Ecom Khalifa

Doing Ecom For A Living

[Ecom Khalifa @EcomKhalifa]

@GlennNieuwenh @winninghunt @scalesurfer_

[Ecom Khalifa @EcomKhalifa]

Just launched Cutroom on Tiny Startups 🚀

Upload one take. Cutroom cuts the dead air, puts b-roll on your exact words, writes the hook and times the captions. Your first video is free, no card.

Listed here: https://t.co/PJHvgDi9eK

cheers @ratheejaisal for the listing

[Ecom Khalifa @EcomKhalifa]

Take notes because this will probably change your Meta performance if you're still dealing with insane CPMs and day-to-day volatility.

Stop starting with 30 random creatives.

Start with the lowest hanging fruit:

Your product.

We write down the 10 biggest objections someone has before buying.

Price.
Trust.

“Will this actually work for me?”
“Why this instead of the thing I already use?”
“Why now?”
“What makes this different?”

Then we make one product-focused creative that kills each objection.

10 objections.

10 ads.

We run all of them.

We use HeyGen To Create 10 spokesperson videos of different Avatars.

CutroomAI handles the editing so turning 10 scripts into 10 finished creatives isn't a week-long production problem anymore.

Then we stop guessing.

The objections that get traction tell us what the market actually cares about.

And this is where most brands get it wrong.

They find a winning objection and make another version of the same ad.

We go one level deeper.

If “why is this different?” is working, we build a VSL entirely around why it's different.

If skepticism is working, we build the next VSL around proof.

If price is working, we go deeper on value and alternatives.

Same objection.
Deeper argument.
Better proof.
Longer format.
Then repeat.

Especially in the US, where CPMs keep getting more painful, you cannot afford a creative strategy built around “make more ads.”

You need a system that tells you what to make next.
Product → objections → test → winner → deeper VSL → repeat.

That's the whole game.

Most brands don't have a creative volume problem.
They have no idea what their next creative should be.

[Ecom Khalifa @EcomKhalifa]

@ssaaammiirr you know there is not much capital when you say you have a shit ton of capital.

[Ecom Khalifa @EcomKhalifa]

@ecommilan 😂😂😂

[Ecom Khalifa @EcomKhalifa]

EDITORS ARE COOKED.

We’re using software now to edit basically ALL our reel content.
VSLs.Spokesperson ads.UGC.Reels.B-roll.
The whole lot.

Faster. Cheaper. More consistent.

And honestly, the output is getting better than what we used to get from editors manually doing the same shit for hours.

If you want access, reply 'EDIT'.

[Ecom Khalifa @EcomKhalifa]

@ecompixsil cola or pepsi

[Ecom Khalifa @EcomKhalifa]

@conortrains Motion is good for this

[Ecom Khalifa @EcomKhalifa]

@alexpgecom why would I? not needed

[Ecom Khalifa @EcomKhalifa]

@abdushodmonov Alhamdullilah

[Ecom Khalifa @EcomKhalifa]

@VinceNijhof Yapping ad

[Ecom Khalifa @EcomKhalifa]

People keep asking:

“How do I lower my CPM?”

Here’s the answer from someone who has spent well over $20M on ads in the last 2 years. My own money, not agency spend.

You don’t lower CPM by fucking around with audiences, placements, bid strategies or whatever hack is trending this week.

You lower CPM by making Meta want to buy you cheaper traffic.

How?

→ Change your offer
→ Increase your CVR
→ Improve your funnel
→ Give Meta better conversion data
→ Create a business that converts better than the competition

Meta is constantly learning which advertisers generate valuable outcomes.

If your traffic converts like shit, Meta has less reason to send you cheap impressions.
If your offer converts significantly better, Meta sees the downstream signal.
Better conversion = more valuable traffic = potentially lower CPM.

And this is why I would NOT recommend obsessing over CPM when you're spending.

CPM is a symptom.

Your offer, CVR and economics are the disease.

Stop trying to hack the auction.

Build something Meta can actually make money from.

[Ecom Khalifa @EcomKhalifa]

@hikmatsharifi1 be aware that it takes times to crack this...

[Ecom Khalifa @EcomKhalifa]

@kdseifu better ads, more social proof, cpa increase till it it becomes uncomfortable for your competitors

[Ecom Khalifa @EcomKhalifa]

@TheBasharAlhajj 🤣

[Ecom Khalifa @EcomKhalifa]

@mannybarbas_ nah

[Ecom Khalifa @EcomKhalifa]

@LazyAdsAI what do you mean?

[Ecom Khalifa @EcomKhalifa]

HOW I SCALED A BRAND TO $2.4M/MONTH IN MY SECOND YEAR OF ECOM

The exact playbook, in the order I'd run it again

In year one I believed three things that nearly killed the company.

That the right campaign structure would unlock scale. That a higher AOV would fix the margin. That more creative volume would fix the account.

All three are half-true, which is what makes them expensive. Structure matters a little. AOV matters a little. Volume matters a lot, but only if you know what you're feeding into it.

What actually took us from a few hundred thousand a month to $2.4M was six systems, built in a specific order, each one making the next one cheaper to run: offer, creative, media buying, landing pages, email, and a layer I think of as the extra 30%.

Every principle here cost me money before it made me any. Examples are from our brand and categories I know well; the mechanics transfer.

PART 1 — THE OFFER

→ Price is a conclusion

The first thing I got wrong was treating price as a number to optimize. It isn't. Price is the last decision your customer makes, and by the time they make it, the outcome is mostly decided by everything that came before.

Apple doesn't sell you a $1,500 phone. Fifteen years of owning one, everyone you know owning one, and two decades of marketing do. The price just closes the file.

So the question was never "what should I charge." It was: what has to happen before the price appears so that, when it does, it reads as the only reasonable ending? Which questions need answering, which objections need to be dead, which proof needs to be seen, and in what order?

When the justification is built properly and the offer makes sense for both sides, price stops being a limiting factor. That's the whole definition of an offer.

→ Sequence it so both sides get their no-brainer

Say you sell a supplement at $80 a bottle. That's the price your economics need. Your customer is a 58-year-old who has tried everything, and your product is five times the category average. Will he try it? If you manipulate him, maybe. But a manipulated customer doesn't stay, and you need him to stay.

Here's the sequence we ran instead. A video ad that opens with the mechanism and why it's different, then stacks proof from the brand and from third parties: customers, a science board, whatever you have. The landing page it drops into kills every remaining objection before the price is visible. Then the offer: a low-ticket trial pack, which after everything he just consumed is a no-brainer. Seven days later it rolls into an $80 monthly subscription.

The trial is the part that makes sense for him. The subscription is the part that makes sense for us. He knew the whole arrangement going in, the product is in his hands, it's working. Staying is the easy decision. That's what a no-brainer actually is: not a discount, but a sequence where each party's easiest yes arrives at the right moment.

It also has to make sense for Meta: a cheap first purchase gives the algorithm a conversion event at a volume it can learn from.

→ Free ad spend: the 5% that fund the 95%

The thing I needed most as a brand was the ability to scale at roughly 1x ROI. Because I wasn't competing with the brand next door. I was competing with VC-funded companies and, for share of wallet, with the five largest companies on earth.

So I built what I call free ad spend: revenue that costs nothing to acquire.

In any audience, about 5% are luxury buyers. They'll buy everything in your catalog if you're in the right place at the right time with a serious offer. Say you sell silk pillowcases. That's what the audience came for. You offer the full bedding set at high ticket, and if they take it, the pillowcase they wanted is free.

You gave them what they wanted in exchange for what you needed. After margin, that order leaves you with about $150 you didn't have to buy from Meta. That $150 goes into ads. Tomorrow's traffic produces two of those orders. Then four.

The compounding gets serious once the high-ticket offer lives at every seam in the journey: right after the price, right after purchase, inside the email flows. Every seam is another self-funding loop.

→ The AOV trap, with the math

Everyone brags about upsells. I did too. It's easier. But you end up in the same traffic arbitrage model with no free cash flow, and here's exactly how that plays out.

Say you get AOV to $70 at a 2.4x store-level ROAS. Solid margin, 20% net. You take 5–10% out. You reinvest 5–10% into scaling. You reinvest 5–10% into product. Where's the profit? More importantly, where's the free cash to grow and sustain growth?

Then the clock runs. A year in, the numbers only work at $85 AOV. Three years in, $100. Meanwhile your audience has seen your product, competitors are undercutting your front end, and there's nothing in the bank. That's how 90% of brands die. Not from a bad product. From never building a cash engine outside of paid media.

I gave myself 3–12 months to build free cash flow. It's the single most important deadline in the business.

→ Let the customer raise the AOV for you

High AOV, as usually practiced, serves the brand, not the customer. In a market where credit is tight and competition is loud, you put the customer first and design the journey so the money comes back to you anyway.

Instead of selling the complete bundle up front, which only that 5% will take, sell the first product in a way that creates the next problem, then solve that problem.

Say you sell dishwashers. Huge problem solved, hours back every week. Now they need tablets. Sell tablets. A month of your tablets leaves residue. Sell the cleaner. They've just spent real money on hardware and they're nervous it'll break. Sell the 5-year warranty.

Four purchases. One acquisition cost. Every one of them arriving at the moment the need is most obvious, from a customer who trusts you more than they did at checkout. Every product creates downstream problems. Map them, sequence the solutions, and the customer raises your AOV for you.

→ Two products, two jobs

Some products exist to acquire a customer. Some exist to retain one. You need both, and the mistake I made early was asking one product to do both jobs.

The acquisition game doesn't scale on its own. CPMs go up. More brands enter and the customer's choice set widens. And every month you're expected to acquire more customers, cheaper. It works when you're new or riding a trend. After 12–16 months, COGS, ad costs, and overhead all move in one direction.

So the front-end product buys the customer. Breakeven is fine. The back-end product retains them and makes the P&L, sold to someone who already trusts you at a CAC of zero. Acquisition is too expensive to treat as the business. Retention is the business.

My bet is that the brands winning the next few years look more like small communities and fan clubs than stores, because when acquisition costs keep climbing, you have to hold a customer for at least six months to make the math work.

PART 2 — CREATIVE

→ Ads are fuel

You can have the best research process in the world, and if I ask why your evergreen ad works, you'll struggle to explain it. If you could, you'd replicate it on demand.

So I stopped thinking of creative as art and started thinking of it as fuel. The farther you want to drive, the more you need, and every 400 miles you have to stop and refill. That's how the algorithm works. We committed to doubling ad volume every month, and that commitment forced the systems below into existence.

→ Test forward, not backward

The way most people analyze winners is backward: look at what worked, guess what triggered it. That's storytelling.

We flipped it. Before an ad launches, we decide what it's testing. A specific hook, a specific angle, a specific creator. If it wins, we know what won, because we chose the variable in advance. Then we produce more of that variable.

And we stopped accepting generalities as explanations. "UGC works." "The trendy hook works." Useless. The real answer is deeper: what awareness stage was the viewer in during the first ten seconds, and what did the ad do to it?

→ Repurpose before you produce

A winning script gets shot with five different creators. A hook with strong retention but a weak body gets married to the body from another ad that held. We never let a proven component die with the ad it was born in.

→ Test the argument before you pay for production

Producing creative is the expensive part, so we moved as much testing as possible in front of it.

Seven angles you want to test? Build a listicle landing page, "7 reasons why," one angle per reason, and put a heatmap on it. Watch where people stop. The page converts better than a standard product page because people like to compare, and you've just tested seven angles for zero dollars.

Ten hooks for a video? Run them as ten statics first. The winner goes to the creator. You've raised the odds on the expensive asset before you paid for it.

And from the start, before any angle gets a dollar of paid spend, it goes to email. Subject line is the hook. Body is the argument. Sent to a slice of existing subscribers under a person's name, not the brand's, so nobody opens out of habit and the replies come in honest. I read CTR, replies, and click-to-purchase. Opens are noise. The list is warm, so absolute numbers flatter you, but the ranking between angles holds when it goes cold.

→ The feed is a sedative

Someone scrolling is getting a hit of dopamine every ten seconds. In that state, their problems are offline. Chronic pain, bad sleep, the thing they promised themselves they'd fix this year. None of it exists for those minutes. The only thing that exists is the next hit.

So your ad isn't a pitch. It's an alarm. You have about five seconds to bring one problem back online.

For someone who isn't sleeping well: "You sleep for 28 years of your life. Do it right," over a product photo they can't quite parse. The first half restores the problem. The second half hands them agency. The photo makes them curious enough to click. That's the whole job of the first five seconds.

→ Match expectations, emotion, and tone from ad to page

When did you last trust a car dealership to sell you groceries? Or want a comedy when you felt terrible? Wrong expectation, wrong emotion.

That's what happens when a static ad hooks someone with an idea, they click, and the first thing they see is "$97. Buy 2 Get 1 Free." Unless they were solution-aware and shopping for exactly that, it's a shock. Expectation broken.

The landing page has to meet the awareness, emotion, and tone the ad created. A two-minute video that walks through problem, solution, and mechanism puts the viewer in a shopping state, so a clean product page with proof and objection handling is right. A short why-focused ad needs a page that delivers the what and the how. If the ad is the what, the page is the how.

Once you're doubling volume monthly, matching every ad to a page gets hard. The fix: start from the ad. Take the evergreen creatives and build the landing page out of what they already say.

→ Monetize every view

Two kinds of people see your ad: the ones who buy and the ones who don't. You pay for both, and the second group is much bigger. You can't directly monetize them. Meta's timing-to-problem math doesn't allow it.

But monetization isn't only money. It's seeding a belief. If the ad talks about the thing that kept them up last night, they'll hear every word, and the right words plant something that turns into a relationship with your solution later. Every view either produces a customer or produces a relationship.

PART 3 — MEDIA BUYING

→ It's poker

The size of your bet is determined by the hand you're holding against the table. The better your assets, the bigger you can spend against competitors.

About 95% of media buyers still believe scaling setups are what make an account work. They think they'll out-think the algorithm by finding setup-driven patterns. That belief has done more damage to brands than any other idea in this industry.

You will not be smarter than the algorithm. It's a system you don't want to constrain. Feed it the best assets you have and give it room.

→ Creative is the targeting

Every ad is a separate salesperson reaching a specific audience. Ad set settings aren't targeting. Creative is.

Cold: unaware of the problem or of your solution. Warm: engaged, aware of you or the solution, not yet ready to buy. Hot: almost bought, one objection away.

We segment creative to match.

Prospecting ads bring new people in at low frequency, as cheaply as possible. Long-form VSLs, longer UGC, longer trend-driven video.

Semi-prospecting ads speak to the solution-aware. Short and mid-length UGC, TikTok-style trend videos, outcome and benefit statics, UGC statics, organic-looking statics, Buzzfeed-style objection videos.

Retargeting ads exist to kill the last objections. Mostly statics.

Master one asset in each segment and you have a funnel that fills itself.

→ Frequency tells you the truth

You don't build the funnel with structure. You read it off frequency, over the last 3–7 days.

Around 1.0–1.3 (depending on market size), the algorithm is buying fresh reach with this ad. It's prospecting.

Around 1.3–1.8, it's still reaching new people but also re-serving the ones who engaged. Mid-funnel.

At 2+, it's pushing the ad hard at a small, hot pool because this creative closes.

Nobody assigned those roles. The ads earned them.

→ Testing: ABO, and the hierarchy of evidence

Before a scaling campaign exists, there's an ABO campaign and real money spent finding the assets. I've never found a better way to test creatives, landers, offers, products, or copy.

Every asset gets its own ad set so budget is allocated separately and the data is clean. Settings are dead simple: broad, Advantage+ audience, exclude purchasers from the last 180 days. Full freedom for Meta. And critically, you test in the same environment you'll scale in, or the test tells you nothing.

Then the decision, in this order.

Spend. Say you test one angle with three variations in an ad set. Which one is Meta choosing to fund? That's a vote from a system that has looked at billions of data points and decided this variation brings new customers most efficiently. Read it first.

CAC / ROAS. Breakeven or better? It's doing the job.

Consistency. Did it deliver at target for the last seven days? One good day is variance. Seven is an asset.

Engagement. You're on social media, whose entire model is pushing engaging content in front of more eyes. Positive engagement is as important as any step above it.

An asset that clears all four goes to scaling.

The hard case is high spend with bad CAC. Don't assume. Look at the whole ad set. Did the lower-spend variations bring customers at target? If yes, the high spender is prospecting and the others are closing on the second, third, or last click. So you analyze the high spender in detail: usually the hook needs more pain or outcome, or the video needs pace, or the hook is fine and people leave the moment the body starts selling.

The other case: you've spent 5–10x breakeven CAC on the whole ad set with nothing. If soft metrics are good, a little more rope. If not, kill it. Marrying an idea is the most expensive habit in this business.

Out the other end of this process you should have 3–5 sustainable winners covering every segment. That's your hand.

→ Scaling: momentum, not tricks

There's no cheat code here either. One consolidated Advantage+ campaign, 7-day click / 1-day view attribution. If we're scaling one hero product through one funnel, there's exactly one of these.

The budget is what matters. Open at 15–20x breakeven CAC or it never becomes an engine. Too low and you get slow growth and lost momentum.

After any big budget increase, expect the account to underperform for a bit. It's a new learning phase in a new environment, not a verdict. Judge on seven days. If it's stable and new assets are flowing in from testing, raise 20–30% every 3–4 days.

And keep 20–30% of total spend in testing, permanently. Finding new assets is an evergreen job. Winners decay.

→ Third-party pages

Run ads from 3–5 pages, not one brand page.

A video from "HealthyLab" is an ad. The same video from "Pain-Free Life In Your 50s" is content. The viewer's prejudice going in is different, so we see higher CTR, longer watch time, better conversion. Some ads that flop on the brand page work on a third-party page because they simply fit there.

Second-order effect: the same person seeing your message from five sources reads consensus, not a campaign. This is one of the biggest low-hanging fruits in almost every account I've looked at.

→ Follow consumption, not your taste

Content that's out of context in the feed loses. Content that looks like what people already consume all day wins. The formats that did the most for us:

Long-form VSLs for products that solve painful problems. Every 9-figure health and wellness library is full of them for a reason.

"3 reasons why I hate buying this" for products with extreme demos. Sounds insane as a sales angle, but people hate being sold to, and this format lets them engage from an emotional state with no buying resistance.

Pop-quiz UGC. Interactive, consumed as organic rather than paid.

The podcast angle. Podcast consumption keeps rising, and a creative shaped like one gets people into the funnel without their guard up.

PART 4 — LANDING PAGES

→ A landing page is a sales process

Everything in this section starts where the ad ends. A landing page is nothing more than a sales process, and DTC has many: landing on the homepage is a long one, landing in checkout is a short one. What matters is that every landing has a purpose and a specific group it's built for.

Supplements are a commodity in a crowded market? A quiz funnel makes the same product feel like a personalized solution. Clothing? A listicle showing seven ways to wear the piece before the collection page, because what you're actually selling is how they imagine it looks on them.

→ What matters to them vs. what matters to you

The system I use to decide which pages to test is three steps.

First, list what raises the probability, in their eyes, that this works. For a supplement: how it works, the mechanism. How it's different from what they've seen. How they can be sure their body is right for it. How people just like them got results.

Second, spot the pattern. Every item starts with "How."

Third, match that pattern with a sales process they don't recognize as one. People love buying and hate being sold. The moment they feel the process, they resist, unless they're actively seeking a solution, which your ads and research already told you.

Then build it. Product page, collection page, advertorial, listicle, quiz, VSL, long-form, short-form. The format follows the pattern.

→ Take the commodity out

Some people think we're in ecommerce. Some think we're in the efficiency business. Every 8- and 9-figure brand knows the truth: the customer can go to Amazon or retail and solve the problem cheaper and with more trust, custom product or not. So we're in the marketing business, and the job is removing the commodity from the product.

The landing page's first fold (headline, banner, opening copy) is where that happens. It's the most important real estate on the site. For a product like a red-light wand, I'd write three or four different first folds around three or four different angles and test them.

One angle: target a specific customer with one specific mechanism. "Serums, treatments, and spa visits are great for your skin in your 20s and 30s. Not at 46. At this age it's about direct collagen stimulation. That's where our award-winning red-light wand comes in: it boosts collagen naturally by 89% and lets you look young again."

That copy takes every commodity alternative in the category, devalues it by naming the reader's age, and presents the product through one sharp angle. After reading it, the Amazon listing looks like the wrong tool.

PART 5 — EMAIL

If DTC is a football team, email is the bench. The better your reserves, the stronger the team, and you send them in when you need help.

It's also how you win the data game. Meta owns data and sells you a slice. If you own data, you pay less for the slice. Everyone knows how to set up an abandoned-cart flow, so I'll skip that. These are the parts that moved the needle.

→ Capture 10% of traffic

This is a numbers game. 5,000 sessions in a week at 10% opt-in is 500 emails. Scale and next week it's 10,000 sessions and 1,000 emails. Nail the opt-in and the snowball gets big. Over the last year we collected around 70K addresses, roughly 192 a day. 192 warm people a day you can reach for free.

→ Close in the first email

If you captured with a discount, go hard on the first welcome email's conversion rate. It's direct revenue, and everyone who buys from it can still be re-engaged later.

→ Segment as narrow as possible

Think of the list as a deck of cards. More cards, more combinations. So we send a broad campaign with open-ended questions about the problems people face. Whichever problem they pick becomes their segment, and the next campaign targets the biggest one. Once you know the dominant problem on the list, use it as the angle for the welcome flow and the post-purchase flow. Click-through goes up immediately.

→ Use the list to create momentum in the ad account

The best thing we did with email recently: build momentum in the ad account before launching a test.

You've got a big batch of creatives and a lot of resources went into them. Before launch, send a one-hour flash sale to 30% of the list. Launch the testing campaign inside that spike. The account enters learning with a conversion surge already underway, and your new ads get credit for the environment you created.

→ Fix the cohort

Your cohort data tells you how customers consume, why they cancel, why they stay. Wrap that in email and 12-month LTV moves a lot. If month-five cancels are overstock, the email offers a pause instead of a cancel and 50% off a new product that might help. Straightforward, and it changes the shape of the whole LTV curve.

PART 6 — THE EXTRA 30%

Four things that are easy to start and change how the brand runs:

→ Influencer whitelisting

Find micro-influencers who make solid UGC. Shoot content with them, get permission for their Facebook and Instagram pages, and run the ads from their pages.

Then a separate remarketing campaign on main events from the last seven days: five ad sets, five influencers. Recent engagers see five different people talking about you in a week and conclude the brand is having a moment.

→ EU localization

If you want less competition and a faster path to number one, look at the EU. For one of our brands, we ran roughly 80% of EU countries on localized product pages and some localized ads, then filtered winners by ROI. The best countries got doubled down: we brought in the most known influencers in each market, which is cheap there, built awareness, and scaled at a 40% net margin, which is nearly impossible in the top five countries. Number one in our segment in under a year.

→ Amazon

Remember the goal of being able to scale at around 1.2x ROI? Amazon and retail are how. Once you're around $300K a month, Amazon snowballs on its own and quietly covers part of your ad spend.

→ Seeding plus cost caps

Seeding works: close 200 micro-influencers a month, they post, you get traffic, social proof, and a mountain of content. The problem is using the content. Throw every piece into a cost-cap campaign. Out of 200, one or two will spend, and you found them without guessing.

THE LOOP

Here's why the order matters.

The offer creates free cash flow, which funds creative. Creative volume feeds the algorithm, which does the targeting. Media buying reads the algorithm's votes and scales the hand it's been dealt. Landing pages turn the ad's promise into a sales process. Email captures the traffic, builds momentum for the next test, and repairs the cohort. The extra layer adds 30% on top of a machine that already works.

Skip a step and the next one gets expensive. Run them in order and each one makes the next cheaper. That's the difference between a brand doing $2.4M a month in year two and one that's still renting Meta's audience at breakeven.

This is still less than 2% of what I have to share.

[Ecom Khalifa @EcomKhalifa]

@chillaurelius ofc bro