# Jacob's X Signal

All the signal from two years of @jforjacob posts in one knowledge base. 230 lessons on ads, creative strategy, and building ecom brands.

By Jacob (https://x.com/jforjacob).

## Chapter 01: Business Strategy

### Product Beats Everything

- **A lot of the time it's not a skill issue, it's a business model issue.** He sat stuck under 8 figures with the exact same skill set, processes, and team he has now: the only thing that changed was the product. If your systems have worked before and you're stagnant now, the fix may be a new product, not more optimization of the current one.

- **The sauce is a six-item checklist and it hasn't changed:** know your customer, write ads that speak to them, produce ads in formats native to platform, have a cold-friendly offer, optimise your business for LTV, and scale ads against LTV, not day-one revenue. Every hyped new tool gets filtered through one question: does it make you better in any of these areas? If not, don't bother. He abdicated from most of the noise and is doing better than ever sticking with the basics.

  The cost of operating is lower than ever and speed to ship is faster than ever, yet the timeline's revenue is flat. Maybe the bottleneck was never speed or cost. Maybe it's always been quality, and everything you're focused on right now does not increase quality.

- **The ecom importance ranking, in order: product and category, offer, retention, CRO and funnel architecture, then creative.** Creative ranks last, from an operator known for creative. Great ads can't save a weak product, offer, or funnel; fix the upstream layers before blaming your ads.

- **Past a certain scale the bottleneck leaves the ad account.** On a call with someone doing much bigger numbers, the advice for getting to 8 figures a month never touched creatives or the ad account: it was all logistics and supply chain. The ecom timeline obsesses over ads while the actual growth lever at scale is ops.

- **Creative is not everything.** The best ads in the world still die against a shit offer, shit margins, a shit landing page, shit supply chain, shit distribution, a shit product, shit product images, or a badly run ad account. Product comes first and is by far the most important thing: complete newbs have hit $100k a day because they landed on an absolute banger, not knowing wtf they were doing with creative. Offer and funnel come next, then creative (maybe). The majority of the biggest scalers he knows run mediocre ads at best.

- **Product is everything; the technical stuff is secondary.** The same operators doing millions per month likely have other brands, or failed brands, running the exact same processes, systems, and media buying. Ask them why those didn't hit the same levels. Everyone loves to talk media buying and systems, but it's all downstream of the product.

### Lean Economics Win

- **One question decides every service partner: do they make me more money than I pay them?** Yes: keep paying. No: stop. It doesn't need to get more complicated than that. People carry crazy, unrealistic expectations for agencies; if the ROI is positive, even small, keep the train rolling. Firing a profitable partner over vibes is stupid.

- **Revenue projections don't change behavior, so skip them.** At a mid-8-figure run rate he has never done one: below target, you keep pushing; above target, you're not pulling back, you keep pushing. The exercise adds nothing. Buy the inventory you can actually afford and let that dictate growth, rather than taking on debt for imaginary numbers and getting caught with your pants down.

- **Most people think they have a creative problem when they really have a business model problem.** Run the thought experiment: if your target ROAS on Meta were 1x or even 0.5x instead of 2x, how many ads you called losers become winners, how much does your hit rate instantly jump, how much more could you spend? Same ads, transformed results. If your economics only work at 2x, the constraint is offer, AOV, and LTV, not "we need better ads."

- **OPEX inflation is the biggest killer of businesses.** The goal is maximum revenue with as few people as possible; boasting about team size is a sign of naivety, the business equivalent of the guy who hits his first $100k profit and signs a penthouse lease. Looks like winning from the outside, actually shackles. Lifestyle inflation takes people down personally, headcount and overhead creep takes down companies: optimize revenue per employee, not employees.

- **There has never been a time when a one-man team could so genuinely compete with big players.** With an internet connection, AI, and some freelancers, anything gets done at lightning speed to high standards. Most big brands are bloated, overpay for everything, move slow, run pathetic margins, and survive on the hope of being bought, yet people model themselves on exactly these brands. Talk instead to the 19-year-old doing $50k a day from their parents' house with no employees: legacy brand operators are mind-blown when shown behind that curtain. Stay lean, stay profitable, stop romanticising the big bloated company. The opportunity lies in doing the opposite.

- **Consumers don't give a shit about your branding: they want stuff and they want it cheap.** Chinese discounters took massive market share despite the huge effort Americans put into branding and "protecting the brand" by refusing to discount; we live in a disposable, discount economy. Bloated brands leave massive gaps underneath themselves: Temu may pay some VAT on imports, but they're not paying $85k for a content shoot, and margin innovation is happening that these brands aren't prepared for. Lean operators out-manoeuvre regardless of tariffs: while an American brand pays $75k to test one product, a western operator with a dropshipping or affiliate background could get 150 brands off the ground with that money. Cutting Section 321 buys air, but the waves keep coming.

  _(posted September 2024 · practice might be outdated)_

### Moats, Copycats, and Crowded Markets

- **"Those who can afford to spend the most to acquire a customer win" is true, but bootstrapped brands don't have that luxury, especially early.** The compensating edge has to be creatives and landing pages so much better than the big boys'. The opening is real: most supplement brands at scale wouldn't believe his CPA numbers are possible, and the majority of category leaders' ads (Gruns, IM8 etc) are objectively mid. Ultra-low CPAs were the scaling phase, not the end state; he's shooting for higher CPAs now, but out-creating the incumbents is what it took to scale so fast from nothing.

- **The new SaaS moat against vibe coding is AI-fueled value density.** Use AI to pack more features into the same subscription, or cut the price: make the bang-for-buck so ridiculous that duplicating the app is irrational. As an operator, take the other side of that trade: when an app's ROI is that lopsided, let the builders build and pay the fee instead of burning your own money, time, and resources vibe coding a clone. You get a friendly UI and support thrown in.

- **The only SaaS that survives vibe coding offers a ridiculously high cost-to-value ratio.** The survival test: is duping it worth even 2 hours of your time? Apps like JudgeMe and Commslayer clear it - you could spend 100 hours building a working JudgeMe dupe, or just pay the $12 a month. Everything a merchant can vibe-code into a workable replacement in an afternoon is toast. Same lens works in reverse for picking which tools to commit to as an operator.

- **Dropshippers hide their brand because they have no moat, not because they're lying.** No unique product, no returning customers, no distribution, no brand: the product, angle, and creative they're running is the entire business, and one whiff means 50 clones the next day with margins instantly shrinking under 1:1 copies. A first-order-profitable business dies by disclosure, same as an affiliate doxing a working offer: it's just suicide. The 2018 era of $5 CPMs and any shitty creative working is gone; modern dropshipping is ruthlessly competitive, which is why the ones who crack it become some of the best marketers and operators in the space. If you have something working with no moat, protect it ruthlessly.

- **In a "saturated" market the education is already done, so stop making generic problem/solution ads.** All your effort should go to differentiation, unique mechanisms, and why you're the better choice, not educating people on a problem they're already well aware of. The easiest move when starting out and struggling: go full bottom-of-funnel across your ads, product images, and landing page copy, and be a little feeder fish sucking up sales from the big players who paid for all the expensive education marketing. Only once you have some scale should you worry about getting sophisticated and moving up awareness levels.

- **When Meta generates the perfect ad for the perfect person, creative stops being an edge.** The prediction: creative assets become like catalog ads. Meta knows what you sell and what content style each user consumes, so it dynamically generates matching creative and you mostly won't need to make your own. A perfectly-served ad everyone has access to is no edge at all, so invest now in what AI can't equalise: distribution, unit economics, supplier relationships, CRO, retention, brand, and community.

  _(posted May 2025 · practice might be outdated)_

- **When something new gets traction, get spend as high as possible as fast as possible, even at the cost of efficiency.** Countless people ripped his new brand 1-on-1: same markets, same creatives, same offer and website. None lasted more than a week, because outbidding them beat sending DMCAs; his budgets were 1000x+ anyone coming in to test the product. Deliberately sacrificing margin means you're pushing the edge of what the current creatives and offer can achieve, leaving no room for anyone to get efficient spend next to you.

  Volume compounds into structural advantage: sheer order volume negotiates better cogs and shipping, so new entrants pay 4-5x what you do and would have to run at a loss at high spend for a long period to take share. Fuck your efficiency in the beginning. Get volume for the best deals, best supplier terms, and favour in the auctions; dial in margins after the copycats give up and you've built a large social following.

### Run a Brand Portfolio

- **With enough diverse creative in the account, a brand can run indefinitely untouched.** One of Jacob's brands does $4-5k/day and he hasn't touched it since March 2025: not a single ad uploaded, email sent, or social post, just the existing creatives running on Meta. Most brands can reach that creative-diversity threshold, and at that point it's literally passive income. He's considering making it a deliberate strategy: 10-15 of these spitting out profit daily while you do nothing.

- **Owning 100% of several brands out-earns a fraction of one famous one.** The highest-revenue operator he knows runs multiple brands and owns all of them, banking more personally than the CEOs of the 9-figure brands people cream over, even at half the revenue. Build the right team and leverage it as an agency for your own brands: no agency runs with one client, why should your team run one brand?

  Wins compound across the portfolio: unlock something on one brand, roll it out across several. Risk diversifies: one brand gets sued, loses a processor, or shits the bed, and the others keep running. It's accepted that the S&P beats picking a single stock, yet people take the opposite approach to their own business assets. And you can fill your own market gaps: if one brand is the value option, spin up a premium version with the supply chain, ops, team, and market insights already in place.

- **The highest-leverage thing in ecom is a reliable team, including trusted agency partners, and then running that team across multiple brands like an agency runs clients.** Crack the formula with one brand and the same team ports across several with minimal additions, just extra customer-service VAs when needed: he was running 3 brands on that infrastructure with a 2025 goal of 7. Laser focus on one brand is a myth past a certain point: instead of hunting the 47th angle to unlock new audiences on a pushed-hard brand, it's much easier to get an extra £200-£300k per month starting something new. Branded fulfilment from China takes enough headache out to make the multi-brand approach far easier.

## Chapter 02: Ads & Creative Strategy

### Better Ads Solve Everything

- **A static built on one sound psychological principle took 3 minutes in Canva, then spent 7 figures.** The principle came from a single @SarahLevinger tweet. The ad got copied one to one by almost every DTC supplement brand, which is itself proof that principle-driven statics are the highest-leverage creative format. Production value isn't the moat; the underlying principle is.

- **A "low volume" brand making 50 creatives a month sits in the top 1% of advertisers by active ads, with 1,000+ active.** Good ads stay running for a long time and compound into a durable active-ad base. Volume-chasers publish hundreds at a time and kill 99%, so their output never accumulates. Ironically, the best way to get volume in your account may be to reduce output: make sure everything you test is meaningful and likely to stay active. Active-ad count is the metric, not ads shipped.

- **Take a full week off from publishing ads and study instead: top TikTok Shop affiliate videos plus direct response copywriting.** Then re-analyse your current ads, laugh at how terrible they are, and remake them. The bar is shockingly low: the ad libraries of 99% of brands and almost every creative agency show nobody knows how to make a good video, and even 9-figure brands barely get past "Do you have x problem? Here's x product that solves it. I swear it's a game changer." A few days of deep diving puts you in the top 1%. Even one day beats the majority.

- **Master creative quality first, then build volume on top.** Nobody learns to juggle starting with 8 balls: you start with 2 and work up, or you never learn at all. Starting a creative system with volume means too much going on, no foundations, no progress. Start with quality, learn the foundations, build the systems around that, then scale output. From someone directly involved in launching more brands than 99.9% of people in ecom, this is the single most important takeaway.

- **90% of problems in ecom are solved by making better ads.** People pour ungodly time and money into things that move the needle 1-2% at best: attribution deep dives, popup debates, "upper funnel events". Drop a couple of bangers into the ad account and you can double the business in a week. The tell: people talking about needlessly complex things usually have objectively terrible ad libraries, big brands included. Check the library before believing the talk.

- **Ignore "the volume era": deep dive every new concept, prove it, then template it for volume.** One month of 21 concepts produced 5 winners; the next month, 9 concepts produced 6 winners because they reused the proven frameworks. The volume-first crowd rushes mid work out the door, maybe hits a lucky winner, has no idea why it worked, can't repeat it, and wonders why they can't scale.

  The 10-step workflow per concept: avatar (age, condition, frustration), core desire (what they want to feel or change), awareness stage (what they already believe), belief shift (what they must adopt to buy), ad angle (frame or narrative), hook or big idea (what grabs attention emotionally), narrative flow (pain to belief to proof to CTA), creative output (format and asset), creative direction (location, scene, specific shooting instructions), destination (existing page or new lander). Track everything in a master Notion database with columns for avatar, desire, awareness stage, format, then chart hit rate, ads launched per month, and hit rate by awareness stage. Link an Atria report per test and tag ads there to build custom reports like spend by awareness stage. _(posted August 2025 · practice might be outdated)_

- **Find a trending organic format, apply it to your brand, sprinkle in a little direct response.** It really isn't more complicated than that. Trending formats are pre-validated attention: the platform is already rewarding them, so don't invent from scratch. Keep the DR layer light, an offer, a CTA, some problem framing. Heavy-handed direct response kills the native feel that made the format work.

- **The top spending ad of all time in the account is a shitty little Canva static made in 5 minutes by a marketing newbie.** No messaging beyond displaying an offer, still running today, and never topped by any of the brilliant ads made by very talented people since. Don't assume production value or cleverness predicts spend: test the dead-simple offer-display version before dismissing it.

### Know Your Buyer: Research, Awareness, Desire

- **When one engagement-heavy ad hogs spend without hitting CPA, don't kill it: build the follow-up touch points.** That ad is great at pulling cold people into your brand; by default they won't convert at target. Analyse the top ad and landing page (use AI if you wish), assume it's all a viewer has seen of your brand, because it likely is, and list every question or concern on that exact journey. Then make hyper-specific ads answering each one. For "omg I'm fitting into my high school clothes after 60 day supplement challenge" that means a full day of eating during the challenge, the workout routine, candid talk about side effects and libido, comment-reply videos, before-and-afters. Trust Meta to sequence them after some optimisation. Do this correctly and you print money.

- **Don't research your audience by searching "x problem" on Reddit: live like you have the problem.** Reddit fails twice, most content is fake and you only hear the karens, so you end up marketing to whatever the most insufferable slice of the population complains about. Do the "method acting" of ecom instead: pretend you have the problem and seek answers the way you would in real life. Getting shredded didn't mean one search, it meant YouTube, Instagram, competing workout philosophies (5x5 vs high volume, 3 vs 7 days, keto vs carnivore vs IIFYM, protein targets), channels, blogs, groups, gym conversations, each question opening another rabbit hole. That depth unlocks segment-specific ads: telling the 5x5 bros they need more volume, re-educating the keto crowd on low-carb thyroid effects, showing the 7-days-a-week crowd that rest gets them there faster. Most ecom guys go one level deep and think it's enough.

- **Don't take media buying or creative advice from big supplement and subscription brands: 90% scale because of how they rebill, not skill.** Some rebill customers at double the initial purchase price and scale at around 0.3 ROAS; even a basic marketer could run multiple 5-6 figures a day in spend on that model, and if they rebilled at the sign-up price the brand would die. For creative inspiration study brands scaling despite a worse business model: front-end-reliant operators who must truly understand and engage their audience, pulling super cheap CPCs and high CVRs. Relying on the first purchase isn't a smart position, but it forges the best marketers.

- **Label every ad in your account with its awareness stage in the title.** Build custom reports in Atria or any reporting tool filtered by those labels, track spend and returns per stage, double down on the biggest one, and fill gaps in stages with high returns but low volume. Most people are flying completely blind on where their money and ad resources go; knowing is the precondition to scaling deliberately.

- **Don't assume audiences get bigger as you move up the awareness stages.** For backpacks the unaware segment is much, much smaller than the product-aware one; targeting unaware or even problem-aware there wastes your time, and almost all effort should go into product-aware marketing: aesthetics, features, materials, pricing differences. It's not a linear scale. For some products the biggest pool is product-aware, for others problem-aware, for others unaware. Analyse market sophistication first, then allocate.

- **Everyone iterates hooks, formats, and creators; nobody iterates the awareness stage.** Take a winner, put it into Google Gemini, ask what awareness stage it would classify it as, then ask how to change it to go up a stage. Can work wonders sometimes.

- **Spend 10 minutes scrolling your feed before briefing any ad.** You'll be amazed how often the idea in your head doesn't look or feel native to the platform at all. Especially true for statics, where designed-looking concepts stick out worst against organic content.

- **Every awareness stage has a current mindset and one specific belief shift your marketing must create.** Unaware: "everything feels normal" must become "there's a hidden problem affecting my life worth paying attention to". Problem-aware: "something's wrong but I don't know how to fix it" becomes "there's a specific type of solution for this exact problem". Solution-aware: "I don't know what actually works" becomes "most options aren't effective, but this one works because of a unique mechanism". Product-aware: "I'm not fully convinced" becomes "this is the best choice: proven, trustworthy, real value". Most aware: "should I buy now?" becomes "now is the right time or I miss out". Map messaging to the shift, not just the stage.

- **Desire can't be created, only channelled toward your product.** "The desire to look good in front of my ex" is not a desire, it's a summary of your ad. Know exactly which core human desire every ad targets, picked from the actual list: to survive, to avoid pain, to be free, to feel safe, to protect what matters, to preserve the self and body, to perform at one's peak, to be in control, to belong, to live with integrity, to live with meaning, to experience joy or vitality. Knowing the core drivers gives clarity and direction to your entire creative strategy.

- **Persona > desire > market sophistication > awareness level > angle > format.** The complete creative strategy stack: "I literally don't think you can fail if you follow this" with business fundamentals in place.

  Start with one persona, the group with the biggest TAM and opportunity from your research, and define the dream outcome they actually want. Then figure out market sophistication: no point going after an unaware audience if you sell backpacks, everyone already knows the problem, so you highlight unique mechanisms, design, functionality, aesthetic. Target awareness stages up to that ceiling: a primarily problem-aware market means designing for problem-aware, solution-aware, product-aware and fully-aware; a product-aware market means focusing mostly on product-aware and fully-aware. For each stage list as many angles as you can, rank by priority, work top down with a checklist. For each angle list every format it can be communicated in and test the same way. Track every ad by awareness stage, angle and format; double down on what works, discard what doesn't.

- **Formats come from what's working organically; angles come from mining real customer language.** Deep-dive competitor reviews on Amazon and Trustpilot, Reddit forums (use Gigabrain to search them), and AnswerThePublic. Watch what influencers in the niche talk about and the questions their followers ask them: those are unanswered objections and desires. _(posted February 2025 · practice might be outdated)_

- **Big brands can afford terrible front-end metrics; you can't, so stop filling your swipe file with their ads.** Inside those ad accounts the front end usually sits below break-even with poor soft metrics too. They survive on Amazon, retail, big organic presence and high LTVs: you are not playing the same game. Study the unknowns running large amounts of ads profitably with no fame and no other channels. The best acquisition accounts belong to dropshippers and hybrid brands run by ex-dropshippers, fully reliant on front-end profitability at eye-wateringly high daily spends, where a break-even CPA just doesn't cut it.

- **Ignore the gimmicky ad concepts "creative experts" hype: Apple Notes checklists, iMessage convo ads, us-vs-them table statics, fake Google image search results.** Go through the ad library of any brand doing well and you very rarely see these, and when you do they didn't run long. The ad library is the truth filter: before copying a hyped concept, check whether scaling brands actually run it and for how long. Longevity of spend is the signal, not Twitter applause.

### Hooks and the Promise of a Payoff

- **Showing the product early or throughout a UGC video is one of the worst things you can do for ad performance.** The irony of the AI UGC race: everyone burns effort getting AI creators to show and hold the product perfectly while the early reveal itself hurts results. Hold attention with the hook and story first; reveal the product later.

- **The complete direct-response skeleton: hook -> authority and credibility -> problem and root cause -> agitate with failed solutions -> validate skepticism -> solution and unique mechanism -> testimonial or social proof -> soft CTA and risk reversal.** The two beats most scripts skip are validating the viewer's skepticism after listing failed solutions and closing soft with risk reversal; those separate it from a generic PAS template.

- **A well thought out text overlay hook implies something is about to go down, and that promise of a payoff dramatically lifts thumb stop and retention.** Zero interest in go-karts, yet a full one-minute watch just to see why "he should get banned"; without the overlay there's no incentive to stop or stay. And yes, test multiple hooks per video: the recent dogma to make every asset substantially different is some of the worst advice in DTC, and text overlays are extremely effective and easy to test.

- **The formula for a truly unaware ad: hook with an interesting story completely unrelated to your product, tie it to the problem you solve, then CTA or keep walking down the awareness stages.** In practice: "In 1946 there was a really interesting study carried out on German soldiers after WW2", two groups, one recovered, one spiralled into addiction, "the thing that differentiated these groups will amaze you". Reveal the differentiator (one group supplemented with magnesium or similar), educate on how and why it works, knock down the alternatives, then introduce your product as the mechanism. Exit to an education landing page with testimonials, or continue the video and bring people further down the awareness stages in-stream.

- **In video retention your hook does 90% of the work, and the mechanism is a promise of a payoff.** A video whose hook implies a girl is about to put on an outfit can hold near-100% watch rate while she rambles about stuff nobody cares about; viewers aren't listening, they're waiting, and the payoff doesn't even need to arrive for the hold to work. Your ad shouldn't be filled with nonsense after the hook, but a hook with a promise of a payoff will do more for your engagement and hold rate than anything else ever could.

- **The "Delayed Payoff & Authority Stack": open a loop by naming your reasons for success without explaining them, stack proof, teach all but the last point, then weld the final secret to your CTA.**

  Hook with a before/after case study and a dramatic claim ("37x more views"), qualify the audience in sentence one ("if you're a real estate agent struggling with cold leads"), then list the "3 reasons" by proprietary-sounding names only ("Influencer clause", "Shorts opt", "Expert-observation continuum"). Delay the payoff with a credibility blitz: rapid-fire client results with hard metrics plus one authority-stack humble brag ("we even helped [Famous Brand]"), aggravating the gap between the viewer's struggle and the proof on screen. Buy trust with a partial payoff: explain the first reasons in genuine, actionable detail, debunk a myth, hand over a real tactic, proving you're an expert rather than a salesman. Withhold the final "most valuable" reason as too complex for the ad and put it in a free training: a low-commitment bridge CTA that feels like the natural next step and filters casual viewers. Then double up with a shortcut CTA for ready buyers: "if you want to skip the learning curve and have my team do this for you, book a free consultation." Works for any service business, consultant, coach, or brand selling a solution to a complex problem.

- **[Hook] -> [Credibility] -> [Problem] -> [Tease possible solution] -> [Why other solutions don't work] -> [Why this one's different] -> [Testimonial] -> [Soft CTA].** The middle beats do the persuasion heavy lifting: teasing the solution, dismantling alternatives, and differentiating your mechanism is the sequence that moves solution-aware buyers, before proof and a soft close.

- **Starting an ad with a question hook like "Tired of stubborn belly fat?" is lazy, uncreative, and about the worst way to open a video.** It immediately feels like a sell; you'd never start a real conversation like that. Go straight into a story that opens a mental loop with a promise of a payoff: "I'm not proud of this but I'm gonna tell you anyway", then the personal story (telling her boyfriend she had a phobia of water to avoid wearing swimwear because she hated her body) before the product appears as the turning point. The feed is wall-to-wall "Is X slowing you down?"; the bar is so low that a natural story hook stands out instantly.

- **Everyone pours effort into scripting while ignoring what the viewer actually sees.** The platforms are hugely visual: staging, direction, scene and environment get almost no attention. No-context videos of somebody randomly assembling, crafting or cooking something, no talking, no captions, suck viewers in for 5+ minutes trying to figure out wtf it is; open with "watch how I make a DIY pet feeder" and they skip. Plant a mistake, controversy, or something viewers "weren't meant to see" and engagement goes insane: people jump into the comments to correct you and argue with each other. Massive hook, retention and engagement rates mean lower CPMs, more eyeballs and more sales for less money. Meta rewards content that engages and retains.

- **Plant deliberate mistakes or controversies in your ads and watch engagement explode.** State a fact that's objectively false, or make the coffee by putting the water in first: people flood the comments to correct you, argue with each other, and joke around. One ad racked up around 8,000 likes and about a thousand comments across placements in two days, and that engagement is what produces ridiculously low CPMs, not targeting third-world countries. Meta is a retention and engagement platform; align your ads with what keeps users on it and you get rewarded. _(posted November 2024 · practice might be outdated)_

- **Your comments section is part of your funnel; moderate and optimise it like the ad itself.** On one placement alone, the top comment on a two-day-old ad had 173 likes while the video itself had 430; imagine the impressions that single comment pulled, never mind the rest. People are in your comments section A LOT. Leaving it unmanaged while you obsess over the ad and landing page is fucking up.

### Statics and Formats That Scale

- **The top spending static in the entire account, 6 figures in weeks, came from a @SarahLevinger copy prompt built on the Von Restorff effect plus future pacing.** The isolation effect says whatever doesn't look, sound, or feel like everything else in the feed wins attention.

  The 3-step hook structure: open with weird, bold, oddly specific phrasing ("There's a man in Finland who hasn't paid for skincare in 10 years, and his skin looks like marble"); future-pace them into a vivid, identity-altering outcome ("Three weeks from now, you won't remember what foot pain even felt like"); land a sharp, slightly uncomfortable insight that explains why their current results suck ("You're not being ignored. You're just blending in"). Operationalise it by feeding the prompt your angle and asking for 3 static ad concepts. This is for the copy, not image generation.

- **Your best emails are your next best ads.** Look at the last 30 days of email campaigns, take the top revenue emails, run the above-the-fold headline and graphic as a Meta static, and turn the remainder of the email into the landing page. Messaging and format are already proven against your own buyers, and ad-to-lander congruency comes free because they're literally the same asset.

- **Face to camera, no b-roll, no editing beyond trimming, TikTok-style yapping is still the consistent top performer.** Highest spend, hook rates, hold rates, average watch time, CTRs and CPA; polish isn't what carries performance. The catch: you need to understand how to actually write a good script and direct well, which is a big but it seems. The barrier is writing skill, not production budget. _(posted October 2025 · practice might be outdated)_

- **Statics can scale at the same level or better than video: the top spender over the last 14 days is a static, on prospecting, not bottom of funnel.** If yours can't, it's 100% a skill issue. 99% of statics are headline plus product image or product plus benefit call-out. Boring, zero creativity, and that's the real gap. The few accounts that actually get the static game: @kamal_razzak, @AbyssusErigo, @harrydelmege_, @DTCMidas.

- **For every video you make, create statics built specifically for that video and publish them in the same ad set.** One or two handling the objections someone could have about things said in the video, one with social proof or authority backing up the claims, one testimonial re-iterating the promised outcome. The ad set becomes a self-reinforcing argument: viewers who saw the video get hit with proof and objection handling for that exact message.

- **Across 5 separate single-variable tests, the positively framed headline got 23x the spend and a £13 CPA versus £78 for the negative frame.** Positive: "Get results like {alternative solution}, without the {negative side effect}". Negative: "Ditch the {alternative solution}, get the {solution} without the {negative side effect}". The positive frame was the runaway winner in every single test, a learning only possible because everything else was held identical: change layout, messaging, framing and colours at once and the only takeaway is "this one got more spend". Put variety in the account batch by batch, not inside individual variations.

- **At scale it's around 15 new concepts a week with 3-5 variations each; starting out, 5 intentional concepts a week does well.** Statics vs video depends entirely on whether the product needs explaining: a jewellery brand scales with statics because the product just needs to look aesthetic, while a supplement brand with strong claims needs video for testimonials and explanation. For most people video will be much more scalable; you doom-scroll reels, not photos. Still keep a mix: a few BOF statics can improve overall account efficiency even in a video-led account.

- **"People only buy benefits, not features" is a blanket statement that breaks against reality.** Amazon find videos stick almost exclusively to features and specs and do huge numbers. There is a time and place for everything: pick the approach that fits the product and context, not the mantra.

- **A plain-background product photo with minimalist text spent around $200k as one of the top ads in the whole account.** Simple "artistic" statics can absolutely work for direct response if the style suits your audience. Match format to audience, not to the trend.

### Creators, UGC, and Creative Sourcing

- **Run multiple independent sources of creative and scale each one horizontally as long as it holds a positive ROI.** The stack: internal creative team, multiple agencies, freelancers, softwares for statics, AI, and seeding, all running at once. Every source sees things differently and makes visually different work, so volume arrives without stretching anybody or compromising quality, and you get concepts one team alone would never produce. If each source has a positive ROI, keep adding sources indefinitely: more strategists, editors and designers each responsible for a different thing, more agencies, more freelancers, more AI.

- **One AI podcast ad becomes dozens of tested creatives.** Make a 3-5 minute AI podcast ad, run CapCut's AI clipping for 15-30 second clips, post them across every short-form platform, run the original long-form in the ad account, and move the highest-reach clips into a single ad set. Organic reach is your free creative test; one video yields dozens of variations, organic exposure, and hopefully a couple of banger ads. The creator world's long-to-short clipping strategy imported into ecom with minimal effort. _(posted January 2026 · practice might be outdated)_

- **The direction you give creators matters as much as the script.** No tripods or phone clamps: the phone gets held in one hand or set down on a kitchen counter, and that one small detail alone gives a much more organic feel and better average results. No staged filming areas: creators shoot outside on a walk, in a car, walking around the house, lying in bed or on a couch, or mid-task cooking, prepping food, doing hair or makeup. Take your "losing" scripts, re-record with this kind of direction, and you'll see better results; the script may never have been the problem.

- **Price paid to creators has zero correlation with content quality.** Comfrt, arguably the most successful creator program of all time, starts creators at a $1k retainer for 30 videos a month, which works out at $33 per video, then moves them to performance-only. Their founder said it plainly: if a creative doesn't perform, they don't pay. Structure creator deals around performance, not upfront fees.

- **At a sub-10% ad hit rate, a $250-per-video-plus-usage creator is a burden, not an asset.** Do the brand's math: $250 x 10 videos is $2,500 for one winner, plus roughly $2,500 spent testing the 9 that didn't work, against an average ecommerce EBITA of 10%, and then 10% usage rights on top.

  A creator charging $50-$100 per 30-second video and doing good work gets paid again and again; brands are actively looking to replace the $250-plus-usage crowd. Smarter still is performance-based pricing: no upfront fee, a % of what the ad spends, and since Meta ads live far longer than TikTok videos a winner can pay you monthly for a year. Quality and punctuality are shockingly rare (creators beg for product, then deliver nothing, or ship terrible work 2 months later with zero product research), so being reliable and reasonably priced is a moat. Ignore the UGC courses teaching you to squeeze brands: it's not sustainable for them, and it's exactly why everyone wants AI creators to succeed.

- **Always dangle the carrot of a long-term partnership when negotiating with creators.** A 600k+ follower creator made 3 great videos, posted to her story, and engaged with the brand's content for $250 plus free product, because she was shown a path through the creator and ambassador program. Even big creators are not as well off as you think and their income is very inconsistent; steady money and a brand partnership are very appealing, so use it. Learn the art of the rizz, stop paying UGC creators, find people who actually know how to make content with an audience you can leverage. And if your product demos without talking (simple demos, before-and-afters), hire creators from Eastern Europe or South America: fantastic work at a fraction of what mid US content costs.

- **Seed wide, let the ad account judge, retain the winners cheap.** Run an Insense seeding campaign with 50+ creators selected on engagement rate, free product in return for a video. Put everything they deliver into the ad account without judging; performance data beats taste. Lock the best onto roughly $1-1.5k/month retainers for 15-30 videos (a quarter posted to their own profiles) or around $50 per video on a deliver-as-much-as-you-want basis. Repeat, cut the fat, offer the best of the best affiliate deals, hold weekly video training meetings: an army of killers scaling paid and organic on autopilot. _(posted March 2025 · practice might be outdated)_

- **Model IG accounts with 25k-100k followers are DMing brands for free samples in exchange for a posted video plus ad usage rights, while UGC creators try to charge $600 per video plus monthly usage.** The UGC market is a serious bubble. Source content from the models: you get content, distribution, and usage rights for the cost of a sample. _(posted February 2025 · practice might be outdated)_

- **Bullet-point briefs left open to interpretation guarantee terrible UGC.** Give granular detail about exactly what you want in the script and the delivery. UGC creators are not your creative strategists: if they were actually good at scripting and content ideas they'd have their own following and be making money from brand partnerships instead of hustling UGC. Strategy stays in-house.

### Testing Is for Learnings, Not Just CPA

- **A true winner announces itself within 24 hours of hitting scaling campaigns.** It picks up spend on its own, comes in 50%+ below CPA target at very low frequency, and lifts sitewide numbers: site CVR went up 17% after this one went live. Low frequency plus cheap conversions means headroom, not a small pocket being exhausted. When you see the pattern, scale budgets immediately and commit big; this single ad got close to 7 figures of spend.

- **Publish creative with forced spend, find anything driving much-below-average CPCs without converting, then build a landing page specifically for that ad.** Cheap CPC plus bad CPA is a landing page problem, not a loser ad: a low CPC is a clear sign of user interest and favour from Meta. For clickbait-style statics, don't send clicks to a PDP; take the topic or headline and continue the story with an advertorial or sales page, and once that page converts at or above normal you've unlocked scale. Build the lander after the creative proves itself so you aren't making a page for every creative. Bid caps, cost caps and "trust the algo" setups will never surface these ads: Meta sees the bounce rate on your regular lander and buries them, which is exactly why testing exists. Testing is not forcing $200 of spend and hoping it scales; it's collecting data, isolating variables, and understanding what drives performance.

- **Judging an ad purely on CPA or ROAS is ridiculous: any winning ad becomes a "loser" instantly with a shit offer or a bad landing page.** CPA judges the whole funnel, not the creative. Hook rate, hold rate, average watch time and CPC are what actually indicate a piece of content is good, and your account likely holds great ads you wrongly binned that just need a different landing page or funnel. If people engage but don't convert, figuring out the post-click experience is your job.

- **Track avatar, awareness stage, desire, creator, format, and editor on every single ad.** Religious tracking beats high-volume spam testing: know the specific elements driving performance and your strategist ends up struggling to make a loser, not even kidding. That taxonomy turns testing into compounding knowledge instead of a slot machine.

- **Test with single-variant changes: a typical testing ad set is one image ad with copy variations only.** If Meta really treated variations as the same asset they wouldn't all get spend, and hook iterations of winners wouldn't consistently pick up spend. Ignore whatever agencies and Meta are saying; you'll make more money. _(posted December 2025 · practice might be outdated)_

- **Don't stop single-variable tests because of the latest buzzword.** Creative diversity comes from new concepts, not from killing small variations of them. Three headline variations take about a minute; skipping them doesn't magically free time for new concepts, and stopping them throttles ad volume, learnings, and ultimately spend. Hook and headline tests will always be high leverage at almost no cost. Ignore creators chasing attention with every new buzzword like Andromeda; nothing has changed in this approach to ads in a long time. _(posted October 2025 · practice might be outdated)_

- **A low CPA you can't explain is worthless.** The goal of testing isn't publishing something cheap; it's knowing exactly why that specific thing worked. Clear knowledge of which messaging, headlines, hooks, imagery, colours and styles drive performance is what keeps CPAs low over a long horizon. This is the unsexy answer to "how do you scale so fast": don't be lazy, do the work, extract meaningful learning from everything you publish, or fight the rising CAC monster forever.

- **When an ad shows good soft metrics but bad CPA, throw up a quick lander dedicated to that specific ad.** Works especially well for click-baity low-CPC statics; people who tell you not to run these usually never have, and those that have, know. Publish, see what picks up spend at low CPCs, then write an advertorial or listicle whose headline continues the bait: static "Man discovers 2 minute miracle that fixed his back pain for good" -> advertorial "How this 45 year old man fixed his back pain with a simple 2 minute daily routine being described as The All Natural Pain Miracle", then go full Sugarman down the slippery slide. Cheap if you've pre-built the formats, and it makes a big difference to hit rate.

- **The exact same ad accidentally uploaded 6 times into one ad set produced night-and-day results.** Three got decent spend, one great ROAS, one decent, one brutal; the other three barely spent. Nothing differed but which pocket of people Meta hit first. Early signals are luck, not verdicts: Meta shows bias to ads that get strong signals after just a few impressions, and a real winner that hits a bad pocket in the first few dollars may never see the light of day again. At decent creative volume, use minimum spends on ad sets in a CBO to force Meta to actually test things. _(posted June 2025 · practice might be outdated)_

- **A testing campaign isn't there to be profitable; it's a controlled environment for isolating variables and understanding why winners win.** Publishing everything into one CBO forces new creatives to compete immediately against ads with hundreds of thousands in spend and tons of engagement, so they get no spend, no data, no iteration: potential winners sent to the graveyard. Find the best version of an ad before putting it to the real test at high budget. Almost all of the biggest winners were "Frankenstein" ads assembled from incremental changes discovered through controlled testing, ads that would have stood no chance launched straight against proven winners.

- **No single metric guarantees a winner, but several soft metrics together have a high correlation.** A fresh ad picking up spend with the makeup of every big winner is worth betting on before sales data lands; study what causes each metric and apply it to future creatives and your winning-ad hit rate goes up. Bonus tactic for products with clear before-and-afters: clip the reveal or final result as your hook, then cut back to the start of the process to tell the story of how you got there. Hold rates go brrr. (Hold rate: % of viewers who, after the first 3 seconds, watched at least 15 seconds or to the end, whichever is shorter.)

- **High CPA doesn't equal bad ad.** Take a proven winner, let someone rip it to a different landing page and offer: theirs shows $100 CPA against your $20, same ad. Slap a bad or irrelevant hook on a winner and CPA-only thinking calls it a loser overnight. Soft metrics diagnose the fix: strong hold rate and CTR with a bad hook rate means make a better, more relevant hook, and suddenly it becomes a "winner", a test you'd never even run looking only at CPA. It could just need a few tweaks somewhere.

- **Launch a cold-start brand with around 10 creatives covering as many main angles as possible in a few different formats, all inside one ASC.** Diversity of angles matters more than polish on any single ad, and ASC is very good at picking up high-intent buyers, giving the best chance of success from a cold start. _(posted February 2025 · practice might be outdated)_

- **Set testing budget at roughly 4x target CPA per day with a $100 minimum.** Target CPA $20: still test at $100/day, the floor. Target CPA $50: test at around $200/day. A flat "test at $100 a day" rule, despite what you may hear, makes no sense for higher AOV products; the budget has to buy enough conversions to judge the test.

- **The launch recipe: at least 10 creatives, budget around 4x target CPA per day, minimum 3 days, ideally over a weekend.** Break-even or profitable: keep it running and optimise on the metrics you see. Unprofitable: hunt for signs of life first, high add-to-carts, high CTRs, high CVR, fix the corresponding part of the funnel and give it another day or two. Only if nothing responds do you move on.

- **Spend is a performance metric.** Testing in ABO gives you no idea of Meta's preference for where to allocate budget, arguably the most important signal there is. One ad with below-average CPA at low budget kept eating spend without performance dropping until it was the top spender in the account, lifting total spend, revenue, and profit; in ABO it would never have met the CPA gate for a budget increase. Huge audience plus really engaging means Meta loves spending on it, and that only reveals itself when the algorithm is free to allocate. _(posted January 2025 · practice might be outdated)_

- **"My ads spent $15 and CPC is above $1 so I killed it" is how you kill winners.** Context-free kill rules give no weight to product price, market, or audience quality; following them would have meant zero brands and zero success. CPC is barely worth caring about at all: want cheap CPCs, go run ads in India and see how that turns out. A fashion-dropshipping background is a disadvantage for brand building, full of bad habits, misread data, and no patience. Unlearn the reflexes first.

- **Media buying debates act like the only costs live inside the ad account.** Your strategy choices drive creative production costs and staff hours that never show in the ROAS column but certainly show up on the P&L. Compare strategies on total economics, including the creative volume each one demands, not just in-platform efficiency.

- **No football team only looks at the final score.** Professional teams track pass completion, distance covered, tackles, shots on target, because "we lost 2-0, how do we improve?" can't be answered with "just score more goals". Hook rate, hold rate, CTR and CPM tell the story behind CPA: yes, CPA is what you optimise for, but just as no team wins with a 10% pass completion rate, you can't expect a winning ad with a 10% hook rate and a 5% hold rate. People who claim they only look at CPA are lying or unbelievably ignorant.

### Cost Controls and Meta's Real Mechanics

- **Regardless of AOV or RPS, CVR still matters: it's a key input to Meta's ad-quality score, and ad quality decides who wins auctions.** CVR goes up -> ad quality goes up -> auction ranking goes up -> lower CPA. Improving on-site conversion directly buys you cheaper traffic.

- **Sitewide CVR halving overnight alongside a traffic spike is a placement problem, not a site problem.** Meta was randomly dumping spend into Audience Network placements with around $0.13 CPCs on high-spending US ads, terrible CVR, and bad CPAs. The signature: CVR down 50% overnight, a large traffic increase, abnormally low CPC. When cheap traffic floods in and CVR tanks, audit placement breakdowns before touching anything else. "Always trust the algo" though. _(posted May 2026 · practice might be outdated)_

- **Highest value with no cost controls worked great; TROAS missed its target by around 50% across $10k over 10 days.** Ironically, cost controls wasted more testing money than anything else. Back to lowest cost CBO with minimum spends. _(posted July 2025 · practice might be outdated)_

- **Read Meta's "best practices" as system-protection, not performance advice.** The mental model: a milk shop with 10 cows and an ever-growing line has to ration supply or the whole thing breaks. Each rationing move maps to a real mechanic. Stop letting people choose which cow: broad targeting and Advantage+ removing control, so worse inventory can be quietly mixed in unnoticed. Charge more the more you take: unexplained CPM ratcheting on aggressive scalers. Reward coming in as a group: consolidation guidance that looks like a deal but delivers less than individuals got before. Maybe it's not what works best for advertisers; it's how Facebook protects finite impressions while herding advertisers with promises of better performance.

- **Triple-firing the pixel to inflate reported conversions dropped performance 30%.** The theory was that doubling your effective action rate wins more auctions; the reality is Meta curtailed delivery, and it appears to cap how well anyone can do to preserve its ad inventory. Supporting anecdote: a $160 AOV brand dropped an ad getting $0.60 clicks and 3% CVR, scaled to $20k/day within days, then CPMs went from £35 to around £150 and never returned to normal, despite some of the best engagement metrics and feedback you'll ever see on an ad. _(posted June 2025 · practice might be outdated)_

- **Meta has performance swings outside your control.** When the platform itself reports bugs, the director of Meta Europe acknowledges it in a meeting, 99% of the operators you know are down, and every account you can see is down month on month, denying it is delusion, not discipline. A bad Meta period doesn't mean your business is failing, but don't let "just make better ads bro" gaslight you into ripping up a working system over a platform-level swing. _(posted May 2025 · practice might be outdated)_

- **Even Meta themselves recommend cost caps now, and "Meta just wants you to spend more" is exactly why their recommendations work.** Yes, they want you to spend more; they achieve that by making you more money, because you spend more when you make more. Deliberately sabotaging advertisers makes zero sense for their business model. _(posted March 2025 · practice might be outdated)_

- **The percent of new visitors from your ads has nothing to do with campaign type; it comes down to spend ratios.** Set up a new-visitor-percentage column in your Triple Whale attribution dashboard, put the same ads in any 2 campaign types at 2x and 1x budgets targeting the same location: the higher spender always gets far more new visitors, and swapping the budgets flips the percentages. The "ASC just retargets" myth was a budget artifact of Meta's launch guideline to give ASC 30% of budget; any campaign on 30% against another's 70% shows fewer new visitors. It's also why consolidation matters: multiple campaigns on the same market and product muddy attribution with cross-over and a shit ton of false last-click signals. Keep it simple. _(posted March 2025 · practice might be outdated)_

- **The "12 touchpoints before purchase" rule is bs.** Almost everything ever sold showed the same number of purchases on a 1-day-click window as a 28-day-click window; if 12 touchpoints were required, the long window would capture far more. When starting out you're tapping bottom-of-funnel traffic already warmed by more established players selling the same thing. You just have to get them over the line.

- **Cost caps and bid caps are bid settings, not optimisation settings.** Target CPA flexes daily and averages out over your attribution window; on the over-target days, what do you think is happening if not upper-funnel reach? The cap level is your choice: break-even CPA $50, account average $30, set a cost-controlled campaign at a $100 target and tell me you're "only going after bottom of funnel". Caps are adjustable levers: plenty of cash, up them and spend at less efficiency; tight, lower them. The setting that actually skews optimisation toward "buy now" people is a 1-day-click attribution window, not your bid type. _(posted December 2024 · practice might be outdated)_

- **The "cost controls can't grow a brand" argument falls apart on five fronts.** They cycle naturally: with target CPA you'll spend well over target for a day or two then come in under, averaging to target over the attribution window, natural cycles of heavier prospecting. The cap is adjustable: want a period of heavier awareness spend, up it and keep control of your margins. Lowest cost is guesswork: you never know if you could have spent more, and budget cuts only happen after you've already lost money. Caps force you to solve for spend, pushing effort into creatives, offer and CRO instead of forcing budget and hoping performance magically improves. And time: Meta is an employee better than anyone on earth, working 24 hours a day for free; checking the ad account several times a day to adjust budgets is not founder work. _(posted November 2024 · practice might be outdated)_

- **Straight from two days inside Meta: they're all-in on Advantage+ shopping and AI, allocating zero resources to manual sales campaigns.** It's like manual campaigns don't exist to them anymore. Their other emphasis: creative diversity is the key to getting the most from ASC, meaning assets that visually look very different from each other, all formats including carousels and DPAs, plus variation in messaging and awareness stages, and it goes beyond video vs image. "ASC is for beginners" is laughable; Meta isn't dedicating the resources of its biggest revenue stream to helping mums sell hand-knitted socks. Look at their public Llama models and imagine what's behind the curtain in the ad algorithms. _(posted November 2024 · practice might be outdated)_

- **Target ROAS as the cost control plus a collection page destination lifted AOV about 20% with CPA at account average, actually slightly lower.** The ROAS-based control lets Meta optimise for order value rather than just any purchase. Cheap experiment, meaningful upside; test it before assuming product-page traffic is always the answer. _(posted November 2024 · practice might be outdated)_

### Account Structure and Forcing Spend

- **"Never turn off an ad" is the dumbest rule in ecom, and people have been psy-oped by it.** Time and again: ads above target eating spend in a CBO get turned off, CPA drops massively, MER improves, and spend gets scaled up as a result. Obvious to anyone outside the DTC Twitter bubble. Turn off the underperformers.

- **Meta's delivery is picking the wrong ads at a near 100% failure rate.** The more you manually intervene, push spend yourself, and turn off the ads Meta favours, the better things go. Same-day proof: surf-scaling 2 ads Meta didn't want to spend on made them the account's top spenders at roughly 50x the budget of other ads with a CPA 4x lower than average. The winners were sitting right there; Meta just wouldn't feed them. _(posted June 2026 · practice might be outdated)_

- **Preaching creative diversity means nothing if one ad is taking 90% of your spend.** CBO will happily concentrate the budget on it. ABO that shit: control the distribution yourself so the rest of your ads actually get spend.

- **Three campaigns kept spend growing at better CPAs while most of Meta was down bad: ABO testing, CBO scaling per market, and a bid-cap graveyard.** Testing: all markets combined, every new concept starts at $100/day, losers cut early on leading indicators like CPATC, winners scaled there and duplicated into scaling campaigns. Scaling: lowest cost CBO per market, winners stacked in with a new ad set each month, "January winners", "February winners". Graveyard: bid cap, all markets combined, all testing losers plus past winners no longer spending get a second life. Attribution is 7DC 1DEV across all campaigns; exclusions everywhere via a synced Klaviyo list of existing customers plus 180-day website purchasers. Structure is still secondary to creative and post-click experience, but this setup has had zero issues. _(posted April 2026 · practice might be outdated)_

- **Don't drop budgets after Black Friday.** Everyone automatically pulls back after the peak, which leaves the auction cheaper for whoever stays in, and Saturday and Sunday consistently outperform: up 35% on the Sunday compared to Black Friday itself, the same pattern year over year. Hold spend from Friday through the weekend. _(posted November 2025 · practice might be outdated)_

- **Running a single CBO and trusting Meta to allocate budget? Take all the losers that never got spend and put them in a new campaign.** That campaign will often become one of your best performers and total account spend goes up. Meta is super autistic: it grabs onto one or two creatives and won't let go for dear life, so budget allocation is not actually optimal and it's not always "just make better ads bro". Proven time and again across his own brands and countless others. This isn't the whole account structure; it's the proof you can't blindly trust Meta's allocation. _(posted November 2025 · practice might be outdated)_

- **Winning ad hit rate pretty much 2x'd from forcing spend to ads, no other changes.** Algorithm-starved ads are false negatives: when delivery concentrates on 1 or 2 creatives, the "losers" were never actually tested. Don't let Meta's default delivery decide which of your ads deserve a real test. _(posted November 2025 · practice might be outdated)_

- **With campaign-level budget, splitting ads into different ad sets now spreads budget far better than consolidating, the opposite of how Meta used to behave.** Everything in one ad set means 1 or 2 creatives take all the spend and most get literally zero, not a single penny. Duplicate the campaign and group ads into different ad sets and previously starved ads pick up a good percent of spend, observed across multiple ad accounts in all markets. _(posted November 2025 · practice might be outdated)_

- **Ad account structure doesn't matter much for scale (as long as you're not doing anything retarded), but it makes a huge difference to how easily you extract learnings and gather feedback.** Judge structure by feedback quality, not scale potential. Tip: the 1 CBO method ain't it.

- **Flexible ads are for consolidating winners, not testing.** Take your current top 10 spenders in a campaign, add them all to one flexible ad within the same campaign, and watch it take over: likely spending more than all ten put together, with a longer life span. Don't be a retard and use them for testing; they obscure per-asset feedback. _(posted October 2025 · practice might be outdated)_

- **Upping budget doesn't automatically drop performance; often it improves it.** Especially in dropshipping niches where you're competing against a bunch of teenagers scared to pump budgets, aggression is an edge in itself. Two stores in one week, budgets doubled multiple times, not 20% bumps, and pretty much every metric went up.

- **The full account playbook: one CBO testing campaign with minimum spends covering every market you ship to, plus ASC scaling campaigns grouped by geos with similar breakeven ROAS.** Worldwide campaigns make no sense: shipping cost and AOV vary massively by country, so working off an average always leaves spend on the table somewhere and runs margins too tight elsewhere. Typical groups: USA, UK, ROW (lower breakeven), ROW (higher breakeven), each with its own targets.

  Testing: CBO with a minimum spend per ad set, each ad set one creative with 3-5 variations, run at least 7 days and at least 4x target CPA. $50 target x 4 = $200 over 7 days, so set the minimum around $25/day, and keep total minimums under 50% of the CBO budget; creative volume dictates the budget (10 new batches a week x $25 = $250, doubled = a $500/day testing budget). Graduation rules, always judged on the last 7 days: stuck at minimum spend with CPA below target, kill it; above minimum at or above target CPA, graduate to scaling; spending way above minimum with CPA below target, graduate anyway and watch overall account MER for a few days, keep it if MER improves. Turn graduated ads off in testing once they pick up spend in scaling, avoiding overlap and false last-click signals while making room for new tests.

  Scaling: ASC with a target-CPA cost control and an inflated budget. Stack your winners and let them rip. Judge results only over your attribution window (7 days for most) since Meta averages to target across it, and never touch a budget again; Meta has far more data than anyone and beats any reactive human budget change. _(posted March 2025 · practice might be outdated)_

- **Hard-won lessons that go against popular opinion.** You don't need a high AOV to scale: lower AOV opens much larger audiences and more markets, and high margins with low AOV can work extremely well. ASC has been better than CBO for a long time, and % of new visitors is not lower with ASC; it's decided by the spend ratios in your account. If you fulfil from China, don't over-focus on the US: seamless worldwide shipping is the main benefit, and done right the US should be under 50% of revenue.

  Never scale with worldwide campaigns: breakeven points differ wildly by country, so you always leave spend on the table in some countries and run margins tighter than desired in others. Lazy and inefficient, and even the people it "works" for could do better. Run a testing campaign covering every market you ship to; efficiency isn't the point there, seeing what works where is, and creatives that flop in the US often work in Europe and vice versa. Don't use flexible ads for launching or testing. Ad account structure matters a lot; meanwhile packaging is not that important, brand matters less than price, and email pop-ups are not high leverage. _(posted March 2025 · practice might be outdated)_

- **If you ship from China but only run English markets, create a new campaign with all your top ads targeting all of Europe with an all-English funnel.** Cost cap around your average CPA, inflated budget, almost no additional effort or risk; you have no idea how much money you're missing. Local-language content is not automatically better: twice German ads were tested against English ads in Germany and both times English won, and most smaller European countries prefer to see English brands anyway. _(posted January 2025 · practice might be outdated)_

- **CBO and ASC have swapped behaviors: CBO now latches hard onto 1-2 creatives while ASC gives new ads a real chance.** Dozens of ad sets getting zero spend in CBO picked up spend and performed well when dropped into an existing ASC. ASC has consistently performed better, tested more than most; if Meta gave ad sets for ASC or breakdowns for flexible ads there'd be no reason to ever run manual sales campaigns again. _(posted January 2025 · practice might be outdated)_

- **Meta heavily favours the flexible ad format, but test normally and promote proven winners into a flexible ad in your cost-controlled ASC.** The top 10 ads from a heavily-spending ASC combined into one flexible ad in the same campaign started spending heavily from day one; a month later it spent more than double all the other ads in the campaign combined, with overall campaign spend up. Flexible ads stay next to useless for testing, so keep the framework: separate testing campaign, 3-5 ads per concept, each concept in its own ad set, so your creative team can break results down and isolate variables, then extract the variations hitting CPA targets and launch them as a flexible ad in the scaling campaign. _(posted October 2024 · practice might be outdated)_

- **ASC only looks like retargeting when your manual sales campaigns starve it of reach.** Higher-budget manual campaigns hit most of the top-of-funnel audience first, leaving ASC little but people already touched by your other ads; turn them off and ask who ASC is going to "retarget" now. It works best as an "all of funnel" campaign fed a wide variety of ads across awareness stages and formats: "top of funnel" ads take the highest spend, lowest frequency and worst CPA but feed the "bottom of funnel" ads sweeping up sales at low spend and low CPA. With a cost control set at campaign level, Meta gets room to spend more at a worse CPA on awareness-driving ads and balances it with the closers, levelling out to your target. Don't run ASC with cost controls until your library can create that funnel; eight months all-in produced stable results, stable CPA, and none of the "fatigue" issues people complain about. _(posted September 2024 · practice might be outdated)_

## Chapter 03: Ecommerce & Brand

### Only a Few Things Move the Needle

- **The needle-movers are all "new": products, landing pages, offers, channels, and creatives, mainly new angles unlocking new audiences.** The busywork list is just as specific: vibe coding apps you can already pay $100/month for, refining your "brand guide", updating SOPs, restructuring Meta for the 100th time, email popups, AI apps, and automating your shitty static ads with AI. If your week is full of the second list, you're not growing.

- **Only four things put more money in your bank account in ecommerce: new products, creative, landing pages/CRO, and retention.** If your time goes anywhere else, audit your life. Perfecting SOPs, building fancy spreadsheets, and vibe coding apps with AI feel productive but don't move the needle. Contingency planning, asset protection, and security matter only once you're at scale: get money first, and never "optimise" when the cost of doing so is greater than the sum of your worth.

- **Branding is not that important, especially when starting out.** People do not think about your brand nearly as much as you think, and the "you need branding for retention" claim doesn't survive contact with real data: Jacob launched a brand in June with $0 on branding, GPT-generated images, and a website built in 2 days, and it's about to hit $1m MRR with healthy retention. Meanwhile people who spend silly money perfecting branding pre-launch get out-scaled by someone using AI images and Shrine theme. The time went to customer research, competitor research, product quality, copywriting, post-purchase education and habit-forming resources, and ads, ads, ads. Know your customers and their issues intimately, have a quality product that solves the issue, and convey it clearly in ads and funnels: that's the whole formula for acquisition and retention. Pretty packaging won't scale you, unless you're in an aesthetic-led category like jewellery or fashion.

- **Seed the social apex, not influencers.** On Taylor Holiday's advice, Jacob tracked down the heads of sororities at the biggest US universities on IG and sent each a big gift box, no payment, no strings. The heads started using the product, then the rest of the sorority, then the rest of the girls at the universities. Gift boxes create genuine usage that cascades; paid sponsorship posts don't. Gymshark runs the same playbook, and it's repeatable: identify the org, find its leaders on Instagram, ship the box.

- **High-AOV stores are the hardest to get traction with and the best to scale.** Expensive products demand more consideration and trust, so early wins come slow: this was the hardest-to-scale store Jacob ever ran. Low AOV gives you easy initial success because cheap products take little consideration, but that's the ceiling trade-off. Once creative and offer are cracked, the high contribution margin lets you push scale further than any low-AOV store can.

- **Ad performance fluctuations are guaranteed, so build the business to absorb them.** On sub-20% margins every bad day shakes you and pushes you into reactive mistakes. With high margins and a strong backend, solid email retention especially, a bad front-end day just means less profit that day, not a loss.

### Sourcing and Fulfilment from China

- **Shopify tells you market-specific exact pricing isn't possible. It is, via the product CSV.** Set the US up as its own market, go to Products then Export, edit the US price and compare-at price columns in the CSV, then Products then Import. US pricing, strikethrough anchoring included, is now fully separate from every other market: exactly what you need to absorb China-to-US shipping cost changes without repricing the rest of the world. And charge shipping for the US. Jacob always does and never sees checkout drop-off if the rate is reasonable. _(posted April 2025 · practice might be outdated)_

- **Chinese fulfilment done right beats the horror stories.** Ordered 14th December, delivered to the UK on the 18th: 4 days, at $3.05 shipping. No storage fees, no pick & pack fees, and 10-day lead times from ordering inventory to it sitting in the 3PL ready to ship, so you can run far leaner on stock than Western 3PL defaults suggest. _(posted December 2024 · practice might be outdated)_

- **If your Chinese 3PL charges you both storage and pick & pack fees, you're getting bent over.** Competitive Chinese 3PLs don't need either line item; treat one on your bill as the signal to renegotiate or switch. Audit the fulfilment bill like you audit ad spend: these fees compound on every order and quietly eat the margin your media buying is fighting for.

- **Validate demand, competition, and product flaws before you ever touch inventory.** Identify high monthly search volume with Exploding Topics or Keyword Planner, cross-reference Amazon US and UK and TikTok Shop monthly sales, then map competitors in the Meta ad library: how long they've sold the product, active ad count, which countries, cross-checked against Similarweb or Semrush for traffic volume and source. Mine Amazon, Trustpilot, and Reddit reviews for dislikes and issues, then exploit the gaps: open markets, pricing and offer gaps, fixable flaws.

  Build the website and ads off that research and run campaigns without stock, or dropship, refunding any sales if you're stockless. Only order MOQs once ads are profitable, customised per your research, and fulfil from a Chinese 3PL for an operating and margin advantage over competitors. _(posted September 2024 · practice might be outdated)_

- **Alibaba manufacturers will hand you market intel for free, and almost nobody asks.** Best sellers, most popular variants, top countries: manufacturers will just tell you what's selling. Going to China is unrivalled for product research, but chats with Alibaba suppliers get you a surprising amount of the same signal from your desk.

### CVR Comes From What You Remove

- **When CVR swings and you swear nothing changed on the site, 99.9% of the time that's wrong.** The reviews widget is the only part of a landing page whose content changes dynamically every day: one day detailed, relevant reviews sit at the top, the next it's "great product", and to the buyer that's a different PDP. Pin your best and most relevant reviews to the top so they never move and the page stops mutating without you. The other known swing factors: different ads taking different amounts of spend, and natural weekly, monthly, and yearly buying cycles.

- **When in doubt on a PDP, hammer all four points of the value equation, especially above the fold and in PDP images.** Dream outcome: the benefit they'll see and how it changes their life. Perceived likelihood of achievement: stats from studies, testimonials, guarantees. Time delay: tell them when to expect results, 1 year vs 1 week is a huge difference in value, and a simple timeline of benefits works here. Effort and sacrifice: if the outcome feels like a chore, value collapses, which is exactly why GLP-1s are so valuable, weight-loss dreams with no effort. In supplements that means fewer pills per dose and "tastes great". Do this and you can't fail.

- **Collect non-customer emails from abandoned checkouts, not popups.** Someone who reached checkout is far higher intent than someone who typed an email into a popup, so the list stays smaller, hotter, and earns more revenue per recipient with less junk clogging it. Bonus: no popup software or email agency claiming credit for sales that would have happened anyway, since 9 out of 10 popup "conversions" would have bought regardless. Killing the popup stops paying attribution tax on those sales.

- **Every extra choice at checkout adds decision fatigue at the worst possible moment.** When Shopify replaced the single "pay now" button with a "Pay as guest" versus "Pay and save my info" choice, CVR tanked, especially for Jacob's mainly 65+ demographic, because each ambiguous option spawns anxious questions: does guest checkout lose order tracking? What will they do with my info? What is Shop, do I need it? Shopify wasn't testing this to improve merchant performance, it was to pull more people into the Shop ecosystem, and the reversal reportedly saved merchants billions collectively. Keep checkout to one obvious action, and watch for platform changes that optimise for the platform instead of your CVR. _(posted May 2026 · practice might be outdated)_

- **Popups tested net negative every single time across 2.5 years.** Jacob reluctantly let his agency set one live, woke up to the worst CVR since the brand started, pulled it halfway through the day, and CVR snapped back to normal immediately. Agencies will swear popups win; other agencies swear the opposite. Your store's data is the only arbiter, and his repeated tests across brands always came out net negative.

- **Name the trusted local courier in your shipping rate and checkout starts feeling domestic.** Pull a Shopify report for sessions by location, add the checkout conversion rate metric, and find markets with high sessions but heavy checkout drop-off. Check your order tracking app filtered for that country, look up last-mile tracking to see which courier actually delivers, then create a shipping profile for that country with the courier in the rate description: "Canada Post Tracked Shipping". Check checkout CVR again in 7 days. A pure trust play that costs nothing.

- **Value props beat offer-only banners.** Changing the homepage banner from just stating the current offer to giving clear value props of the brand drove an almost 18% increase in profit per site visitor, tested to 97% statistical significance. Sometimes it's the simple things, but prove them with a real test, not vibes.

- **High CVR usually comes down to the things you don't do.** First, CVR isn't even the right target: most people could double it by halving prices without making more profit. The wins come from subtraction: strip unnecessary steps, offers, text, and elements, because overwhelming sites feel like a chore to consume and add friction. Then run the mum test: hand your phone to your mum or grandmother, ask her to buy something, and don't intervene. Wherever she struggles, simplify. Afterwards quiz her: does she know exactly what she's getting, the return window, the guarantee, how much she saved, and when it arrives? Whatever she can't answer, make clearer.

- **Split test a 60-day return window against 30 days.** When Jacob did, CVR and RPSV went up almost 20%: the extra guarantee removes purchase risk in the buyer's head. The downside is theoretical: still not a single return after 30 days, so the perceived generosity is nearly free. A one-toggle split test with a measured ~20% payoff, worth running for everyone.

- **High-quality product photography is one of the cheapest needle-movers available.** Around £750 of photography and renders drove a 20% CVR increase plus a modest AOV lift on Jacob's first brand, and the efficiency ripples through every ad in the account: breakeven ads turn profitable, profitable ads start printing. Scammy funnels and shitty product images can still convert, and "it works bro, the messaging matters" is technically true, but a landing page that doesn't look crafted by the Nigerian Prince himself would make meaningfully more.

### Landers, Listicles, and Split Tests That Lie

- **Page changes affect Meta itself, and your split testing tool can't see it.** Intelligems says CVR and RPV are up 10%, you roll out, and nothing changes or things get worse: the change moved CPMs and shifted targeting on the ad side. A test can show a variant beating control by 20% while CPMs for traffic going to that variant rose 40%, and the tool will never show you that. Monitor your in-platform ad metrics for a week or so after every rollout, otherwise the "winner" you thought you had could actually be a loser. Not enough people ever think about this.

- **Your listicles and advertorials don't need images.** Copy carries the page. Treat that as a permission slip to ship text-only pre-sell pages: skipping images removes a production bottleneck, so you can test new angles faster instead of waiting on visuals.

- **AI landing page builds are so fast there is no excuse not to have a unique lander for every winning angle.** The play: a new angle going straight to PDP shows signs of life, you feed the ad transcript to AI, it builds a simple congruent listicle, live in 20 minutes. Jacob then duplicated the original ad into the same campaign pointed at the new lander: the duplicate outspends the original at half the CPA, now doing roughly 3x the original's daily spend despite launching 5 days later. Don't perfect every funnel before launching. Ship scrappy, check for signs of life, then invest the time and money. _(posted March 2026 · practice might be outdated)_

- **Split test pages from the ad account, not with on-site tools.** Anyone running more than one page from an ad account has seen that the exact same ads at the exact same budget produce different CPMs and CPCs on different pages: the page itself moves auction costs. On-site split-testing tools miss that side entirely, so a "10% lift in rev per visitor" can roll out as a 20% CPM increase across the account, a net loss the tool never saw. Testing from the ad account is the only way to capture both conversion performance and what Meta charges you to send traffic there. _(posted January 2026 · practice might be outdated)_

- **Analyse your listicles with Heatmap and let revenue data reorder the page.** Re-order the list of reasons by rev per session per reason instead of gut feel, then find the reasons with outsized scroll drop-off and swap the copy and/or image. Each fixed drop-off point is recovered revenue. Tool: @heatmapme. Easy money. _(posted December 2025 · practice might be outdated)_

- **Never split test a new landing page at the site level straight away.** If the page is shit and it's getting 50% of your traffic, you eat a big fat L. Graduate it instead.

  Step 1: cheap pulse check. Take your top 3-5 existing ads for the angle, put them in one ad set in an ABO campaign pointed at the new page, just to confirm it's not terrible. Don't skip this and jump straight to duplicating a proven ad: Facebook will spend heavily on a creative it already knows and likes even when the page is bad. Step 2: if the pulse check looks promising, duplicate your top ad inside the main campaign and send the copy to the new page, watching whether it picks up spend and outperforms the original. Step 3: publish all new ads in two versions, one to each page, and see which consistently wins. Step 4: once the new page consistently outperforms, publish only to it. By now you have enough traffic to justify on-page split tests of headlines, copy, and images. Squeeze the page, then repeat the whole process.

- **Before you pay anyone silly money to build a website, grab a basic Shopify theme and browse the Section Store app.** You can assemble almost anything you'd pay an agency or freelancer for within hours at almost no cost. Hours instead of weeks, near-free instead of thousands, and the saved money and time go into creative and offer testing where they actually compound. _(posted June 2025 · practice might be outdated)_

- **Ad-based CRO is goated: base the advertorial copy on a winning ad angle already in your account.** The landing experience continues the exact message that already converts. The mechanic is simple: write the advertorial from the proven angle, duplicate the winning ad, swap the destination. The duplicated ad became the biggest spender in the account on day 1. Fast to ship too: around 1 hour on Pagedeck from signing up to having the advertorial live. _(posted January 2025 · practice might be outdated)_

- **The majority of customer surveys are complete bs when read literally.** Jacob's product-page survey asked what was holding visitors back from buying: number one answer, the price. His thank-you-page survey asked the main reason customers bought: number one answer, the price. Same store, same answer, opposite conclusions. Treat surveys as directional and validate against behaviour, orders, refunds, and session recordings before acting on what people say.

### Subscriptions, Churn, and Refunds

- **Churn came down 25% from operational moves, not marketing.** The changes: moved to US fulfilment, since faster and more reliable delivery shows up directly as retention. Gave customers more delivery-frequency flexibility, so people adjust cadence instead of cancelling. Changed factories to get the highest quality and most expensive version of the product, at a price competitors wouldn't pay even if they could get the same pricing, which they couldn't. Plus unshared work in the customer journey and product experience. The levers for churn live in ops and experience, not the ad account.

- **Free gifts with subscription tested worse on every metric.** A split test of gifts vs no gifts showed a staggering drop in CVR, AOV, and RPV on the gift side, and fewer people even opted to subscribe. Gift bundles create confusion and clutter, which Jacob had long suspected watching other subscription brands pile free stuff onto their offers. Simple always wins. Split test it before copying the gift-stack meta.

- **You don't need capital, projections, or months of floated losses to build a subscription brand.** Jacob started with a 250pc purchase order and $100/day in ad spend and hit 100,000 orders and 7 figures/month in about 5 months, profitable, no debt, even with a 25% PayPal hold.

  The formula. Fulfil from China at the start, non-negotiable: minimal costs and fast lead times let you react to performance without large POs or data-less projections. Do not enable subscriptions at launch, they always lower CVR, and early on more conversions and data beat take rate. Test offers until you're first-order profitable, which took him about a week at $100/day, with quantity breaks doing the heavy lifting. Around $1k/day in spend, introduce subscriptions: CVR drops but take rate is high, and you stay first-order profitable so the first small rebills can't catch you with your pants down. When rebills hit, scale ads at break-even and take profit from repeat revenue, which compounds fast: 45% net going into month two of rebills. Only scale ads at a loss once you have 3 months of LTV data and cash in the bank, because that data tells you the CAC you can actually afford. Startup costs were zero: website reused from an old template and built in 2 days, product images done with AI and Canva. "If you think you need to take on debt or raise money, sorry but it's a skill issue."

- **Returns are often an education problem, not a product problem.** Customers were returning after about a week because they had difficulty using the products, so Jacob posted a couple of tutorial videos on YouTube for existing customers. The videos helped massively with return rate, and a year later had racked up almost 100k views: post-purchase education doubling as evergreen top-of-funnel content, for free.

- **Diagnose when and why refunds happen, then fix it in the post-purchase flow.** Analysis showed most returns came 2-3 weeks after delivery and the main reason was customers struggling to use the product. The fix: recorded educational videos linked from the welcome flow, plus explicit expectation-setting that most customers find the product takes 2-3 weeks to get used to, and after that it's a breeze, reframing the struggle as normal. Refunds fell from 3.91% in August to 1.04% in October, a 73.4% decrease, saving thousands a month from listening to customers and a few small educational changes.

### Team, Agencies, and Not Overhiring

- **Make every manager-level hire fully responsible for the people beneath them: hiring, onboarding, training, and even paying them.** Jacob's customer service manager hires agents, trains them, and pays them from a single payment Jacob sends him; headcount changes arrive as a heads-up, not a task. Same with the creative strategist, who owns the video editors. The people under your managers never contact you and you do zero management of them, which strips enormous day-to-day complexity out of the business. It's an agency structure without agency prices. And culture doesn't require knowing everyone: Tobi, Bezos, and Musk don't know all their employees, and those companies have no culture or vision problem. Standards set at the top trickle down through management.

- **Good agencies compound, but most people cut them before they've figured out the business.** Jacob's first month using @aaronmtrx for creatives was decent, not amazing; six months later they had almost all the top spenders in his account. Yes, plenty of agencies happily take your money and churn you a couple of months later. But agencies need ramp time the same way a new brand or product launch does, and good ones have been key to his growth. The failure mode isn't only picking bad agencies, it's firing good ones too early.

- **Give customer service agents personal affiliate links that pay a small commission on sales.** It turns comment-section support into an incentivized acquisition channel: agents actually help potential customers instead of just clearing tickets. Scope it tightly though. Jacob's rule was comments-section only, and one agent still went rogue, building pages impersonating the brand plus an army of DMers pitching "a better deal" through his affiliate link. Incentives work so well you need explicit guardrails on where they can be used. Fire him or promote him, GG.

- **Before you bolt an incentive onto a customer service team, map how each KPI gets gamed.** Average response time encourages rushed responses. Return and refund rate makes agents stonewall refund requests and piss off customers. Trustpilot rating makes them pushy about begging for reviews or over-pleasing customers in ways they shouldn't. Tickets answered produces shallow replies that don't fully solve the issue, or drawn-out interactions to farm ticket count. Every obvious metric creates a perverse behavior; design the incentive around the gaming, not the dashboard.

- **If your brand is under a year old, be careful hiring and making big investments off Q4 numbers.** The surge gives a false sense that you've cracked the code, that you finally nailed your creative or offer and it's happy scaling from here, when it's most likely just Q4 purchase intent, especially with highly gift-able products. The cliff is real: friends have gone from $250k+ a day down to $1k a day after Christmas. Stretch your current resources until after Christmas and see where demand settles. You don't want to be left standing with your dick in your hands.

## Chapter 04: Writing & Content

### Copy Craft Fundamentals

- **Read every piece of copy out loud after writing it: if you cringe or stutter, tweak it.** The out-loud pass catches sticking points that read fine on paper but sound unnatural spoken, which is exactly what matters when a creator delivers the line on camera. Copy that survives the test comes out more authentic and natural in video ads. Almost nobody does this, so the edge is cheap: it's easy to get ahead by being one of the people who actually try.

- **Study TikTok for formats and hooks; study Schwartz and Ogilvy for subject matter.** That split is the whole creative education. Delivery evolves on-platform: what stops the scroll today lives on TikTok, so learn it there. Persuasion doesn't evolve: desire, awareness, and what to actually say come from the classic direct-response canon. _(posted March 2025 · practice might be outdated)_

### Details That Read as Amateur

- **Capitalising every letter in headlines makes them harder to read, so write headlines in sentence case.** People recognise words by their shape as well as their letters, and all-caps destroys word shape: in the study he recalls, readers could identify shape-blocked words in lowercase but almost never in uppercase. People mostly skim your page and ads, so even a small increase in headline processing time costs you.

- **"1+1 sale" isn't a thing in English markets: it's "buy one get one free."** The phrase is a Dutch ecom tell no English brand would write. Localise your offer idioms, not just your translations - copy that's technically translated but phrased like no native brand breaks the illusion of a local brand and quietly costs trust and conversions.

- **"Save 30% off" doesn't make sense: it's either "Save 30%" or "Get 30% off."** The two constructions are mashed together, and the mistake is everywhere at ad-account scale. Micro-copy errors read as sloppy; picking one and republishing is a two-second fix that buys free credibility.

## Chapter 05: Monetization & Offers

### Spend Into Your Economics

- **Stop asking "how can I cut my CAC in half?" and start asking "how can I afford to double my CAC?"** The first question leads to endless creative and media-buying tinkering. The second forces you to fix the business model itself: AOV, LTV, and offer economics. It almost always produces the better outcome.

- **If your NCROAS is consistently above 2x, you're spending like a little bitch.** It's not that getting above 2x is hard: consistently sitting there means you're leaving scale on the table. With good unit economics and LTV you should never want it that high - push spend until efficiency drops toward your true break-even and take the bigger absolute number, because volume of profit beats percentage of profit. Jacob has never scaled an ad that stayed consistently above a 2x ROAS, while making millions.

- **The "$75 AOV minimum in 2025" rule is guru fiction.** Jacob's new US brand did $620k in its second full month at a $40 AOV with roughly 30% net margins, tracking to $1.2m in September. He'd already run a $30 AOV product to $615k/month in its second month just to prove dropshipping wasn't dead when tariffs first hit China, and the most impressive operator he knows in ecom is nowhere near $75 either. Stay off YouTube and podcasts for ecom benchmarks: most content creators have never seen inside the ops of a brand run by a degen ex-dropshipper or affiliate, so they don't know what's possible. _(posted September 2025 · practice might be outdated)_

### Charge for Shipping

- **Upping shipping price increased both CVR and RPV.** A 5% increase doesn't sound like much, but on $220k-$250k per month in shipping revenue alone it's a meaningful lift. Shipping is a price like any other, so split test it; most brands never touch it. And every time Jacob has tested free versus paid shipping, paid wins. Crazy that people run free shipping, never mind skip testing their prices. Moral of the story: charge for fucking shipping.

- **Jacob switched one market to free shipping and it became his worst performer.** CVR dropped; charging shipping again brought CVR, AOV and everything else back to normal. Charge for shipping, there are literally no downsides: CVR won't drop, AOV goes up, and you can leverage the threshold to incentivise higher AOVs. Drop your front-end price, add the difference back as shipping, and convert more people at the same revenue. You'll also capture more emails for abandoned-cart flows since the charge lands at checkout. "Free shipping" is really not the incentive for the customer you think it is.

- **For the love of God, charge shipping: it's essentially free AOV, and the second-order benefits compound.** A lower front-end price means a higher percentage of visitors hit checkout, so you capture more emails and more revenue from email flows. Triggering that lower-funnel checkout event also gets more profiles matched on Meta versus just a page view. And shipping is refund armour: the bigger the share of your price that is shipping, the less refunds hurt, because you don't refund shipping. The list is endless. _(posted September 2025 · practice might be outdated)_

- **Charge shipping, especially at low AOV: the math is brutal in your favour.** On a $24.99 supplement with $8 total COGS and 3% processing, $2.99 shipping moves margin from 65% to 77% and break-even ROAS from 1.54 to 1.29; $3.99 shipping takes margin to 80% and B/E ROAS to 1.24. That can be the difference between losing money and being profitable on the front end, and nobody who actually wants the product bounces for $3-4.

  Don't just bake it into the product price. The lower sticker gets more people to checkout, which means more emails collected. Sunk cost works for you: they've decided they want it and entered email and shipping info, so a small charge at that point won't put most people off. And shipping is a justifiable cost buyers understand, unlike a shipping insurance upsell.

- **Set your free shipping threshold just above your most common order value, not your AOV.** AOV might be $50 while the most common order is actually $30: a $35 threshold moves AOV more than the default AOV-plus-$5 at $55, because the mode is where most customers actually sit. Obvious in retrospect (h/t @jordanhill11), yet most people default to AOV math. Jacob implemented it on one store and it's having the desired effect.

### Offer Anchoring and Price Psychology

- **A no-brainer deal is engineered from the ad down, and 50% off isn't it.** Show creators in your ads all using a 3-pack; you'll even get complaints that you don't sell one, so everybody hits the site already wanting a minimum of 3. The base offer is a deliberately inflated anchor: a 2-pack at $39.99 priced at 90%+ margins, next to a 4-pack at $49.99. Almost nobody takes the 2-pack, and when they do the margins are silly. You could never sell that 2-pack alone at such markups because the perceived value isn't there, but as a quantity-break pair it outperforms selling either option separately. The 4-pack looks like great value purely because of the anchor. That's a no-brainer.

- **An offer that's too good destroys trust.** Jacob accidentally left a 100%-off automatic discount live, meaning everyone reaching checkout got the product literally free, and conversion rate dropped 46%. Customers price-check your credibility: when the deal is unbelievable, they believe the catch instead of the offer. Discount depth has a trust ceiling.

### Post-Purchase Upsells Are Free Money

- **In post-purchase upsells, don't try to convert non-subscribers onto a subscription: upsell a 6 or 12 month supply upfront at a discount.** The bulk supply will likely be worth more than the LTV of your average subscriber, and you collect the full amount on day one instead of waiting up to a year for a subscriber to drip it to you. Framed correctly, the take rate can be surprisingly high. Don't fall into the trap of assuming subscription is always where the money is.

- **Test your fucking post purchase upsells.** One week after setting up his post-purchase funnel on Shopify, Jacob was at $22,000 from 2,750 visitors and holding, with a split test on the one-click upsell tracking toward another $20 of revenue per visitor. The margin is free AOV: the customer has already bought, so there's zero acquisition cost on the incremental revenue, and the funnel isn't set-and-forget. _(posted May 2025 · practice might be outdated)_

## Chapter 06: AI as Leverage

### Automate Before You Hire or Subscribe

- **Learn N8N (or hire someone who knows it) and replace hundreds or thousands per month of SaaS with custom automations built exactly how you want.** Every Facebook ad launching tool either had big limitations or cost more than a VA, so Jacob built his own in N8N: ads dropped into a Google Drive folder called "to upload" auto-launch to Facebook, sub-folders become ad sets, files become ads, names inherit from folder and file, launched files move to an "uploaded" folder and Slack confirms with details. Next layer: an image-gen automation connected to a custom GPT that knows the brand and scrapes ad libraries, reviews, and Reddit threads, launching AI-generated ads on autopilot for roughly the cost of OpenAI credits. Jacob's verdict: N8N is the best value tool in the world. _(posted May 2025 · practice might be outdated)_

- **Flip the default from "I need to hire a..." to "I need to automate...".** Before any new hire or SaaS subscription, ask whether the job can be automated in-house instead: MCPs are coming to eat everything, and much of what SaaS tools charge monthly for is already automatable. The pace is the argument - nothing has advanced as fast as AI/MCP tooling did in a single month, and betting on automation over headcount is a bet on that curve continuing. _(posted April 2025 · practice might be outdated)_

### Prompting That Actually Works

- **The fix for gibberish on AI-generated product labels is one dumb line: "make sure there are no spelling mistakes on the label."** Works 95% of the time. Every sophisticated approach failed first: spelling out the exact label text, providing a separate image of just the label, even a JSON file describing the text, styling, colour, and positioning. Nobody knows why the dumb instruction beats the elaborate methods, but when prompting fails, naming the failure mode you don't want can outperform describing the output you do. _(posted July 2025 · practice might be outdated)_

- **Don't ask ChatGPT to recreate a reference image: ask it to write a prompt describing the photography style of a photo you like, then reuse that prompt as a styling guide for any product.** The extracted style prompt becomes a reusable asset across products and brands (Jacob demoed it on Hims, Nutrition Geeks, and Goli). His supplement example: editorial-style flat lay with top-down curated composition, hard lighting with defined shadows for drama and contrast, monochromatic background tied to product colours, texture and light play creating specular highlights for a tangible, premium feel. Results were amazing. _(posted May 2025 · practice might be outdated)_

- **End every AI copywriting prompt with "make this the best you possibly can with absolutely no room for improvement."** Jacob swears this 10x's the output every single time. No elaborate system prompts needed: one maximal-quality instruction at the end shifts the model's effort more than prompt length ever does. _(posted April 2025 · practice might be outdated)_

- **To replicate a brand's photography style with ChatGPT image gen, describe the subject in one sentence and let a full style guide do the heavy lifting.** Jacob got close to Dore and Rose's look with the subject "a woman with voluminous, wavy hair" plus a five-part style guide: Lighting (bright high-flash, 90s paparazzi / vintage Victoria's Secret, flattens shadows), Setting (luxurious vintage bedroom, ornate decor, "old money meets glam"), Styling (silk and lace lingerie, matching robe, deep red sleep mask), Mood (confident, relaxed, "rich girl in a hotel suite"), and Photography Genre Mix (boudoir + editorial + lifestyle/luxury fashion). Reverse-engineer any photography you admire into those five components. _(posted April 2025 · practice might be outdated)_

### AI Creative Production: Video and UGC

- **After multiple courses and dozens of guides, the best and easiest AI video ad workflow is just Google Flow with 2-3 lines of natural language, maybe a Nano Banana reference image.** First clip: a simple text-to-video prompt describing the character and what they say ("A middle-aged, well educated looking white man wearing smart casual clothing in a podcast studio talking on a podcast. He says 'My whole life I struggled with getting in shape, and now after 20 years of research I know exactly why that is.'"). Additional clips: save a frame from the original and use frame-to-video, telling it what the character says next.

  The 100-line JSON prompts on skin texture and 10-step multi-tool workflows are course-seller theatre: they don't change whether the video is good, the process got overcomplicated so people could charge for it. One exception: @maxxmalist's course is the best Jacob has bought for video creation, though the copywriting lessons could be better. _(posted January 2026 · practice might be outdated)_

- **AI UGC taking less time is a misconception: the wait is shorter but your hands-on time is far higher.** AI UGC means reference images, prompting, re-prompting voiceovers, wrangling character movement and camera, many manual steps per ad. Regular UGC is one message to a creator once the brief and script are done, then a payment when finished. Think stir fry vs slow cooker: the stir fry finishes faster but needs constant manual input, the slow cooker takes longer but frees your day. Judge creative workflows by your time input, not calendar time.

- **Anchor every clip to one saved frame.** For consistent VEO 3 characters (podcast ads etc): prompt in Google Flow, generate your first clip, save a frame from it, then use frame-to-video with that same frame for every subsequent clip. Keeps things 100% consistent and kills the character drift that breaks multi-clip formats. _(posted August 2025 · practice might be outdated)_

- **Jacob paid for Icon to test the hype: it doesn't do anywhere near what it claims. Avoid.** The "100 ads a week" promise delivered 3 ads in 2 weeks, all bad. "Launch ads with one click" is a blatant lie, with huge limitations on what you actually control when launching, and the team no-showed the onboarding call. Judge AI ad tools by shipped output, not demos. _(posted May 2025 · practice might be outdated)_

- **Don't build your creative pipeline assuming human UGC pricing and scarcity hold.** AI creator tools like Arcads have reached mind-blowing quality, Creatify isn't beating around the bush either, and UGC creators don't realise how limited their days are. Price and structure creative production expecting AI creators to keep closing the gap. _(posted October 2024 · practice might be outdated)_

### Agents That Run the Business

- **Before launching any product, build a master research document with Claude and work it systematically until Meta tells you what to double down on. "This methodology has never failed us to this day."** Step 1: one massive Claude prompt assessing market sophistication, competitors, reviews on Amazon, Reddit, and Trustpilot, common complaints, possible ad angles, estimated market size, and pricing strategy; the output goes into a Claude project alongside a summary of Breakthrough Advertising and a doc on maximising hook and hold rates.

  Step 2: ask the project for the top 5 audiences in priority order, each with a description, main pain points, and core desires. Step 3: per audience, rank which awareness levels to target first (some audiences start product-aware, others problem-aware). Step 4: get the best ad angles for each awareness level of each audience. Step 5: make ads for the top 2 audiences x top 2 awareness levels, usually statics first in as many formats as possible, leaning on what already works for other products and knocking competitors on their known complaints. Step 6: launch at $100/day for signal, not profit; double down on what Meta favours, adapt the landing page copy and images to the winning angle, take working angles to UGC and video production, and if nothing sticks, pull more angles and audiences from the research and keep working through it. _(posted June 2026 · practice might be outdated)_

- **Commslayer MCP + Claude made customer service the most AI-revolutionized part of the business, and the AI agent's CSAT beats every human agent.** Jacob is confident the setup can resolve 90%+ of tickets. The self-improvement loop is one prompt: ask Claude to look at Commslayer's last 30 days, count tickets the AI agent resolved, surface the most common escalation reasons, and update the agent's guidance and actions so she resolves those herself. Claude did the analysis and the updates in about 5 minutes, unlocking roughly 19,000 more AI-resolved tickets per month. Audit escalation reasons, patch the guidance, repeat. _(posted May 2026 · practice might be outdated)_

- **Any tool whose job is moving assets between your systems is now automatable with a general agent.** Jacob's Manus automation checks his Notion creative database daily for new ads, downloads the creatives from Frame, uploads them to his Facebook testing campaign, then marks each one "uploaded" in Notion with the date. Runs automatically at midnight every day. A general AI agent replaced a dedicated paid tool: "Bye bye Adnova I guess." _(posted March 2026 · practice might be outdated)_

- **Brief the agent with your top ad, not a blank prompt.** Jacob gave Manus a link to his top-spending ad and one listicle URL he liked. It analysed and transcribed the ad so the lander copy was congruent with the traffic source (the thing that actually converts cold traffic), one-shotted the listicle format in HTML, generated the images with Nano Banana Pro unprompted (no API key or Google account given), pushed the page to a Shopify draft theme once granted access, then re-coded the one-block page into editable theme-editor sections, each step one-shotted. All CTA and button links checked out correct, and it even placed his site's new-customer discount code in the announcement bar on its own. _(posted March 2026 · practice might be outdated)_

- **Export your Meta campaign data and let a reasoning model hunt for correlations you'd never eyeball yourself.** Jacob downloaded the raw data from his top spending campaign over the last 30 days and asked ChatGPT o3 to find whether any creative metrics correlated with ROAS and spend, not to summarize. It generated charts automatically and turned his own account into a research dataset: correlation analysis on actual spend instead of guessing which soft metrics matter. _(posted April 2025 · practice might be outdated)_

## Chapter 07: Mindset & Execution

### The Inner Game

- **Genuinely believe everything you attempt will be massively successful.** Not optimism, identity: that you are eternally lucky, blessed by God, favoured by algorithms, carrying an x factor nobody, including you, can explain. It is a self fulfilling prophecy - the belief changes how you attempt things and the attempts confirm the belief.

- **The most productive 48-hour cycle: work through the first night, sleep around mid-day, wake around 8pm fully rested.** That buys 12 straight hours of total silence while everyone else sleeps: no Slacks, no WhatsApps, no emails, no calls. The silence is the mechanism - not more hours, but relocating your working hours to where interruptions cannot reach.

- **One of the best things a founder can do is go on a full information diet.** Podcasts, YouTube, Twitter: mostly bs that distracts you and causes analysis paralysis. Only a few core things move the needle in any business and you already know what they are, so get the reps in on those. There is no point overloading yourself with new info before you have internalised and mastered the basics you were already given.

- **Actively trying to get rid of stress strengthens it.** Stress is fear-based and the foundation of fear is avoidance: even temporary relief from avoidance tactics reinforces the feedback cycle, so you are literally making it worse. Do the opposite - when stressed or afraid, actively try to prolong it and make it stronger, then let it burn out on its own. What you resist will persist; handling stress this way is a real operator advantage.

- **Your attention is the only valuable commodity you have. Guard it and deliberately seek out boredom.** Multi-billion dollar companies were built from stolen attention and keep fighting over it; without conscious intervention you are fighting Goliath and you will get wrecked. Curate your environment to make distraction hard, seek the boring and un-stimulating, and start small: the 10 minutes in the shower you would fill with music, let your mind wander instead. Boredom is where the work happens - ideas come flooding, problems get worked out, creativity blossoms, and you build intuition the infinite pit of hyper-stimulating scrolling is eating away.

### Just Hit Publish

- **$100k in a month proves everything you need to do $100k in a day.** Your customers exist, you acquire them efficiently, your website and ads convert; factors like Meta delivery efficiency play a part, but the main gap is having the balls to up the budgets. I have fundamentally changed my business and brought it to new heights numerous times from simply saying "fuck it" and raising budgets when the numbers did not justify it, with no other changes. The nerds will never understand this.

- **Reviewing ads is probably the biggest waste of time in ecom.** Guys get assets delivered and spend hours on "can you change this font size" and "can you move this overlay 1mm to the left," delaying the only feedback loop that counts: spend and performance. Just hit publish and let the market review it.

- **If you could click your fingers and the thing was ready to launch, would you launch it?** If yes, all your doubts are just disguised laziness. When the only thing stopping you is the work itself rather than the idea, the hesitation is not strategic caution.

- **The finger-click test exposes cope.** When deliberating a brand idea, ask: if you could click your fingers and the store was built, creative assets complete, product ready, ads made, all you need to do is hit publish, would you do it? If yes, every "what if" is just cope for not wanting to do the work. Stop deliberating hypotheticals and test - all my lucky breaks came from ploughing ahead without a second thought.

### How Real Operators Work

- **The alpha is always in doing the opposite of the masses.** How do you expect to get extraordinary results doing the ordinary? By definition you need to be doing what others aren't. The creative volume meta is the case study: everyone preached volume, 1000 active ads became the norm, nobody had an edge, and quality collapsed in the volume chase, so those ads get raped in auctions by people who researched, planned and directed theirs.

  I deliberately dialled volume back further as others abandoned quality, widening the delta between my ads and the competition. My 5 ads a week beat competitors publishing 300+, and we scaled a new brand from zero to 7 figures a month in its 3rd month. Where the market moves, move the opposite direction; the sheep are exit liquidity.

  _(posted January 2026 · practice might be outdated)_

- **The average 19-year-old dropshipper has personally reps'd the entire ecom skill stack while the average 30-year-old "brand founder" lent money and paid people.** The stack: 100+ products tested, 50+ stores built (stuff agencies charge $50k+ for, done in a day), Breakthrough Advertising known by heart, thousands of hours of graphic design and video editing, thousands of customer service tickets handled personally, ad accounts scaled better than paid professionals, lean cashflow management with no funding. Give those kids the capital and credit access startup founders get and they would wipe the floor with them. The leap most never make is reinvesting profits into building something long term; the ones who do are running rings around your favourite Twitter gurus.

- **Money in your bank account is the only metric that really matters.** You raised AOV? Cool story bro, how's your bank account? Increased signups, grew creative output 200%? Same test. Don't complicate the scorecard into 1000 shades of grey with proxy metrics: it's black and white if you let it be.

- **Being a good technical marketer matters less and less.** Complex data, deep cohort analysis, nerdy attribution work: all that nitty gritty pales in comparison to being a good copywriter who can make genuinely good ads. So few people are actually doing that, which makes copy and creative the higher-ROI skill to stack.

  _(posted August 2025 · practice might be outdated)_

- **Two bootstrapped 9-figure friends I talk to daily share the same traits: always the student, humble, fast, always testing, working a plan.** They are sponges despite crushing it, asking advice even from people well below their level, taking the meat and leaving the bones anywhere they can. They move ridiculously fast: in the time it took me to hire a graphic designer, one had planned and started a warehouse expansion based on demand from the previous weeks. They experiment heavily and accept losing money doing it, knowing the alternative is stagnation and the death of the business.

  The last trait is a clear long-term plan. Knowing exactly what path they are on lets them zoom out and course-correct when things go wrong; being reactive leaves you stuck asking "well, what do I do now?" and repeating the same stuff on a loop.

- **A rough Meta market is not an excuse to stop attacking every lever - here is one week's output.** Around 600 new ads published, 2 new winning landing pages found, 2 new ad accounts and 2 new pages tested, 2 ad accounts restructured (big difference). Margin got attacked too: boots on the ground in China for new suppliers and lower COGS, shipping renegotiated for a 15% reduction. Diversification ran in parallel: a wholesale order secured, 7 new affiliate creators onboarded, an Amazon listing live, a new photoshoot arranged.

  _(posted May 2025 · practice might be outdated)_

- **A Kalodata rip can get your first bag, but understand you still know nothing about ecom.** You made money copying and pasting: you can't make ads, build and manage a team, or do retention, and you don't even understand why the thing you sold works, just that it did numbers on TikTok. Most of these operators own none of their assets and are one defenceless DMCA away from broke. Don't blow the money on cars and a fancy apartment - stay in your parents' house or a cheap place, use it to buy your time and freedom for the next year, and work out how to build something legit that doesn't rely on another brand paying affiliates for your success.

  _(posted April 2025 · practice might be outdated)_

### Filter Your Inputs

- **Pretty much all the alpha in ecom comes from the few, not the many.** When the trend says everyone is down, study the people swimming upstream and figure out what lets them scale while others stagnate: far better to ask questions of someone at $1m/month growing 200% MoM than someone doing $20m/month and declining. Managing more spend gives someone a good read on the general landscape, but it doesn't mean others haven't cracked the platform right now, and dismissing someone because they challenge a big reputation or a bigger spender is how ideas go stagnant. Question everything no matter who it comes from, test it yourself, listen to everyone: take the meat, leave the bones.

- **No amount of guru "sauce" will ever equal funnel hacking successful brands.** The winning playbook sits in public view: the ads, landers, and offers of brands already at scale. Study those instead of buying courses.

- **The people with time for consistent content output usually aren't the experts.** If they were, they'd be doing the thing instead of talking about it: the people talking are usually larping, the people doing are too busy doing. That's why everything from "thought leaders" is old, rinsed, needlessly complicated, or flat out doesn't work. The best education is doing anything to work for or help someone actively running the business model you want, and running it well.

- **Many of the massive revenue screenshots you benchmark against are black hat.** Cancer claims, hidden re-bills, obscured "VIP membership" rebills: tactics a legitimate operator won't touch sit behind some of the loudest numbers on the timeline. New operators see those numbers, can't get close, and conclude they have a skill issue when in reality it's a morality issue. Filter who you let preach at you before you copy anyone's playbook.

- **Most free content is made to sell, not to help.** Viewers get the advice paying the creator the highest commission, and end up with bloated SAAS bills, overly expensive partners, and no idea better or cheaper alternatives exist even when the creator knows about them. The filter: never listen to anything promoted on a podcast, rarely trust referrals from agencies (they are all in bed with each other and their SAAS partners), and be very sceptical of anything pushed by large accounts, especially when co-ordinated across several - that's a paid campaign, not an endorsement.

- **Most DTC misinformation is misaligned incentives: run everything through the filter of who's saying it and what they gain.** An agency's objective is maximum money for minimum work, so it will never promote anything requiring more manual hours and less margin. An info seller's objective is convincing you their strategies are right and everything mainstream is wrong, and when a better way appears they are too invested to concede. Brand owners should mostly listen to other brand owners: you rarely see them arguing over creative testing or cost caps with each other, they just want whatever works, while agencies are ready to die defending their philosophies.

### Your Circle Is the Asset

- **Agency sales calls are a leak vector for your playbook: reveal as little as you can.** Multiple times Jacob has done a call, seen his store cloned within days or weeks, and traced it back to friends of the agency owner. Find your trusted group of proven partners and stick with them instead of shopping your secrets to every agency pitching you.

- **Get around people better than you, and when the chance comes to help them, refuse payment.** Anytime you think you know a lot, talking with people ahead of you reveals things you wouldn't have considered in the realms of possibility, so do whatever it takes to work with them, do favours, build the relationship: the opportunity is the payment. Most people miss these openings by trying to extract something from every interaction. Jacob repeatedly declines offered payment ("happy to help, no need to pay me") and it keeps turning into something massively more valuable than any fee. Give, deposit good will, ask for nothing. You reap what you sow.

- **Skill is the magnet, not outreach.** Skip the networking events: get very good at what you do and watch your network magically appear. The people worth knowing seek out demonstrated competence, and it's the only networking strategy that compounds.

- **When someone further ahead helps you, figure out what you can bring to the table for them.** They have more to offer you than you them, so find something anyway: it has served him very well. Copying your mentor instead is a four-way loss: you can't compete with them, they'll never help you again, you traded a long-term relationship that mainly benefited you for short-term cash, and you never actually applied what you were taught. Jacob has had 4 people he mentored 1-on-1, shown the brands, ad account, and backend, decide the best move was to copy him.

- **Don't do business or share sensitive information with anyone financially worse off than you.** When they see a chance at quick cash it's blood to a vampire: morals get abandoned, and like a junkie needing a fix they'll beg, borrow and steal to get it. It's ridiculous how many "friends" get caught trying to screw you over for short-term money. Trust is a function of what betrayal would cost them. Keep your circle water-tight.

## Chapter 08: Tools & Tactics

### Research Shortcuts: Steal Validated Demand

- **Tagged posts are a pre-vetted creator directory.** Go to your favourite brands' Instagram profiles and open their tagged posts: anyone tagging a brand has already proven they make brand-friendly content in your niche, no platform, agency, or cold outreach list required. Raid brands you admire rather than competitors' follower lists, because their tags surface creators whose style already matches the quality bar you want.

- **Let mega-influencers fund your demand generation.** Keep an eye on what the Kardashians and Jenners are selling, then offer a cheaper alternative: they already paid for mass awareness, the most expensive part of marketing. The play only works with a real branded approach, never a bare knockoff.

- **Expensive agencies pre-filter winning products for you.** Browse the case studies and client lists on the websites of agencies you know charge big retainers: any brand that can afford them is almost certainly doing real volume, so their public client pages are a free directory of validated products and niches.

### The Stack That Actually Wins

- **Distrust promoted tools by default: too much software gets attention purely because big names take backhanders to push it.** The unsponsored, unaffiliated stack that actually earns its keep: Commslayer (customer service), Section Store (store building), Judgeme (reviews), Atria (swipe files, creative analytics, asset management, AI image ad gen), Upstack Data (first party tracking), Triple Whale (UTM tracking, profit tracking, AI agents), Archive.com (social listener), Seal (subscriptions, great for lower volume), Loop (subscriptions), Track123 (parcel tracking and PayPal tracking sync), Aftersell (post purchase upsells), Disputifier (chargeback management), Kaching Bundles (bundles and offer testing), Intelligems and Elevate (split testing), Shopify Collabs (affiliates), Google AI Studio (video ad analysis), and Heatmap.com (amazing value, probably the most underutilised ecommerce software on the market). Klaviyo deliberately excluded: still in use, but not worth the money and "needs disrupted badly." _(posted July 2025 · practice might be outdated)_

- **Shrine Pro still always wins.** After testing every theme imaginable, custom builds included, the results keep coming back the same: "I don't like that it's true." Don't sink money into bespoke storefronts before testing the boring winner; if a widely-used conversion-focused theme outperforms custom builds, spend that budget elsewhere. _(posted May 2025 · practice might be outdated)_

- **Not a single Replo product-page template beats a standard Shopify theme template.** Even the "CRO Top Pick" pages could be rebuilt by a 13-year-old Shopify newbie with default sections, and the big-brand templates are a trap: they are pulled from brands like GymShark and Baggu, which often have the worst websites. Pre-made templates should let you quickly do things you couldn't do in Shopify; if they don't, run Impact theme supplemented with Section Store instead. _(posted September 2024 · practice might be outdated)_

### Run the Numbers, Then Play Hardball

- **When platform support stonewalls you, CC your lawyer into the escalation email and name the regulator.** After weeks of back and forth with Shopify support over frozen payouts, asking the agent to copy the lawyer on the escalation email to the payments team produced the response waited on for weeks within minutes. Make the paper trail the threat: state you're not leaving the chat until the escalation email is sent with your lawyer copied in, otherwise all chat logs go to the financial ombudsman. And verify the email actually gets sent, because the first "escalation" never came.

- **You can't diagnose creative or funnel problems from CPA alone.** Set up custom metrics in Meta ads manager for CVR, AOV, thumb stop ratio and hold rate: Columns > Customise columns > Create custom metric. It takes minutes, permanently upgrades every report you look at, and it's amazing how many people run accounts without them. _(posted October 2024 · practice might be outdated)_

- **Worldwide campaigns force one CPA target onto unequal countries.** A country where you can afford a $50 CPA sitting in the same campaign as one where you can only afford $40 means you're leaving serious spend on the table. The fix: download the last 60 to 90 days of invoices from your fulfilment company, compile into one spreadsheet, upload to ChatGPT and ask for average shipping cost per country, then do the same with 60 to 90 days of orders to get AOV and items per order per country. For each country subtract average shipping cost and product cost from AOV, then build campaigns around groups of countries with similar margin, not geography. _(posted August 2024 · practice might be outdated)_