The ecom scaling framework.

An ecom business has four levers — most operators only pull one.

Market. Product. Ads. Funnel. They sit on a slider, not a list. If your product is a 10/10, mid ads and a mid funnel still print money. If your product is a 5/10, ads and funnel have to be cracked to compensate. Most operators try to fix every problem with a product change. Wrong lever, every time.

It's not the product. It's the audience. They sold to the wrong people — usually no people at all — with the wrong message at the wrong moment. If you sell to everyone, you sell to nobody.

The frame: don't look for a winning product. Look for a frustrated audience with money who has already tried other solutions that didn't work. That's the most valuable position in ecommerce. Everything else stacks on top.

Find the gap before you find the product.

We use Claude to dig into Amazon 2-star and 3-star reviews, Reddit threads, Quora questions, and TikTok comments — pulling the language frustrated buyers actually use. We're not looking for product ideas. We're looking for the gap between what exists and what people actually need.

You're looking for the complaint that keeps appearing across multiple platforms, from people who already tried to solve it, where the current market response is weak, generic, or overpriced. That intersection is the gap.

Once you have a gap, validate the product fits.

A good product passes a short checklist before you ever launch an ad. Skip these and you'll find yourself fighting structural problems no creative can fix.

Offer = People + Perceived Value + Packaging.

If you only remember one thing, remember this: a banger offer is the foundation. Ad creative, landing pages, retention flows — all of it lives on top of the offer. If the foundation is weak, nothing built on top of it holds up.

Use the value equation. Strong offer = high dream outcome × high believability, divided by low time delay × low effort. Optimize each variable until the customer can't say no.

Avatar work — research before you write.

Before you write a single ad, build a single landing page, or design a single funnel — get inside the head of the person you're selling to. Not the demographic. The actual person. Their language, their fears, their failed attempts, their internal monologue at 11pm.

This is the upstream work that determines whether anything downstream converts. Most brands skip it and try to compensate with creative volume. Volume doesn't fix a misread avatar.

Run this once per category. Save the output. Every ad, lander, and email script for the next 12 months pulls from this document.

The avatar swap is the fastest profit lever you have.

Same product can serve completely different people. The avatar you pick determines the price you can charge, the LTV you can extract, and the CPA you have to live with.

Take a sleep supplement. Sold to anxious 25-year-olds, it's "a melatonin alternative" — competitive, low-margin, tough CPA. Same product, repositioned for stressed-out parents of newborns? It's "the 90 minutes of extra sleep that saves your marriage." Wealthier audience, longer LTV, far less price-sensitive.

Build a dedicated PDP, dedicated ads, dedicated messaging tuned to the second avatar. The product didn't change — the person buying it did. The pattern repeats across categories. Lower NCPA. Higher LTV. Often a 2–3× lift in net profit. A better avatar is worth more than a better creative.

Most operators waste weeks debating whether a creative is "good enough" to launch. Wrong question. The market decides what's good — not you. Your job is to put it in front of the market quickly, read the signal, and act.

The 72-hour read.

Three signals. That's it. Don't overthink it.

End of day 3, ask one question: which 2–3 creatives produced purchases at or below target CPA? Those are winners. Cut everything else. No sentiment. No "this one looks good." The numbers decide.

Quiz funnels — how to make people validate themselves.

People buy when they feel understood. A quiz is the fastest way to make someone feel understood at scale.

The mechanism: imagine walking into a doctor's office, sitting down, and the doctor immediately says "you have X, here's the prescription." You'd doubt it. Now imagine they take your blood pressure, ask about your symptoms, look in your ears, and then say the same thing. Total trust. The diagnosis is identical. The journey changes everything.

Build it on Heyflow — skip the email gate.

• Use Heyflow as a bridge page only — not a hosted store
• Remove the default email opt-in step (kills completion rate)
• End the quiz with a redirect to your dedicated PDP
• Each question should be tappable buttons, not text fields
• Total quiz length: 8–12 questions. Less than 8 feels shallow. More than 12 drops off.

Once you have one validated avatar, the next test isn't a new product — it's a new avatar for the same product. Spin up a dedicated CBO + dedicated lander with messaging targeted to the new avatar. If CPA holds, you've found a second audience. Same product, multiplied revenue.

The structure is boring on purpose.

• One CBO per business objective — one product, scaling across the big 5 markets (US, UK, CA, AU, EU)
• One ad set per ad idea
• Three creatives per ad set — same idea, different visual hook
• Broad targeting. Open audience. Let the algorithm do its job.

Two scaling modes — Default & Surf.

Default — the 20% SOP. Five minutes a day. That's the entire scaling protocol.

• Did we hit yesterday's KPI (your acceptable CPA)? Yes → increase CBO budget by 20%.
• No → wait. Do nothing for 3 days. The algorithm self-corrects 80% of the time.
• Still under target after 3 days? Decrease 20%. Stay calm. Cut creative, not budget.

Surf — when momentum is real. Hook rate above 35%, CPA crushing target by 30%+, ROAS climbing day over day. That's a wave. Don't tap it for 20%. Ride it.

• 50–100% same-day budget jumps for 2–3 days while the signal is hot
• The moment KPI starts drifting — back to default mode, no exceptions
• Surf scaling captures momentum the algorithm has already validated. The risk is not surfing when it's there, not surfing when it isn't.

Most operators waste two hours a day "checking ads" and tweaking. The 5-minute SOP is enough. Fewer decisions = better decisions.

TOF · MOF · BOF — the three jobs of a creative.

One CBO doesn't mean one type of creative. Inside the campaign, you're running three distinct jobs at once.

• Top of funnel — unaware / problem-aware. Hooks like "The 3pm crash isn't normal" or "If you're over 35, you've felt this." Connect with the feeling first, introduce the category second.
• Middle of funnel — product-aware, needs proof. Hooks like "5 things nobody tells you about [category]" or "Why we don't use [common ingredient]." Pattern-interrupt with contrast, then deliver mechanism + risk reversal.
• Bottom of funnel — sold but waiting on price. Hooks like "We made too much — please help us empty our warehouse" or "Restock dropped. 48 hours only." Only works once you have brand presence. Don't run BOF on day one.

The mix that's working in 2026.

• ~50% static — half educational infographics, half direct-offer statics
• ~50% video — UGC, founder talking-head, listicle, white-label collaboration posts
• ~50% of video output should be collaboration ads from creator accounts — not your own brand page

The biggest unlock right now is the white-label / collaboration ad. You post the ad from a creator's account with your brand as a collaborator. Native, trusted, piggybacks their audience. Most brands are running zero. It's the fastest CPA reduction available.

Most brands stay small because their creative pipeline is broken. They wait on freelancers, go back and forth on deliverables, spend $3k and get six creatives in three weeks. Meanwhile the brands scaling past them are shipping 50 ads a week.

The asymmetry is the entire game.

Morning — scripts with Claude.

Run this four times with slightly different problem framings. Two hours in, you have 35–40 scripts. Cut to the 10 sharpest hooks — the ones that make you stop mid-scroll when you read them out loud.

This is a starter prompt. Operators inside our community work from a deeper library — gap research, full avatar pipelines, hook generation, advertorial scaffolds, post-purchase flows. Start with what's above and you'll get 80% of the value.

Afternoon — visuals with Higgsfield + ChatGPT Image 2.0.

Higgsfield handles cinematic product footage and lifestyle sequences. One morning of work adds 20+ high-fidelity variations to your batch — Tier-1 agency production value at near-zero cost. Use Nano Banana Pro for the photo-realistic statics.

ChatGPT Image 2.0 (in thinking mode) handles photorealistic statics with rendered text — the format that's quietly outperforming AI video right now. The workflow that ships:

• Brief generation in Claude — feed full PDP screenshot + product photos as context
• Claude writes the prompt — use case, format, scene, subject, camera, lighting, styling, text, constraints
• Render in ChatGPT Image 2.0 (thinking mode, not instant — this matters)
• Feed result back to Claude for critique
• Regenerate with refinement — usually one revision is enough

The 80 / 20 iteration loop.

Once you have a winner, replicate the angle across every format and hook variation you can generate. From one winning creative, extract 12–15 variations inside 24 hours:

• Rewrite only the first 15 words — test 5 different hooks on the same body
• Cut the same script as a square format with B-roll behind it instead of talking-head
• Add the hook as bold text overlay in the first 3 seconds, remove the spoken intro
• Localize for a new country — run the same script in a different language
• Cut the best 10 seconds as a short-form version with a harder CTA

80% of new creatives iterate the winning angle. 20% test brand-new angles. The 20% means your next winner is already in testing before the current one fades.

Retention isn't only subscription.

For consumable categories — supplements, beauty, food — subscription is the architecture. The big greens-powder brands all run the same play: subscribe is the obvious choice, one-time is intentionally worse. 50%+ off + stacked free gifts on the subscribe path. One-time purchase priced $20+ higher with no extras. The math closes the decision before the customer consciously makes it.

For non-consumable brands — apparel, accessories, lifestyle, beauty drops — subscription doesn't fit. Brands like Nuamore retain through other mechanics:

• Frequent collection drops — new items every 4–6 weeks keep returning customers in the loop
• Bundle offers — reward AOV at checkout, increase basket size on first purchase
• VIP / loyalty tiers — early access, free shipping, exclusive items for repeat buyers
• Email + SMS storytelling — keep customers in the brand world between purchases

The mechanism is the same: design the post-purchase journey so the customer stays in your world. The execution differs by category.

Gamified popups — the new conversion lever.

Static popup forms are cooked. The mechanic that's working in 2026: gamified popups — Alia, Spin Wheel, Scratch Card. The user reveals their reward instead of being handed it.

• Visitor lands → popup says "scratch to reveal your offer" or "spin the wheel"
• They tap once → reward is revealed (welcome offer, free shipping, gift, percentage)
• The reward feels earned, not given. Engagement bias kicks in.
• Email and SMS opt-ins flow naturally after the reveal

Same offer. Different psychology. Earned discounts feel different from given ones — and the conversion gap is real. We've seen 2–3× lift swapping a static popup for a scratch-card flow.

Don't stack a second discount inside the popup. The acquisition offer already lives on the site. The popup gates it behind opt-in. Most brands quietly destroy their margin layering 10% off on top of an existing welcome offer.

Welcome flow density — day one is the warmest they'll ever be.

When someone signs up after clicking your ad, they are the warmest they will ever be. Every day after, they only get colder. You have a 24–48 hour window to convert before someone else does — Meta is feeding that customer your three closest competitors the moment they click.

The cadence: 3–5 emails on day one. Daily SMS for the first two weeks. Then daily email campaigns ongoing. That sounds aggressive. It works.

• Educational infographics over walls of copy
• Social proof in every email
• Offer in the hero of every email — never make them scroll for it
• Mix in occasional founder text-only emails — they crush, because they look human

The daily SOP — five minutes, not two hours.

• One question every morning: did we hit yesterday's KPI?
• Apply the scaling rules from §03 — Default 20% or Surf when momentum is real.
• Done. Move on to creative production, supplier ops, B2B.

Most operators waste hours daily refreshing the dashboard. Five minutes is enough. The rest of the day is for building the next thing — not babysitting the current one.

B2B is the cheat code most brands ignore.

D2C is volatile. Ad accounts get banned. CPMs spike. Attribution breaks. B2B is forecastable revenue with higher margins and smoother volume.

Once your brand has any presence, B2B starts falling into your lap — corporate gifting, hotels, spas, boutique retailers, subscription-box partners. It's also a marketing channel: someone sees your product in a hotel spa, sees an ad two weeks later, converts on D2C. Omnichannel compounds. Single-channel plateaus.

Ouassim — $52,000 in 7 days.

Ouassim came to us with a product and an audience but no real structure underneath it. No creative testing system. No way to diagnose what was breaking when ads went sideways.

We didn't find him a magic product. The framework made every decision faster and every test smarter. Within seven days he was running healthy profits with a system he could repeat.

Ouassim — $52,089.50 in 7 days, healthy profits returning

What our own ops look like.

Yesterday — Order Revenue $130,794, up 11.78% vs previous period

$3.3M months. $100k–$140k days. Six years operating. Real brands, real supplier problems, real ad accounts, real chargebacks. Not rented lifestyle — running operations.