Live Call Reference — Confidential

Namgu Kim — One Smile Orthodontics

Diagnostic Conversation — The 43-Lead Question
Date: April 9, 2026 Frame: Diagnostic, not closing Duration: 30–45 min screen share
THE FRAME: This call is not a pitch. It's two operators looking at the same problem together. The job is to learn the conversion truth, design the response, and earn the next move on its own merits.
The Question
The text we sent April 9 reframed the entire next call around a single unanswered number. That question is the call.
The text Namgu received
"Did you ever pull the actual conversion number on those 43 leads from month one? How many became booked consults, how many turned into contracts?"
That number turns the rest of the data into something we can act on. Until we know it, every recommendation is theoretical. The call's job is to surface the answer (or the absence of an answer) and design what comes next from there.

Why this is the right question and not a pitch: 43 leads is a top-of-funnel metric. It tells us nothing about whether the funnel actually moves people to chairs. Until we know consults-booked and contracts-signed, we don't know if the leak is at acquisition (cold), qualification (warm), front-desk handling (the human layer), or close (clinical/financial). Each leak is a different fix. Asking him this proves we think like an operator, not a vendor.

What we already know about his answer
Namgu has no formal attribution. His front-desk EA asks patients where they heard about the practice (and possibly notes it on an intake form). That is the entire system.
Confirmed context from Kyle, April 9. This means the call won't be a discovery of which branch — we already know it's Branch A. The call's actual job is to (1) confirm this on the call without making him feel called out, (2) name the failure modes of human-recall attribution as proof of expertise, and (3) propose what real measurement would reveal that he is currently flying blind to.
The Real Frame — Human Recall Is Not Attribution
Because we already know Namgu's answer, the call is no longer a discovery. It's a structured conversation about why the system he has does not produce the truth he needs — without making him feel attacked for it.

The trap to avoid: Walking in with "you have no attribution" energy. His EA is a real person doing her job. His intake form (if it exists) is a real artifact. Calling them worthless makes Namgu defend them. Make the EA the hero of the conversation, then surface the structural reason the recall system can't tell him what he actually needs to know.

Step 1 · Acknowledge what exists

"Tell me how she captures it right now."

Open with curiosity about the existing system. Does she ask verbally? Is it written down? Does it go into a spreadsheet, the patient management software, or just her head? Is the intake form a checkbox list ("Google / Friend / Instagram / Other") or a free-text field?

Listen to the answer. The shape of his system tells us exactly which failure mode is dominant.

Step 2 · Name the four failure modes

Why human-recall attribution lies — even when the human is great

(1) Recency bias. Patient says "Google" because that's the last thing they touched — but the actual decision was made when they saw a Reel three weeks earlier and then googled the brand name to find the website. Last-touch buries first-touch.

(2) Conflation. "A friend told me" — but the friend told them after they'd already seen the ad and were primed. The ad earned the receptivity. The friend gets the credit.

(3) Social desirability. "Friend" sounds better than "Instagram ad." Patients unconsciously upgrade their answer to whatever feels more flattering for the practice.

(4) The ones who never came in. The EA can only attribute the patients who showed up. The 30+ leads who clicked, called, didn't book, didn't show — they are completely invisible to her system. That's where the leak is, and it's the part no human-recall system can ever see.

Step 3 · The pivot to measurement

What real attribution would reveal in 30 days

Lead-source tracking (CallRail or equivalent) on every inbound number, tied to Meta's ad ID. Pixel + conversion API on the landing page tied to form submissions. A simple spreadsheet the EA fills in once a week tying intake forms to ad source via the lead's phone number. None of this replaces her. It gives her a system that catches what her memory cannot.

Concrete promise: 30 days after install, Namgu will know — not guess — which of his 2 ads is producing real chairs vs. real noise. He'll know what his 43 leads were actually worth. He'll know if doubling Meta budget was the right call. Right now he is operating on the EA's confidence, not on data.

The line that bridges into the offer (only if natural) "What I'm describing isn't a Cakesmash thing. It's table stakes for any practice spending real money on Meta. The reason we don't see it in your setup isn't a failure of your team — it's that nobody has had the job of putting it in. That job is the first 30 days of the Founder Pilot, if we end up working together."
Edge case · He surprises us

If the intake form turns out to be more structured than expected

If Namgu pulls up the form mid-call and it's a checkbox source list that's been entered into his patient software for the last month, that's a real (if imperfect) signal. Acknowledge it: "Okay — so you actually have something. Let's look at it." The four failure modes still apply, but he has a starting denominator. From there, treat the conversation like a Branch B from the original frame: walk the funnel by stage, find the leak, propose the fix.

Numbers Cheat Sheet
Glance reference. The unknowns are intentional — the call is built around the gray cells, not the white ones.
Case Value
$5K
Confirmed Apr 1
Leads (30d)
43
34 calls + 9 form fills
Consults Booked
?
The call's job
Contracts Signed
?
The call's job
Meta Spend (30d)
$1,275
Now ~$2K/mo post-Apr 1
Best Demo CPL
$20
35–44 cohort
Founder Pilot
$3,997
/mo · 6mo
Cakesmash Breakeven
~0.8
Cases/mo to pay for itself*

* Breakeven math: $3,997/mo retainer ÷ $5,000 case value = 0.8 cases/mo to cover the retainer at the doctor's gross. Does not account for case profit margin, COGS, chair time, or attribution lag. Honest enough for a verbal reference, not for a written deliverable.

Founder Pilot — If It Surfaces
This is NOT the call's purpose. Only walk through it if Namgu asks "okay, so what would working together actually look like." If he doesn't ask, do not pitch.

Founder Pilot

$3,997 /mo
6-month commitment · 90-day exit ramp · LA practices only
  • Monthly diagnostic against patient revenue, not vanity metrics
  • One half-day production at One Smile per month: 3–4 ad-ready creatives + 4 organic Reels
  • Lead-source tracking install (CallRail or equivalent) + landing-page measurement
  • Monthly creative diagnostic on what's working and why
  • Direct line to Kyle for creative and strategy calls
  • Founder pricing locked for the full 6 months. Standard tier is $5K/mo.
Talking-Point — Only If He Asks

"It's a 6-month engagement at $3,997 a month with a 90-day exit ramp built in, so if it isn't working at month 3 either of us can walk. The first month is mostly measurement and one production day. By month 2 you have new creative running and a real funnel dashboard. By month 6 you own a creative library no SGV competitor can match in a year. That's the shape of it. But honestly the reason to do it isn't the deliverables — it's that you'd have someone in your account every week who actually thinks about your business instead of running platform defaults."

The One Goal
The goal flexes with the branch. Read this last, right before you dial.

The goal: Get Namgu to verbally agree that the next 30 days should be about installing real measurement, not about doubling spend further. Never end the call with "I'll think about it." End with one of these three:

Best case: "Yes, let's do the Founder Pilot. Month 1 is measurement install. Send me what I need to sign." Verbal yes on the engagement.

Middle case: "Walk me through what installing measurement would actually look like — what does it cost, what does it touch, how long until I have data." A booked 30-min follow-up scoped specifically to the measurement install, with the explicit decision criterion: "if the first 30 days of data shows a real conversion leak, we sign the 6-month. If it shows everything is fine, we walk and you've still got the measurement."

Floor case: "Let me think about it." Do not let this be the closing state. Push back gently: "Totally fair. What specifically do you need to think about? Is it the measurement piece or the engagement piece? Because those are two different decisions and we should separate them." Then either get a yes on measurement OR get a date for the next conversation.

The closing line

"Here's what I want to leave you with. You've been spending real money against a number you can't see. Your EA is doing the best a human can do with this, but no human-recall system can tell you what your 43 leads were actually worth or where the 35 you didn't close went sideways. The first 30 days of working together fixes that. Whatever you decide on the bigger engagement, the measurement is the part you cannot afford to keep skipping. Want to start there?"