Deck Armor

Franchise funnel

Cold Facebook ad to a bought territory, every stage that is measured and every stage that is not.

data is 721h old, STALE 8:55 AM EST, Nov 9

What to do

Worst first. Every row is derived from the numbers on this page.

25% of registrants show upALARM
204 of 818 people who booked a sitting were in the room, counted off Meet's own reports rather than off CRM tags. That turns a $22.93 application into a $81.85 conversation, and it is the single biggest multiplier on this page. Every point of show rate is worth more than any change to the ads.
These numbers are 721 hours oldALARM
audit_webinar_funnel.py has not completed since then, so every figure below is that stale. Check data/webinar_funnel_guard.log.
The CRM cannot name 93 of the 204 people who cameWATCH
They joined under a display name or an address the roster does not hold, so the matcher refuses to guess and tags them neither way. That is the right refusal, since a no-show text to somebody who attended is the one visible error on a $25,000 decision. But it means no follow-up lane can reach them: they are in the room and invisible to every sequence that keys on webinar-attended.
7.5% of applicants move past AppliedWATCH
46 of 614 cards, 144 still on Applied, 424 parked Cold. 23% of the 204 in the room. $362.97 per applicant past Applied. Of the ones still on Applied, 93 never turned up and 10 were never graded.
22 registrants carry no attendance tagFYI
Counted in the show rate above through Meet's report, not through the CRM. They are excluded from anything that keys on the tag.
No territory has sold yetFYI
$16,696.65 spent, 728 applications, 0 closed. There is no cost per sale and no ROAS on this page because there is no sale to divide by, which is a different statement from the ads losing money.
Applications cost $22.93, against a $50 lineOK
728 applications on $16,696.65 of spend. The top of this funnel is not the problem and spending more on it buys more of the same. The leak is below.

Where the money is

Every tile divides the SAME pot of ad spend by a different stage.

Ad spend
$16,696.65
2026-08-09 to 2026-10-10, Meta
Applications
728
CRM contacts, ground truth
Cost per application
$22.93
target $50
Show rate
25%
ceiling. 204 of 818 booked were in the room; the CRM could name 111, which is the 14% floor
Cost per person in the room
$81.85
same spend, divided by who showed
Territories sold
0
46 off the entry stage

The whole flow, one dollar to one sold territory

Each bar is the share that survived from its own named denominator, not a share of the impressions. The shortest bar is the worst leak.

1. ad impression
$16,696.65 spend
994,334
top of funnel
2. ad click
CTR 1.63%
16,244
1.6% of impressions
of 994,334
3. landing page view
89% of clicks arrive
14,392
88.6% of clicks
of 16,244
4. registration submit
989 started the form, 688 finished it (70% form completion). Meta-attributed; the CRM is ground truth, see stage 5
688
4.8% of landing page views
$24.27 each
5. CRM contact created
0 SSG probes excluded; 1 carry no session date
728
100.0% of Meta-attributed registrations
$22.93 each
6. confirmation delivered
HOLD THE JOIN LINK. 728 of 728 were reached at all, 0 received nothing, 1 were texted but hold no room link because they never replied
727
99.9% of registrants
of 728
7. reminder sequence
registrants whose contact carries at least one reminder-rail tag for their own sitting: 24h email 525, same-day email 624, T-20 SMS 467. 82 got NO rai
644
88.7% of registrants whose sitting has run
of 726
8. attended the webinar
204 humans Meet saw across 19 sittings. The CRM could name 111, which is the floor it reports
204
24.9% of registrants whose sitting has run
$81.85 each
9. post-webinar follow-up
registrants who received anything AFTER their sitting ran. 170 wrote back. 530 of 728 have replied at any point
677
93.3% of registrants whose sitting has run
of 726
10. NDA executed
91 sent and never signed. Denominator is GHL's own document total (239), not the nda-signed tag. From webinar registrants: 64 completed, 69 outstandin
88
39.8% of NDAs actually sent
$189.73 each
11. onboarding meeting booked
registrants whose opportunity sits past Applied on a forward stage (424 more are parked on Cold and are NOT counted). 0 of the cohort reached the Onbo
46
6.3% of registrants
$362.97 each
12. closed / franchise sold
WON and belonging to a webinar registrant. 0 won on either board in total
0
0.0% of registrants
no cost per unit at zero

Show rate, sitting by sitting

Counted off Meet's own attendance reports. What the CRM could not put a name to is its own column, never dropped.

Show rate %, every human Meet saw in the room
050100708-281809-043409-117409-182309-254010-023910-092110-162910-23010-30
SittingBookedShow rateIn the roomCRM could nameIt could not name
2026-08-28157%110
2026-09-014540%18810
2026-09-0411918%22220
2026-09-089022%20173
2026-09-116234%211011
2026-09-1513823%321715
2026-09-181974%14113
2026-09-2211320%231013
2026-09-255723%1358
2026-09-292921%642
2026-10-022040%844
2026-10-062623%642
2026-10-091839%725
2026-10-131429%440
2026-10-161921%422
2026-10-20180%000
2026-10-231429%404
2026-10-271100%101
2026-10-3010%000
All sittings that have run81825%20411193

What one of each thing costs

All out of the same spend. Green only where a target exists.

Cost of oneHow manyEachWhere the line comes from
Application (CRM registration)728$22.93Corey's line, 2026-08-09: $50 per application
Person in the room204$81.85every human Meet saw. Divided by the CRM's 111 tagged it would read $150.42, which is a price on our matcher, not on the funnel
Signed NDA88$189.73no agreed target
Prospect off the entry stage46$362.97no agreed target
Sold territory0-nothing has sold yet, so there is no cost per sale to report

Every franchise opportunity, by stage

Denominator is GoHighLevel's own opportunity list.

Pipeline stageOpportunitiesShare
Franchise Applications / Cold42662.5%
Franchise Applications / Applied19628.7%
Franchise Applications / Pending NDA172.5%
Franchise Applications / Pending FDD162.3%
Franchise Applications / Talked to Dave71.0%
Franchise Onboarding / P2 Payment & File Setup50.7%
Franchise Applications / HQ Visit Done40.6%
Franchise Applications / LVM - Dead after 2nd Attempt40.6%
Franchise Applications / Pending talk to dave20.3%
Franchise Applications / Planning to sign20.3%
Franchise Onboarding / P1 Inquiry & Qualification20.3%
Franchise Applications / HQ Visit Scheduled10.1%
All franchise opportunities682100%

Ground truth is the CRM. GoHighLevel holds 728 franchise registrants; Meta attributes 688. Meta can only see the ones it can match back to a click, so the CRM is the higher and the real number, and cost per application is calculated on it. Show rate is counted off Google Meet's own attendance reports, not off CRM tags. The tag is a floor by construction: the matcher refuses to guess, so somebody who joins as a first name and a last initial, or off a different address, is tagged neither way. Over the sittings that have run it named 111 of the 204 humans in the room, so the honest answer is a RANGE: 14% is the floor and 25% the ceiling, and the gap is people nobody can attribute either way. The ceiling counts every non-host person Meet saw, including short look-ins and people booked for a different day who came anyway. 9 of our own side, and another SSG client sitting in to watch the format, are excluded outright and by name in config/attendance_observers.json; the per-sitting composition is in data/attendance_occupancy.json. Thresholds. The only agreed number on this page is $50 per application (Corey, 2026-08-09, read from client_config.AD_OPTIMIZATION_BY_ADSET). Show rate, cost per attendee and cost per NDA have no agreed target, so they render neutral rather than green. No revenue panel. Nothing has sold, so there is no ROAS to report and a $0 revenue tile would read as a loss rather than as a sale that has not happened. Source. audit_webinar_funnel.py, walked 2026-10-10 12:49 UTC. This page computes no funnel number of its own.