Five stages, two levers, and one part everybody gets wrong.
The whole model on one page. Part 1 is the business: who to sell to, what to charge, how to get clients. Part 2 is the machine that delivers it. Part 3 is where the experts contradict each other and where they're selling you something.
12
videos distilled
6
creators, not one guru
5
stages in the machine
2
levers that move money
5
things they disagree on
Start here
The whole business, one picture
Two loops. The top loop gets you paid. The bottom loop keeps you paid. Most people only ever build the bottom one and wonder why they have no clients.
Scroll the diagram sideways on a phone.
Part 1
The business
Nobody pays you for a machine. Sort this half first.
The model
01What you actually sell
One word decides whether this business works: not leads, but appointments.
The old way
Agency hands over a spreadsheet of leads. Client is bad at phoning them. Nothing sells. → "Your leads were rubbish."
→
What you sell
You call, qualify and book them into the client's calendar yourself. → You control the conversion, so the blame has nowhere to go.
Sell the outcome, stay vague on the parts. "We help kitchen remodelers get high-value consultations using AI sales systems" beats "we run Facebook ads". The second one is a commodity anyone can undercut. Ads, pages, CRM and bots are plumbing — never the pitch.
Who to sell to
02Cast a net first, then pick up the rod
"Niche down" is advice for agencies that already have data. Starting out you don't know who you're good with — so go wide to find out, then go narrow on purpose.
Do you already know an industry from a past job, hobby or family business?
Yes → start there
That's your unfair advantage. Skip the net entirely. One student went into motorsport because he'd worked in F1 and made 250k in four months.
No? Then cast the net.
Pick 3–5 niches. Work with anyone who fits.
Drop the ones you hate, keep the ones that pay. You're buying information, not just revenue.
Once you know who pays → pick up the rod.
Go narrow, deliberately
Broad = a different strategy per client, pricier staff who can handle all of it, thinner margins, more stress. Narrow = copy-paste. Narrow is calmer and richer.
Which niches, and in what order
Stage
Niches
Why these
Prove it Start here
High-ticket home services — solar installation, roofing (bigger firms), full renovations and extensions, commercial cleaning.
High-ticket health — cosmetic dentistry, aesthetic clinics, weight-loss clinics, specialist gyms, high-end med spas.
Barely touched by tech, so the fear of missing out on AI is loud and real. Easy to sign while your confidence and case studies are thin.
Scale it Later
Legal — personal injury, family law, insurance and loan claims. Financial — tax prep, wealth management, debt consolidation, credit repair. Education — online courses, coaching, online trainers. Plus recruitment, staffing, B2B software.
Already buy leads, often badly and expensively. Bigger budgets, longer lifetime value, and they'll do pay-per-lead deals. But they'll shred you without proof.
Screen every prospect on four things
👥
2–10 staff
Bigger firms are too sophisticated for a new agency. One-man bands can't afford you.
💰
Over ~USD 20k/month
Below that your fee eats their profit and they'll churn even if they sign.
📍
Single location
Not a franchise. You need to reach the person who can actually say yes.
🔁
Repeat customers
That's where lifetime value lives — and lifetime value is what pays your retainer.
Know your ceiling before you commit
10,000
Gyms in the country (everything that exists)
6,000
Pass your four screens (~60%)
300
A realistic 5% share of that
RM1.8m
Per month, at RM6k each. That's the ceiling.
Do this maths before you commit to a niche, not after. It's the difference between a niche that can carry your goal and one that runs out at RM40k a month.
What to charge
03Price the value, not your hours
$500–1,000
First 1–3 clients. "Don't charge more than that." A case study is worth more than the money. A free trial for a testimonial is a fair trade here.
$1,000–2,500
Proven, with results. Typical local-business range. A gym tops out around $1,500–2,500 once established.
$2,500–10,000+
Established and niched. He reports one client at $50,000. Rare — but the ceiling is the value you can prove, not the hours you work.
Developing market = smaller retainer. He says it outright: same service, lower number, until your proof is strong enough to sell overseas.
Why a cheap lead isn't a cheap lead
The number that sets your price isn't what a lead costs. It's what a lead is worth across the client's whole relationship with that customer. Almost no agency works this out, and they underprice themselves forever.
$100
Teeth cleaning the front-end offer
$300
Teeth whitening upsell
$5,000
Invisalign upsell
$20,000
Implants upsell
$600
What one $10 lead is really worth once you track it all the way through
Track it backwards and your job changes. Once the client sees that a RM40 lead becomes RM2,400 of lifetime revenue, they stop haggling over your fee and start asking how fast you can spend more. Clients who understand their lifetime value spend more, stay longer and complain less.
Getting clients
04The outreach maths
This is the part the delivery videos skip. Client acquisition is arithmetic, not luck.
How often each channel books you a meeting
Cold calling
~4% · 25 dials = 1 meeting
DMs
~1% · 50–100 per meeting
Cold email
~1% · 50–100 per meeting
Door to door isn't on the chart because he gave no number — only that it beats everything. Hardest, most intimidating, and that's exactly why it works: nobody does it any more. One student signed 6 clients in 2 months this way.
One client a week, worked backwards
125
Dials a week — 25 a day, five days
4% book
5
Meetings booked
60% show
3
Actually turn up — they Google you first, and early on you look thin
~30% close
1
Client signed. Every week.
Bad close rate? Multiply the top, don't quit. At 10% closing you need 3 weeks per client. Triple the dials to 75 a day and you're back to one a week. 75 dials is about 5 hours for a beginner. These numbers are a dial you control, not a verdict on you.
Three rules that decide whether outreach works at all
Bank 200–500 leads before you start calling. Run dry mid-week and you stall, then you never restart.
Never pitch during outreach. The only goal of a call, DM or email is to book the meeting. Pitching cold means no prep, no visual presence, no leverage. If they push: "Great question — I don't have time to do it justice now. Free tomorrow at 10?"
Spend 70% of your time on what already works, 20% improving it, 10% testing new things. Most people invert this and never compound.
Getting past the receptionist
Try 1 · assume you know them
"Can you pop me through to John?" The second you sound like a salesperson, they do their job and block you.
→
Try 2 · be vague
"John knows what it's about." Don't explain. Explaining is pitching.
→
Try 3 · honest, still no pitch
"I've got some clients I want to send his way." Never rude — that person can lock you out permanently.
Cold email limits worth respecting: 15 a day per address to start, up to 30 after a couple of months. 2–3 addresses per domain. Plain text beats HTML. Warming always on, or you get blacklisted and lose the domain. DM tools cap around 50/day per account and break platform rules — use spare accounts, never your own.
Closing
05The sales meeting: four moves
A conversation that solves a problem, not a pitch. In every meeting somebody is in control — make sure it's you.
1
Attention
Open with something only research would tell you — a trip, a hobby, their town.
Then set the frame, out loud.
"I'll ask some questions about your business. If I see a way we can help, I'll make you an offer at the end. Sound good?"
2
Identify
Where are they now? Where do they want to be?
Ask revenue directly, or ask goals first if that's easier.
Listen properly. Ask follow-ups.
"What's stopping you solving this yourself?" … "What's it costing you to stay where you are?"
3
Solve
Answer only the pains they named.
Listing every feature is the top mistake — every extra thing is a new reason to say no.
Short statement, then shut up and let them ask.
"We help local gyms get more free trials using AI sales systems." Then say nothing.
4
Close
State the price. Then stop talking.
Whoever speaks first after the price loses.
Adding features after the number screams you don't believe it.
"It's RM6,000 a month." Silence.
An objection is not a rejection. It's a request for more information, and usually a sign something earlier was weak. Over half his closed deals had one first.
They say
You say
What you're doing
"I need to think about it."
"Of course. Is it a value question or a price question?"
Splits "I don't see it working" from "I can't afford it". Two completely different problems.
"I need to ask my wife / partner."
"Are you personally in? … So what could they say that would change your mind?"
Usually a polite mask. Move the blame off them and the real doubt comes out, where you can answer it.
"I don't have time."
"You don't need any. We run it all. You just turn up to the appointments."
Reframes the service as removing work, not adding it.
Never end a meeting with nothing booked. Can't close today? Book the next meeting before you hang up. And when they say yes, put the onboarding call in the diary within 3 days — 5 at the outside.
Expect to close roughly 1 in 4 early on. He quotes 20–30% in one video and 10–20% in another almost in the same breath — so treat ~20% as normal and don't take the misses personally.
Part 2
The machine
What you build once someone has actually paid you.
The machine
06Five stages, left to right
01
Traffic
Where are the client's buyers already spending attention?
Meta · Google
→
02
Capture
How do we turn attention into a name, phone and email?
Form · Landing page
→
03
Store
Where does every lead live so nothing is lost or double-handled?
GoHighLevel
→
04
Set
Who chases every lead within a minute, all day, every day?
AI setter on WhatsApp
→
05
Build
Once you know the niche cold, what else can you automate?
Custom systems
Stage 5 is where you stop being replaceable. Once your systems are inside the guts of their business, leaving you is expensive. It's also the difference between selling a service business and selling a tech business. Ignore the exact multiples the videos quote — that's sales talk — but the direction is right.
What a working campaign looks like
100,000
Ad impressions
1%
1,000
Leads
50%
500
Appointments
20%
100
Customers
Sanity check. The 50% booking rate is his single best account (1,300 leads · 45% booked · 63% replied · 7% opted out) and he says on camera it is not typical. Plan on 20–30%. Never quote a client the ceiling.
Stage 1
07Pick one traffic source. One.
Do people buy this on impulse, while scrolling?
→ Meta ads default
Local, consumer, cheap first offer. Cheapest leads, weakest intent — you're interrupting people, not answering them.
Do they go and search for it when they need it?
→ Google ads
Urgent, expensive, or business-to-business. Nobody buys a lawyer on impulse. Dearer per lead, far stronger intent. No lead forms — you must build a page.
Does the client already have an unused email list?
→ Email it. Free traffic.
A bonus, not a business. It runs out.
Are you already good at content or cold outreach?
→ Organic. Otherwise don't.
Slow to prove. A bad place to start with no experience.
The profit rule. Every extra service needs another person to run it. Agencies hit RM10k a month, keep RM2k, and wonder why. Do one thing. When a client needs something you don't sell, hand it to another agency — they send work back, and you've built a referral network instead of a payroll.
Budget floors that actually matter: USD 5/day per creative is the bare minimum and you test three creatives, not one. Two of the six creators independently land on USD 25–35/day as the realistic starting budget — that's what local businesses actually commit and it's enough to learn from. Keep the audience above 200k or the ads burn out in weeks.
Stage 2
08Everything is one dial
The cleanest idea in all twelve videos. Every setting in a lead campaign is the same single trade-off. The name for it is friction: how hard you make it to become a lead.
Launch with the fewest hurdles, get 5–10 leads, take the phone calls. Then add friction based on what those calls told you. Start strict and you never learn what you filtered out.
⚠
Never stay on the left
You'll think it's working because the lead count looks great. Then nobody answers the phone. They were tapping mindlessly while scrolling and forgot within seconds.
The five levers, cheapest first
One typed short-answer questionbest value — even something trivial like "which town are you nearest to?". Typing snaps people out of autopilot and into thinking. Huge quality lift, almost no drop-off. Use this always.
"Higher intent" form setting — adds a confirm-and-slide review screen. Bots can't slide it, so bot leads drop to zero. Use this always too.
Phone verification code — kills every fake number. Also costs you real leads, so not on day one.
A qualifying question — filters directly on fit. Add it once the phone calls have told you exactly who's wasting your time.
Website form instead of the Meta form — far fewer leads, far more serious ones. Leaving the platform and still filling it in is real intent. Use when quality, not volume, is the problem.
Ask "is this okay?", not "what's your budget?" — "Our jobs start at RM20,000, is that alright?" filters properly. "What's your budget?" asks someone to price a thing they haven't seen from a company they just met. You'll get a made-up number that means nothing, and you'll scare off people who'd have paid more once they understood the offer. (One creator disagrees — see Part 3.)
The capture ladder — climb it in this order
1 · Meta lead form
Lowest friction, lowest quality. Live in five minutes, cheapest leads you'll ever get. Always your first test. Meta only.
2 · Straight-sell page
Simple offer people already understand. Straight to "book now", with proof on the page.
3 · Lead magnet page
High-ticket or professional niches. Free valuation, calculator, voucher, guide.
4 · VSL funnel
A video that sells and filters at once. Best when you can explain the offer on camera. Most work to build.
If a plain lead form can't get results, a fancier page won't save you. The offer is wrong, not the page. One exception: professional or corporate niches, where you can skip straight to the lead magnet.
Free money on the thank-you page. Never leave it as "You're all set" — that's a full stop. Say "You're almost done" and give the next step: book the consultation, call this number, claim this code, bring a friend free. You already paid for that lead.
Two traps: once a Meta form is published you cannot edit it, only duplicate it. And never set an end date on a campaign — once it completes you can't restart it, you have to rebuild.
Stage 1 · testing
09How to test ads after Andromeda
Meta's newer ranking engine moved where the thinking happens. The old habit — one ad set per audience — now actively hurts you. This is the most current material in the set.
The old way · dead
One ad set per audience segment. Ads written for that segment. You judged the ad set. The ad set was the brain.
→
The new way
The campaign looks inside every ad set and matches one ad to one person. Your creative is now your targeting. Say who you want in the ad and let Meta find them.
So group ad sets by format, not audience
Meta plays favourites between formats. Put an image, a video and a carousel in the same ad set and the video swallows the budget — the image never gets a fair test and you learn nothing. Give each format its own ad set, three ads inside each.
One campaign — the brain
Ad set 1 · Static images ✅
Ad A — text-based, dark
Ad B — text-based, white
Ad C — different headline
Ad set 2 · Talking head
Ad A — hook 1
Ad B — hook 2
Ad C — hook 3
Ad set 3 · Customer-style
Ad A — angle 1
Ad B — angle 2
Ad C — angle 3
Iterate inside the winning box, never across boxes. Statics winning? Your next test is another static — same format, new headline or colour. Never drop a video into the statics ad set.
The kill rule. An ad must spend 1–3× your target cost per lead before you're allowed to judge it. Target RM130 a lead? Killing at RM60 spent tells you nothing. At RM260 spent with a weak click rate and no leads — kill it with confidence.
Scale slowly. Raise the winning ad set $30 → $45, hold 3–4 days, then → $60. Jumping the budget resets the learning and you start over.
Settings, one line each
Detailed interest targeting: blank for local campaigns. The creative does the targeting now.
Untick Meta's "use as a suggestion" boxes on age and location, so your limits are rules and your data means something.
Gender: always "All" — it also catches people who never set one.
Placements: automatic. Restricting to the main feeds costs more, they're premium slots.
Carousels: rarely worth it. The algorithm doesn't like them and they cost more.
Name ads only by what makes them different. Long names are a beginner tell and make the numbers unreadable.
Don't chase rich people. Someone who wants your thing beats someone rich who doesn't. Plenty of wealthy people got there by not spending carelessly.
Stage 4
10The AI setter, built properly
Text beats voice today, and AI beats humans on text — not because it's smarter, but because it never sleeps and never skips a follow-up.
🧩
One sub-account per client
Never let two businesses share a bot's memory. Mixed context is how a bot quotes the wrong prices to the wrong people.
💬
Open with a yes/no question
Low friction. Once someone replies once, they keep replying. Open with a hard qualifying question and they put the phone down.
3️⃣
One to three qualifying questions
No more. More questions mean better leads and more drop-off. Re-ask what the form asked — people lie on forms.
🔔
Two to three auto follow-ups
The single biggest leak in any human team. Setters quietly skip follow-ups and you can't police it. A bot can't skip.
📅
Let the bot read the calendar
Offering real times and handling "no, can Saturday work?" converts far better than sending a booking link.
📚
Feed it the site plus every FAQ
People ask about ten things they didn't opt in for. If the bot can't answer, it loses the thread and the booking.
Use a female sender name — his own split tests say it converts better. Cheap test, free upside. And use WhatsApp, not SMS — higher reply rates, and it's where Malaysian buyers already live. Two of the six creators note their Asia and Middle East clients simply insist on it.
The part everyone gets wrong
11The lead quality engine
⚠️ The trap
Meta gives you the cheapest possible version of whatever you ask for. Ask for leads and it finds people who love filling in forms and never buy. The ads look brilliant. The sales never come. Meta did exactly what you told it to.
Fix 1 — climb the optimisation ladder as budget grows
< $20/day
Leads. No gate. You'll get junk — accept it and climb.
$20–50
Qualified leads. A filter question on the form. Only a yes fires the event.
$50–100
Qualified applications. An application before the booking page.
$100–1,000
Qualified calls. A human approves each booking. See fix 2.
$1,000+
Sales. Only if the sale closes inside 7 days — Meta's attribution window.
Ignore the "50 conversions a week" rule. One or two good events a day beats fifty junk ones. Junk data doesn't train the machine, it poisons it.
Long sales cycle? If people take 30–90 days to buy — webinars, challenges, mini-courses — optimise for qualified leads, not booked calls. Optimising for calls on a slow funnel attracts people who book fast and buy never.
Fix 2 — the human pixel gate
The single best trick in the twelve videos. Instead of firing the conversion the moment someone books, a person approves it first.
Fix 3 — say who you want, out loud, in the ad
Meta transcribes your video, reads your image, and reads your text to decide who should see it. If you never name your buyer, Meta can't help you find them. This matters more after Andromeda, not less — the creative is the targeting now.
Say it 1
In the headline
Say it 2
In the body copy
Say it 3
Spoken out loud in the video
Say it 4
As a yes/no question on the page
The surprise: he expected his cost per lead to jump when he narrowed the call-out to "online coaches making 5K a month or more". It didn't move. Still USD 15–25 a lead — but now they were the right people. Filtering is nearly free. Most people just never try it.
Where results come from
12Only two levers move the number
Lever 1
Speed to lead
Every minute that passes, the lead goes colder. Call inside 15 minutes and you're in a different business to the agency that calls in 3 days. Meta leads are lower intent than search leads, so they decay faster. A human team cannot answer everyone within a minute at 11pm on a Sunday. A bot can, forever.
Lever 2
Lead quality
Junk leads make every number downstream lie to you. Cheap cost per lead, healthy booking rate, zero sales — and you can't see why, because the rot happened at the top. Fix the data you feed Meta and every metric under it becomes honest.
Creative and offer decide whether the ad works. These two decide whether the business works. Most agencies obsess over the first and never touch the second.
Part 3
Judgement
Where six experts contradict each other, and where they're selling you something.
Read before you copy anyone
13Where they flatly disagree
Five real fights. Nobody flags them, because each video is filmed as if it's settled. Each one is a test to run, not a rule to obey.
Should you ask for budget on the form?
Yes — PlattenUses "What's your monthly cleaning budget?" as his example qualifying question.
VS
No — VidellThey haven't met you or seen the value, so the number is meaningless. Ask "our jobs start at X — is that okay?" instead.
My call: the "is that okay?" version wins, and it isn't close. Same filter, no guesswork, and you don't scare off someone who'd have paid more once they understood the offer.
Dynamic creative: on or off?
On — BauerUpload every image and headline, let Meta find the winning combination. Less work.
VS
Off — VidellTurn off all the Advantage+ enhancements. You lose control and can't tell what actually won.
My call: off while you're learning, on once you're scaling something proven. Early on you're buying information, not just leads — and scrambled combinations destroy the information.
One conversion location, or several at once?
Just one — VidellPick instant form or website, so the numbers in the report actually mean something.
VS
Run both — StentonSome convert on a form, others on a site. Give them both. Bauer adds phone calls for local trades.
My call: one while testing, both once you know your baseline. Adding phone calls is genuinely right for local trades — plenty of people would rather ring than type.
Niche down now, or cast wide?
Wide first — PlattenBelow USD 10k a month you don't have the data to choose. Work with anyone who fits, learn who pays.
VS
Narrow nowStandard agency advice — and the same guy's own conclusion, that narrow means higher margins and less stress.
My call: not really a contradiction. Wide is how you gather the data; narrow is what you do with it. The mistake is staying wide once you know.
Interest targeting: blank or not?
Always blank — Videll & DarbyPost-Andromeda your creative does the targeting. Interests just handicap the machine.
VS
Add some — BauerWorth adding interests when you're targeting a whole country.
My call: blank for local radius campaigns. The wider your geography, the more a starting hint helps. Also note Bauer's video is the least current on Andromeda — weigh it accordingly.
Follow the money
14Where they're selling you
🎣
7 of 12 are one man's funnel
Seven Platten videos all end at the same paid mentorship, and the tools he raves about — the CRM, the setter, the voice bot, the scrapers — carry affiliate links. He admits it on the scrapers, to his credit. The frameworks are genuinely good. The tool worship is bought.
🔗
Every creator sells the same CRM
Four of six push GoHighLevel with a discount link; one white-labels it and resells at half price. That's a commission programme, not a consensus. It may still be the right tool — just don't mistake agreement for evidence.
📈
The headline numbers are best cases
"45% booked", "one meeting per 100 emails", "0 to 750k a month" — every one is followed by "not typical" if you listen closely. Plan on the boring middle.
📅
Check the dates
These span about 18 months. "Manual calls beat AI voice" is the oldest claim here and is probably stale. The Andromeda advice is newest and contradicts some older targeting tips. When two clash, the newer one usually wins.
Four questions nobody answers in twelve videos: who owns the ad account when a client leaves · what the client sees in a monthly report · what happens when month one flops and who eats the ad spend · how many accounts one person can run before quality drops. Those four decide whether the business survives its own growth.
Applied to Fastcom
You already own most of this
Not a plan for starting something. A checklist against a machine you're already selling, with a signed price and clients waiting.
✅ Already built
Meta ads, GoHighLevel, landing pages, forms, WhatsApp flows, conversions API, a live brief and a confirmed retainer. Stages 1, 2, 3 and most of 5 — plus the thing these videos treat as the hard part: you already have clients.
⚠️ The real gap
Stage 4. No AI setter answering every lead inside a minute with automatic follow-ups. That work is human right now, so delivery is capped by headcount you don't have.
🎯 Free wins first
The pixel gate and one typed short-answer question on every form. No new tool, no new spend, and both hit lead quality directly.
Where all of this came from
15The 12 sources
Every claim on this page traces back to one of these. Click through when you want the original wording, or when something here disagrees with what you're seeing in an ad account.