You built a successful med spa. So why does this still feel like it's running you?
Credits her before asking anything. It flatters without measuring, which is fine for a hook.
Case study / Ad dissection
Past last click.
All the way to the sales call.
Last-click attribution tells you which ad got the lead. It goes quiet the moment that lead books a call, which is exactly where you find out if they could ever buy.
We join the whole journey. Ad, comments, form, calendar, sales call, every follow-up. That shows who your best clients are, what brought them in, and what your worst-fit leads said in their own words. Then we change a few lines of copy to turn the wrong people away, and keep the ad that works.
- Ad
- Page
- Form
- Calendar
- Call
- Verdict
What changed
Most reporting stops at the click. The sale happens on the call.
Attribution tells you
Which ad got the lead
- Which ad got the first or last click
- Cost per lead, cost per booking
- Six dashboards, each with one piece
- A winning ad you are scared to touch
Dissection tells you
Who the lead turned into
- Which ad brought in your highest-value clients
- What the worst-fit leads said on the call, word for word
- The exact seconds of the ad that attracted them
- What to swap, and what to leave alone
The short version
One great ad. The wrong people booking.
A med spa consultancy's best ad was filling the calendar with owners too small to buy. Here is what the full trail showed.
80%
of all leads
came from one 90-second video
3 to 1
Unqualified to qualified
on the calls it booked
1 word
8 wrong calls
traced to one second of script
4
Passages swapped
the proven lines kept word for word
Blurred for client privacyThe ad
Video 11, the original take
The co-founder, straight to camera for 89.6 seconds. It was the best ad the account had run, so the brief was never to replace it. The brief was to find out why the wrong people were booking, without breaking the part that worked.
Spend
$14,559
70% of account spend
Leads
135 of 169
at $108 each
Calls graded
46
read and graded for fit
Verdict mix
27 to 9
unqualified to qualified
01 / Connect the data
Six systems, one contact record.
Can a sales call be tied back to the ad that caused it?Nothing below works unless a call verdict can be matched to an ad. So the first job was the join. Each system holds one piece, and the contact record in the CRM is where they meet.
- MetaThe ad runsSpend, click-through, three-second views, saves, landing page views.135 leads reported
- Landing pageThe click is taggedEvery visitor from this ad carries the ad's own tag.815 page views tracked
- Application formShe states her revenueThe form asks for annual revenue before the calendar appears.125 contacts, 112 answered
- CalendarShe books 45 minutesBooking and show status are written to the same contact.86 booked
- Call recordingThe call is recordedFull transcript of the founder's call with the owner.51 calls transcribed
- Call gradingA verdict is written backEvery transcript is read and graded for fit, without seeing the form answer.46 verdicts on Video 11
Why grading blind matters. The form answer is what she said about herself before the call. The verdict is what the transcript showed during it. Because the grading never sees the form, the two are independent reads of the same person, and they can be checked against each other.
02 / Find what works
Every ad hooks about the same. Only this one converts the page.
Is the hook the thing that works?The working belief was that the opening line pulled people in and the rest failed to filter them. If that were true, the fix would be a rewrite of everything after the hook. The account data says otherwise.
| Ad | Watched 3 seconds | CTR | Page view to lead | Saves | Leads |
|---|---|---|---|---|---|
| Video 11, June | 28.4% | 3.17% | 16% | 45 | 60 |
| Video 11, August | 28.0% | 3.29% | 10% | 30 | 44 |
| Static carousel | n/a | 3.32% | 13% | 11 | 15 |
| Video A | 24.6% | 3.40% | 7% | 1 | 5 |
| Video B | 27.7% | 3.84% | 6% | 2 | 4 |
| Video C | 36.3% | 1.09% | 7% | 4 | 3 |
| Video D | 21.2% | 3.67% | 2% | 1 | 1 |
Ads above $400 spend, names withheld. Three-second bars on a 0 to 40% scale, page-view-to-lead bars on 0 to 20%. A third Video 11 ad set is left out because its page tracking broke. The carousel is static, so it has no view rate.
The left bars are nearly the same length for every video, and a newer ad beats Video 11 on the opening three seconds. The right bars are where it pulls away: same page, same form, same offer, two to five times the completion. Saves tell the same story, 30 to 45 against 1 to 4 for the other videos.
What this ruled out
The asset is the middle of the ad: the burnout confession and the systems line. So the middle could not be touched, and any fix had to be surgical.
03 / Test the funnel
The funnel does not separate them. Both groups reach a call at the same rate.
Does the funnel filter out the wrong owners?The form asks for revenue, so every contact can be split by her own answer. Under $500K is below the qualifying floor. Each panel follows one group from form fill to a call that happened.
Said under $500K
Below the floor. 69% of everyone who answered.
Leads
77
Booked
51 · 66%
Showed
16 · 21%
Said $500K and above
At or above the floor. 31% of everyone who answered.
Leads
35
Booked
29 · 83%
Showed
8 · 23%
Lead to booked
66%vs83%
Booked to showed
31%vs28%
Lead to showed call
21%vs23%
125 contacts from this ad. Thirteen gave no revenue answer and are left out. Show status is the CRM field as recorded.
Owners above the floor book a little more readily, and then the gap closes. By the time a call takes place the two groups have arrived at 21% and 23%. Because the under-floor group is more than twice the size, it supplies 51 of the 86 bookings and 16 of the 26 calls that showed.
What this ruled out
Conversion rates cannot tell the two audiences apart. The numbers looked healthy all the way down. The difference had to be found in what was said on the calls.
04 / Listen to the calls
The wrong callers kept naming one item from the ad.
What did they say once they were on the call?The call verdicts confirm what the form answers suggested, from a source that never saw the form.
Unqualified
Qualified
Pre-launch
Fit unknown
Then the call summaries were searched for what brought them in.
The sales team had said the video talks about memberships. It does, for about one second, as the third item in a list of six. Reading all 51 graded call summaries for that word produced this.
51
graded call summaries, all ads
15
mention membership
11
of those came from Video 11
8
of those were unqualified or pre-launch
Unqualified or pre-launch
Xxxxx · solo provider who also owns a salon
She saw a Xxxxxx ad about memberships and booked the call.
Xxxxxx · $15k a month
Exploring memberships and recurring revenue models.
Xxxx · $20k to $25k a month
Interested in membership strategy.
Xxxxx · pre-launch, no revenue
Prioritizes memberships over new client acquisition.
The owners the offer is for
Xxxxxxxx · qualified · $60k a month, 5 providersA real fit on every measure. She came in wanting her membership programme relaunched and turned the engagement down on price. She was buying the item, and the offer is the whole system.
Owner B · multi-location, seven figuresThe profile the offer is built for. Used in step 05 as the test reader for each line of the script: would she recognise herself in it?
Quotes are from the call summaries, written from each recorded transcript. Names blurred for privacy.
What this showed
A list of six deliverables works as a menu. Callers booked for one item on it, and the item that appeals most to the smallest operator is the one that promises recurring revenue without more patients.
1 word.
8 wrong calls.
…a lifetime client system, a membership framework, a reputation engine…
About one second of a 90-second ad. Here is how we found the rest.
05 / Mark up the ad
The 89.6 seconds, line by line.
Which seconds of the ad are doing it?With the call evidence in hand, each line was read against two real owners from the pipeline: Owner B, multi-location with a full team, and Owner A, working three days a week at $240k a year. The bar is drawn to scale.
0:000:301:001:29
Keep word for word
Defect
Weak
Light edit
...you're still the last one to leave, the first one stressed, and the only one doing everything that nobody else seems to be able to do.
A portrait of a solo operator. True for Owner A, false for Owner B, who runs several locations with a full team. The owner the offer is built for cannot say yes to it.
I managed xxxxx xxxxxxxxx and xx xxxxxxxxx, and I'll be completely honest with you when I said I burnt out.
The credential, and the honesty the middle depends on. Also the only scale numbers in the script, and they belong to the seller.
It wasn't that I was working too hard. It was that I had never built the systems that would let the business run without me at the center of everything.
The line earning the saves and the form completions found in step 02.
I'm Xxxxxxx, the co-founder of Xxx Xxxxxx Xxxxxxxxxx... I now work with med spa owners across the country to do the same exact thing.
Fine as it is, but it never says which owners. One phrase gets added.
We install the growth infrastructure that turns a busy practice into a scalable, profitable business.
The only qualifying words in the ad, and "busy" is self-assessed. A solo injector with 40 clients a month is busy.
...a revenue leak audit, a lifetime client system, a membership framework, a reputation engine, a trained front desk, and a team that performs...
The menu from step 04. One word, about one second of airtime, eight wrong calls.
Owners who go through this stop being the bottleneck and start being the visionary.
Proven line. It stays.
...see if you qualify... We're very selective about who we work with, because we commit and we guarantee results. Apply now...
A gate with no stated bar filters nobody. It reads as a dare, and the owner with $8k months books to find out where the bar is.
The structural finding
The script contains no number the viewer has to meet. An ad that assumes nothing about her fits everyone.
06 / Make the swap
Four passages swapped. The proven lines stay word for word.
What changed?One rule governed every new line: could a solo injector say this about herself? If yes, cut it. The new script describes a world she is not in and never tells her she failed a test.
00:05The recognition test
Was
You did everything right. You built something real, and somehow you're still the last one to leave, the first one stressed, and the only one doing everything that nobody else seems to be able to do.
Now
You have the team. Injectors, a front desk, an ad budget going out every month. The rooms are busy, the work is excellent. And the number at the end of the month looks almost exactly like it looked last year.
That is not a work problem. It is that every decision still comes back to you. And a practice where every decision routes through the owner has a ceiling with her name on it.
Assumes injectors in the plural, a front desk and a live ad budget. The bottleneck idea survives, attached to scale.
00:40Who she works with
Was
I now work with med spa owners across the country to do the same exact thing. We install the growth infrastructure that turns a busy practice into a scalable, profitable business.
Now
I work with med spa owners who are already doing real volume and still cannot get past their current number.
Screens without a revenue figure being spoken aloud.
00:48The offer
Was
In just 90 days, we build the full engine together: a revenue leak audit, a lifetime client system, a membership framework, a reputation engine, a trained front desk, and a team that performs because they have the tools.
Now
We do two things at the same time, and it only works as both. We bring more of the right patients in, and we build the system inside your practice that keeps them, so the patients you already paid for come back and spend more. Same team, same rooms, a bigger number.
One mechanism, nothing to pick off and shop for. The wording is lifted from the ad's own written copy, so the video and the text now say the same thing.
01:06The close
Was
Click the link below and see if you qualify for a free strategy call with our team. We're very selective about who we work with, because we commit and we guarantee results. Apply now, and let's find out if you're ready.
Now
Book the free 45-minute growth assessment. We'll find where revenue is leaking in your practice and map what to fix first, second and third. You keep that plan whether we work together or not.
Names what she gets. An unnamed free call is a calendar slot.
Kept exactly as recorded
- You built a successful med spa. So why does this still feel like it's running you?
- I know that feeling because I lived it...
- It wasn't that I was working too hard. It was that I had never built the systems...
- Owners who go through this stop being the bottleneck and start being the visionary.
About 72 seconds against the original 90. The new cut runs in its own ad set at equal spend against the original.
07 / Be straight about it
What the script alone cannot fix.
And where this evidence is thin.- The bigger lever is routing. The form asks the revenue question and the calendar ignores the answer. 51 of 86 bookings came from owners who had just said they were under $500K. At $517 per showed call, the 16 who showed cost roughly $8,000 of founder time. An under-floor answer should go to a lower-tier offer or to nurture.
- The sample is small. Nine qualified calls and eleven membership mentions are enough to point at a passage. They are not enough to size the effect, which is why the new version runs head to head against the old one.
- It is judged on cost per qualified call. The rewrite is meant to reduce lead volume, so cost per lead would mark the winner as a loser.
- Two counts do not reconcile. The CRM marks 26 calls from this ad as showed, while grading holds 46 verdicts for the same ad. They come from different records and were not matched one to one.
The lesson
Your ad picks who calls you. Read the calls and you get to pick too.
- Cost per lead hides the problem. Every number looked healthy until the calls were read.
- Your worst-fit leads tell you what to cut. In their own words, on the call.
- Fix the seconds, keep the winner. Four passages changed. The lines that earned the saves did not.
Client anonymised, names and footage blurred. Figures are from the client's own Meta account, CRM and recorded sales calls, June to September 2026.
Your turn
Do you know what your best clients said before they bought?
Book 25 minutes. We will look at where your ad data, your calendar and your sales calls live today, and what it would take to read them together.
Book 25 minutes