Session 28: Acquire Leads With Paid Ads — Sell the Click, Match the Temperature, Feed Andromeda Variety, and Do the Money Math

Warm outreach (parking lot) -> cold outreach (side streets) -> content (city driving) -> paid ads (highway). Ads amplify whatever offer you have, so they amplify a bad one too: 'the offer answers, what do we sell? Advertising answers, how do we get people to buy.' The assumed prerequisite is the ICP and value-equation work from the monetization session. The economics framed as a casino: 'how can we buy money at a discount?' - $3-5 back per $1 in, his own 8x ROAS across April, May and July - and 'everything boils down into math, efficiency, and us becoming the house.' Cautionary tale: a $2M/year chiropractor spending $30k a month on Facebook with no tracking and no second channel had only 'dipped a toe' in a $15.1B industry.

'The job of an ad is not to sell the product, the job of an ad is to sell the click.' Relay race: each runner runs one leg and passes the baton; one runner trying all four collapses the system. Ad sells the click -> landing page sells the opt-in (contact details plus motivations) -> book-a-call -> thank-you page -> sales call closes (B2B/service model). A VSL on the landing page pre-sells; lead nurture has two paths after opt-in. The failure mode is any ad that asks for the booking or the sale directly from cold traffic.

Cold = strangers scrolling Facebook/Instagram: no pricing, no booking CTA - educational content, a hook, or a lead magnet to a landing page. Warm = they visited the page or watched the video: a specific offer with proof, and 'do not break the scent trail' - the offer in the ad must be the offer on the call. Hot = email list, referrals, past clients, content consumers: frictionless, direct - 'book a call with us.' This is 'the textbook approach, especially true for B2C'; B2B buyers are 'in problem-solving mode' and will respond cold to a specific enough offer. His B2B static ad - 'Looking for 5 AI agencies, we'll build your clients' projects in 21 days flat, or you don't pay and I give you $5,000' - produced booked appointments at $64 on 6k front-end packages; anatomy: specific call-out (cocktail-party effect), clear offer and timeframe, bold guarantee, scarcity (5 spots), proof (12+ case studies). Rule: B2C always temperature-matches; B2B may go editorial or direct.

Three parts to every ad on every platform. Call-out: the cocktail-party effect - 'Attention, gym owners in Dallas', 'HVAC companies with 5 to 15 trucks', 'Property managers drowning in tenant calls.' Value: the value equation - dream outcome x perceived likelihood, divided by time delay x effort and sacrifice; worked example 'we helped CloudTech go from 50 to 200 qualified leads per month in 30 days' maps each term (200 leads, case study, 30 days, 'we did it for them'). CTA: 'Learn More' beats every other button on cold Facebook/Instagram because it is the least threatening; 'Book a call' and 'Buy now' scare cold prospects.

From Sabri Suby's 'Sell Like Crazy', adapted to B2B: brains see 10,000+ ads a day and filter them; an ad dressed as a magazine article slips through. Eugene Schwartz: 'the copywriter's task is not to create mass desire, but to channel it.' Four headline archetypes: personal journey ('$0 to $833,000 per month in 4 years'), industry expose, transformation ('lost 77 kilograms eating ice cream'), outcome-in-time ('first profitable investment property in 60 days'). What breaks the frame: logos, prices, dates, phone numbers, 'book a call / request a quote / get a demo', stock photos, heavy branding. Seven primal desires - curiosity/intrigue, shock, direct benefit, implied benefit, fear, vanity, self-interest - hit at least one, two is better (divorce ad = curiosity + fear; weight-loss = shock + implied benefit). Trigger words that 'double or triple click-through': exposed, confession, revealed, horror, secret, alarming, shocking, hidden, banned, forbidden, unbelievable, devastating, outrageous, finally, urgent, warning. Use editorial about 3 in 10 times.

Sculpting: (1) start boring - '7 things to know about financial planning'; (2) name the outcome, not the topic - '7 things you must know before hiring a financial planner'; (3) add a trigger word - '7 alarming things...'; (4) add a parenthetical open loop - '(number 3 will blow your mind)'; (5) write 20 variations, test, then build a 'Franken headline' from the winners. Model headlines on magazine covers (AARP, Google News), not on other ads. Fascination templates: 'How to [outcome] without [pain]'; 'The [adjective] secret to [result] that [authority] doesn't want you to know'; 'Why [common advice] is actually destroying your [valued thing]'; '[Number] alarming signs your [thing] is failing'; 'What [famous entity] knows about [topic] that you don't'; 'The [N]-minute trick to [result]'; 'Warning: [behaviour] may be costing you [loss]'; 'Finally revealed: the method behind [result]'. Usable as headlines, subject lines, landing bullets, video hooks, LinkedIn openers.

Static images are cheap, fast and consistent ('right around the middle'); video is volatile - a bad one 'will absolutely bomb', a good one 'wildly outperform'. Lower CPM on statics means more tests per dollar. Workflow: 15-20 static iterations to find the messaging and angle, then turn the winners into video to scale. His 8x ROAS came from statics alone. Three static architectures: guarantee-led, situation call-out, contrarian hot-take - his best performer is a screenshot of his own tweet: 'the fastest path to AI revenue right now isn't building anything, it's the AI middleman model.'

A hook = call-out ('this is for me') + value promise. Types, from a ~$7,000 Hormozi playbook: labels ('Dental clinic owners, I need to tell you something'), yes-questions, if-then conditionals, shocking statements, open curiosity loops, story openings, contrarian statements, nonverbal (odd setting, movement, text overlay). 'Labels and statements are the most performative' - ~80% statements, test questions and if-thens. Script that 'works for 80% of video ads': 0-3 s hook (label + shock/curiosity), 3-10 s problem/agitation, 10-25 s solution/mechanism, then proof (numbers, names, results), then one low-commitment CTA.

Intent-based (Google): highest intent and highest cost per lead; captures existing demand, cannot create it; right for bottom-of-funnel local services - plumber, lawyer, dentist. Interruption-based (Meta, LinkedIn, TikTok): interrupts entertainment to create desire 'that didn't exist 5 seconds ago.' LinkedIn: the only platform with exact job-title targeting, for $10k+ B2B or $50k+ enterprise ABM, at '$8-12 per click, not per lead.' Meta: 3.3 billion users, the cheapest leads of any major platform ($1-5 typical), the most advanced targeting AI - 'the final boss' - and works for B2C, B2B, local, e-com, agencies, coaching, SaaS. Q&A adds: YouTube is hard without a channel; X untested; Reddit doubles as an organic route into LLM answers; OpenAI's new ad product has no data yet but low competition.

Old Meta was a paper map - you guessed the audience with stacked interests and one ad. Andromeda is GPS with live traffic: go broad, it finds the buyer, but it needs many roads to test. 'Conceptually different' means different person, background, setting AND a different angle - not the same video with a new first line; the spy analogy is that look-alike creatives get clustered and die together. Minimum 8-12 conceptually different creatives; cited test: one ad set with 20 creatives beat five ad sets of five by 17% more conversions at 16% lower cost. Creative fatigue sets in after 2-4 weeks even for winners. Meta reads five signal types: visual (faces, colours, motion), language, audio (tone, pacing), engagement (stop/click/watch/bounce), conversion.

'Qualium' (quality x volume, attributed to Hormozi) is 'the single most important strategic concept in Meta paid ads in 2026.' Wrong way: one ad for an average customer, $5k, hope. Right way: one creative per psychographic segment - his own: sales professionals entering AI, agency owners turning away 6-7 figure deals, young operators starting out. Structure on a low budget: Ad Set 1 = proven winners at ~70% of budget, Ad Set 2 = new concepts at ~30%; a new creative in the winners' set gets 'starved of spend', so tests need their own set.

Bad friction is a locked door with no handle - slow pages, broken links, 100-question forms, tiny mobile text. Good friction is the bouncer: wrong people leave, right people feel exclusivity. Tools: a VSL on the landing page (filters tyre-kickers, pre-sells), a 3-5 question qualifier (revenue, employees, biggest challenge), an application-style form ('Apply for a partnership' instead of 'book a free call'). It is a dial: zero friction = many cheap, often unqualified leads; maximum = fewer, better, higher show and close rates. Story: a Meta ban (VPN-triggered, not content) forced a fresh account and pixel; the same low-friction forms that had delivered 8x ROAS on a trained pixel produced 4-5 no-shows out of 6; adding friction fixed it. Start low to gather data, add friction when quality drops; B2B services should always carry a VSL.

facebook.com Ads Library lists every ad a brand is running; sort by longest-running or most impressions - 'if they're still spending money, it's making money'; ignore fresh ads, they have proved nothing. Per winning ad log format (image/video/carousel), hook, who is called out, CTA, and where the click lands (page, magnet, video) - 'we're not copying their ad, we're reverse-engineering their strategy.' Demo: Hormozi's ads and an HVAC competitor. Exercise: find ads running 30+ days in your niche, screenshot, log the four fields. Creative-effort rule 70/20/10: 70% proven hooks (own data, competitor research, organic winners), 20% variations on winners, 10% wild experiments.

Phase 1 track: pixel, conversions, parameters before spending, or 'you're gonna get cleaned out.' Phase 2 lose: a fresh pixel does not know who to target; the losses are college tuition. Phase 3 scale: spend as much as possible on the winner 'until it doesn't work anymore.' Metrics: CPM, CPC, cost per lead / per booked appointment, CPA, LTV:CAC at least 3:1. Worked: $500 -> 50,000 impressions -> 500 clicks ($1 CPC) -> 25 leads ($20) -> 5 calls ($100) -> 1 client at $5,000 ($500 CPA) = 10x; $5,000 spend -> 10 clients -> $50,000, 'same system, different scale.' Client-financed acquisition: $10,000 deal, $500 ad cost, $3,000 fulfilment = $6,500 profit plus $500/month x 12 = $6,000 recurring - 'every customer funds the acquisition of the next.' His benchmark: a good ad books an appointment for $150 or less. Homework: feed the transcript and slides into Claude rather than prompting generically.

Unnamed in speech (on-screen text reads 'Techify'): an offer-creator agent and an ads-writer agent. He uploads his 'Partner Launch' offer document, runs the ad sprint - competitor research through the Ads Library (visual plus transcript analysis of their ads), ad teardown, ad-copy generation, then image-ad headlines with a suggestion to render them in Nano Banana. Hook taxonomy from the Hormozi playbook is baked in. A 'generous free tier' exists and he promises to extend limits for the cohort. Q&A closes with platform comparisons, qualifying competitor winners, tracking ROI, and his LinkedIn/Instagram.