Tonight we eat steaks and lobsters

Airport lounge in Denver, last month. 36-year-old founder waiting on the same (elayed) flight. Doing $940K/month on insoles for people who stand all day. Asked him what moved him from $180K/month to $940K/month in 12 months. "I stopped filming the product. I started filming the end of the shift." His old ads were clean studio shots: foam layers, arch diagrams, a podiatrist quote. Flat at $180K/month for 9 months. Then he tested a 74-second ad of a 47-year-old ER nurse in a hospital locker room at 7pm, peeling off her socks. "On my feet since 6:40 this morning. I bought the $400 orthotics. I bought every gel thing in the pharmacy. This is what 12 hours does." It hit $8K/day in 10 days. Film the moment pain peaks.

By far the best event I’ve ever visited. Room full of killers in the Ecom space. Thanks @RomanEcom and crew for bringing Ecom events to the next level. #shenzhen #ecommerce

An ad lives or dies in its first line. This is how we write ours. Every hook passes 4 checks: ideal customer, curiosity, large TAM, something different. Then the AI system: Level 1: basic prompt Level 2: add a swipe file Level 3: paste in an emotion wheel, tell it to get the customer upset, then offer the fix "Write me 10 more like this." Polish in a Google Doc. Never ever ship the first output.

Our full creative stack (in the exact order a concept moves through it): - Perplexity: research - TikTok: find concepts - Trendtrack: competitor ads with data - Sandcastles: outlier scripts and hooks - Poppy AI: analyze videos without downloading - Gemini: creative analysis - Higgsfield: AI hooks and statics - Claude Code + Meta API: creative reports - Notion: organize everything - Claude: put it all to work 10 tools.

Breakfast at a retreat in Lisbon. Met a 44-year-old guy next to me doing $2.1M/month selling OTC hearing aids. Started 4 years ago. "I stopped selling to the guy who can't hear. I started selling to his wife." Banger. For 2 years his team ran ads aimed at the user: audiologist explainers, tech specs, side-by-side price charts. Stuck at $520K/month for 6 months. Then he tested an ad where a 61-year-old woman films her husband on the couch. TV volume bar at 48. It did $15K/day in 2 weeks. Refund rate 6.2%.

Raised the price by $20. Started charging for shipping. Killed the zombie campaign and moved its budget into the main CBO. Every instinct says those 3 moves tank conversion. We're back.

3 changes that fixed our margins without touching a single ad: - Killed free shipping - Raised the single-item price and left bundle prices where they were - Switched 3PLs and saved over $1.50 per order CM2 and CM3 both went up. If you raise the single item price, it makes the bundle the obvious buy.

HIRING: Supply Chain & Inventory Manager 8-figure streetwear brand · multiple suppliers in China Responsibilities: inventory management, purchasing & season preparation, demand forecasting. Remote · USD 2–3k/month base, with growth into multiple brands. 3+ years proven experience required. Apply: forms.gle/UB8qH7aqXvWhBwDq6

"I stopped showing my jeans. I started showing everyone else's." A friend of mine was stuck at $260K/month for 8 months. He was running for 2 years fabric close-ups, stitch counts, and lifestyle shots on job sites. Then he tested an ad where a 39-year-old electrician lays 6 pairs of torn jeans across his truck tailgate. "Same brand my dad wore for 30 years. His lasted 5 years. Mine last 5 months." It killed in the first week. $30K/days. He counted 24 active variations: welders, roofers, framers, each with their own pile of dead jeans.

Everyone around me is running 99% AI UGC ads. I think it's a mistake. We ran the same ad with and without a real creator attached. 16 purchases vs 1. - Real creator/influencer partnership: 16 purchases at €30.33 CPA - Without: 1 purchase at a €54.51 CPA Meta algo moved almost all of the spend to the partnership version  €485 went to it, against €54.51 for the other). Partnership ads are winning across the board for us right now. I mean, you can obviously get really far with UGC AI ads, but Meta still prefers real accounts. That’s also why livestreams are blowing up atm - it’s the ultimate proof of humanity on a feed full of AI. And we see it in our own numbers as well. So we're going all in on real people. My current thesis is: the more AI there is on my feed, the more real accounts pushing products go up in value. And we are betting big on partnership ads.

Founder dinner at a DTC summit in Austin, last year. 38-year-old founder talking about doing $1.9M/month selling custom-fit night guards for people who grind their teeth. Started 3 years ago. He said his growth went from $310K/month to $1.9M/month in 16 months. For 2 years he ran the same playbook as every other dental brand. Dentist endorsements, 3D-scan tech explainers, close-ups of worn enamel etc. He hit a ceiling at $310K/month and stayed stuck for 7 straight months. Then he tested an ad where a 58-year-old woman sits in a dental office parking lot and holds an invoice up to her phone camera for 87 seconds. "This is my 3rd night guard. $785, $810, $840. Same piece of plastic every time. The dentist takes a mold, mails it to a lab, the lab charges him about $100, and he charges me $800. I know because I called the lab myself." 0 product mention until second 64. It hit $11K/day in the first 10 days. He since produced 40+ ads in the same receipt-in-the-car format. I pulled his brand on GetHookd to scan the lineup and counted 27 active variations. Different actors, different bills (one is $1,150 from a prosthodontist), and every single one opens on a number. Refund rate 4.6%. AOV $149, since most buyers take the 2-pack. The customer's own bill is the an amazing ad, probably the best ad you'll ever run.

I’m convinced that huge wins in e-com are downstream from finding battle-tested offers. Here are the numbers from an ad set we ran this week: - €4.69 to get a customer. - €142.56 on their first order. By far the biggest win of the week. We did 4 ad sets, each on a €150 budget. Here are the numbers from Meta: - Ad set 721: €88 spend, 8 purchases, €11.05 CPA, 12.58 ROAS, €138.91 AOV - Ad set 722: €87 spend, 15 purchases, €5.77 CPA, 24.87 ROAS, €143.56 AOV - Ad set 723: €98 spend, 12 purchases, €8.16 CPA, 19.79 ROAS, €161.42 AOV - Ad set 724: €89 spend, 19 purchases, €4.69 CPA, 30.42 ROAS, €142.56 AOV ROAS between 12 and 30. CPA between €4.69 and €11.05. AOV between €138 and €161. Pay attention to the AOV. €150 average order is the main reason why low CPAs turn into high ROAS. With numbers like these, the offer is validated. No questions about it. We've done a lot on the creative side lately too. But a strong offer is what you should search for first. When you hear people saying they are doing offer testing, what they really mean is they’re doing offer hunting. Offer testing means only 1 thing: you keep looking for a new offer that works, constantly. Get the offer right first. Once it hits, creative gets a lot easier. Offer is king.

The maths on hiring a creative strategist that people get wrong. Work out what a strategist has to produce to pay for themselves. If a single winning concept spends 6 figures at target CPA, one hire generating 3 a year has justified a €70k package several times over. People compare the salary to their payroll instead of to the spend it unlocks.

2 things decide what your brand is worth at exit (neither is revenue): Whether the margin holds without you personally managing it, and whether a buyer can rebuild what you have. A brand doing €500k a month on custom product with owned supplier terms is worth more than one doing double that on a product anyone can order tomorrow.

"My product doesn't solve a problem, so this doesn't apply to me." It applies. The problem is just not clinical. True Classic sells t-shirts. The problem is looking soft in a photo. The desire, the targeting and the format all point at that, and none of it requires a medical claim. If you can't name what your buyer is trying to stop feeling, that's the actual gap.

A hook that scales does one of 3 things in the first 2 seconds. Names the exact frustration the buyer felt this morning. Shows a result before explaining anything. Says something that contradicts what they believe about the category. Everything else is a title card.

Test your Q4 offer before Q4. Labour Day and the early fall sale exist so you can find out which offer structure your audience responds to while the CPMs are still cheap. Discovering your Black Friday offer on Black Friday is how brands spend the most expensive traffic of the year on a guess.

Should you do a quiz funnel? Quiz funnels work for one reason that has nothing to do with the quiz. By the time someone answers 6 questions about their own situation, they've spent 5 minutes describing their problem to you. The recommendation at the end lands differently than the same product on a cold PDP. The personalization is real, but the commitment is doing most of the work.

You can generate 200 ad concepts this week. Most operators now can. Volume stopped being the constraint the moment everyone got the same tools. What's scarce is knowing why a specific line makes someone feel something. That doesn't come out of a prompt, and it's the only part of the process that still separates accounts.

Your cancelled subscribers are the cheapest consultants you'll ever hire. Survey every single cancel personally. Not a typeform blast. A direct question with a human reading the answers. Their reasons are your next product iteration, already sorted by priority. It's the last step of the 30-day playbook and the one most operators skip. Comment PLAYBOOK and I'll send you the whole system.