digits into the teens and never return to baseline.

MAY 8

Dead lead reactivation goes live

Trigger link email and SMS with lead scoring. 616 clicks, sixteen booked demos, zero media cost.

JUN 29

Highest demo week of the year

22 booked demos in a single week, against a pre-engagement average of eight.

Where it stands

The $11,190 question.

Ninety percent of the ad budget is still running on instant forms at $175 a demo. The campaign I built runs at $30. This is the case I am making to the client with their own numbers, and it is the same case I would make on day one of any account.

Scenario
Spend
Cost / demo
Booked demos
What the budget does today, on instant forms
$11,190
$175
64
Same money, moved to the model that works, priced at 3x my current cost to be conservative
$11,190
$90
124

I have priced the recommendation at three times what my campaign actually costs, because a $1,176 ad will not hold $30 a demo when you put eleven thousand dollars behind it. It still comes out to roughly double the demos on the same budget.

Whether that budget moves is a business decision, and it is not mine to make. My job is to make sure the number is on the table, in writing, with the data behind it. It is on the table.

Most funnels aren't broken. They're unread.

Paid media, creative, automation, and attribution in one head, so nothing gets lost in the handoff between four people who each own a slice.

Available for select engagements


Funnel Repair Company