Free · no email required
How to build one of these, and keep it advertised.
Thirteen chapters on building a brand in this category and getting it advertised — suppliers, processing, fulfilment, organic, and why ads die here. All of it learned running our own brand, paying for it ourselves. Read it, use it, ignore us. It's yours either way.
About 25 minutes. Or skim the chapter you need.
Part one · Building the business
Chapter one
Why this category is different
Most advice about running ads is written for people selling t-shirts. It assumes that if your ad is good and your offer is good, you'll be fine. In a restricted category, that assumption is wrong, and it's wrong in a way that costs people entire businesses.
Here, there is a step before performance. Before anyone sees your ad, an automated review decides whether your business is allowed to advertise at all. It doesn't read your ad in isolation. It reads your ad, your product fields, and the page you're sending people to. If any of those trip a rule, the ad dies — and if it happens enough times, the account goes with it.
That changes the order of operations completely. In a normal category you optimise for conversion and worry about policy later. Here you build for review first and optimise second, because an account that can't advertise has no numbers to optimise.
The thing to internalise: your creative is rarely what killed your ad. It's usually the least likely suspect. Almost everything in this guide is about the parts nobody looks at.
Chapter two
Suppliers, and the three ways people lose money
This is the part everyone underestimates, and it's the part where a first-timer has no leverage. Good suppliers don't advertise. They come to you once you're already selling volume, which is a real problem when you're not selling anything yet.
So the honest position for someone starting out: you're going to be evaluating people who all sound credible, without the one thing that actually tells you who's real, which is a track record of other people ordering from them repeatedly. Everything below is about not paying full price for that education.
Mistake one: bad product
You can't tell by looking. Everything arrives in the same vials and the same boxes. What you find out later, from customers, is whether what's inside matches what the label says. By the time you know, you've shipped it, and the damage is to your brand rather than your inventory.
The fix is independent verification before you sell, not after. Not the supplier's paperwork — your own. That costs money and time up front, which is exactly why people skip it, and it's the difference between a brand that lasts and one that quietly dies from reviews.
Mistake two: trusting the cheapest quote
Price is the easiest thing for a bad supplier to compete on, because they're not paying for the things that make a good one expensive. When one quote is dramatically under the others, that gap is the answer to a question you should be asking.
People send money to a stranger because the number was good. It is, by a distance, the most common way money disappears in this business.
Mistake three: too many products
This one doesn't feel like a mistake while you're making it. Launching with a wide catalogue feels like giving customers choice. What it actually does is spread your cash across a dozen slow sellers, so you're out of stock on the two things people actually want and sitting on the rest.
Start narrow. Find what sells. Put your money there. You can add range later from profit instead of from savings, and by then you'll know which additions are worth it.
What we do here: introductions to suppliers we already order from, so you skip the part where you're evaluating strangers with no leverage. That's most of the value of a launch, and it's the one part of this guide we can't hand over in writing.
Chapter three
The wall nobody warns you about
Most people build in this order: pick products, build a store, then try to take payments. They hit the wall at step three, after they've spent the money and the time.
Mainstream payment processors decline this category. Not "make it hard" — decline. And the worse outcome isn't the decline; it's getting approved by someone who didn't look closely, taking orders for two weeks, and then having the account frozen with your customers' money inside it while you work out what to tell them.
Do it in this order instead
- Register properly first. An LLC, a tax ID, a business bank account in the business's name. Processors won't seriously consider an individual, so this isn't optional and it's the slowest part to fix later.
- Get a processor to say yes before you build. High-risk processors exist for exactly this. You'll pay a higher rate per transaction and probably sit under a rolling reserve, where a percentage of your money is held for a period. That sounds bad until you compare it to a frozen account.
- Apply to more than one. Approval is not a coin flip you get once. Have a second option before you need it.
- Then build.
The high-risk rate is the thing people balk at. Look at it against your margin rather than against what a normal store pays and it stops being the deciding factor. The deciding factor is whether you can take money at all.
Chapter four
Shipping it yourself, longer than you'd think
People assume they need a fulfilment company on day one. You don't, and outsourcing early is how a lot of small brands hand away their margin before they have any.
You can run this out of your own space for a long time — genuinely into the low hundreds of orders a day if you're organised about it. Labelling, packaging, printing, packing, dropping off. It's time consuming and it is absolutely worth it, because in the early months your time is the cheapest thing you own and your cash is the scarcest.
What you get by doing it yourself
- Control over how it arrives. Packaging is the only physical contact a customer has with your brand. A warehouse will not care about that the way you do.
- You see problems immediately. Wrong item, damaged, running low — you find out while packing rather than in a review.
- Margin stays with you during the months when margin decides whether you continue.
- You learn what your operation actually costs before you pay someone else to run it.
When to stop
When packing is stopping you doing the thing that grows the business. That's the signal — not a volume number. If you're spending your day at a table instead of on marketing and suppliers, the maths has flipped.
How deep you go on presentation is up to you. Your own labels, your own bags, your own inserts. It costs more and it takes longer, and it's a real difference customers notice.
Chapter five
Organic is not a substitute, it's a multiplier
Video does something ads can't. It puts you in front of people who weren't looking for you, at no cost per view, and the ones who follow you stay reachable forever.
The mechanism is simple and most people miss it: a video that performs sends people to your profile, and the profile sends them to your site. You are not converting on the video. You're building an audience that keeps arriving long after you posted.
We've had a single video pass a million views. The traffic from something like that doesn't stop when the video stops trending, because the followers are still there.
Why it matters more here than in normal categories
In a category where paid ads can be switched off by a policy decision you don't control, an audience you own is not a nice extra. It's the thing that keeps revenue coming while you sort out an account. Brands here that only run ads are one review away from zero.
How to think about it
- Volume beats polish. You are looking for the video that works, and you find it by posting.
- Same rules apply. Everything in the later chapters about what gets flagged applies to organic too — accounts get restricted, not just ads.
- Anything that performs is worth paying to show more people. A video that already proved itself is a better bet than one that hasn't.
- Do both. Organic builds the audience, paid finds people the audience never reaches.
Chapter six
Pricing, refunds, and the boring truth about service
Pricing
Products in this space sit anywhere from about twenty-five to two hundred dollars, and where yours land depends on what it is, what it costs you, what you're competing against, and what your packaging and service justify. There is no formula worth printing, because the answer changes per product and per brand.
What is worth saying: price after you know your real per-order cost, including shipping, packaging and what the processor takes. People price off product cost alone, then discover their margin lives somewhere else entirely.
Refunds and chargebacks
These are mostly a service problem wearing a finance costume. Orders that arrive quickly, packaged properly, matching what was described, from a brand that answers messages, don't generate many disputes.
You get what you give. Give slow shipping, poor packaging, and silence when someone asks a question, and you'll get the refunds and the chargebacks that come with it. It's not complicated and it's not luck.
The questions you can't answer
You will be asked things you should not answer. Anything about how to use a product, what it will do for someone, or what's right for their situation. The answer is a polite decline every time, and it doesn't cost you the sale as often as people fear.
The brands that get into trouble here are the ones who answer helpfully in a DM. Nothing is more damaging in this category than a well-meaning reply.
Part two · Getting it advertised
Chapter seven
Your store is being read
When you submit an ad, the platform does not only look at the ad. It fetches the page you're pointing at and reads it. Everything on it. Product names, descriptions, disclaimers, the footer, the pop-up you forgot about.
This is where most rejections in this category actually come from, and it's why people find them so confusing. The ad looks clean. The ad is clean. The page it points at is what got flagged.
The language that causes trouble
- Anything describing how a person would use it. Quantities, schedules, how it's taken. This is the fastest way into a prohibited category.
- Anything promising a change in someone's body. What it will do for them, what it did for someone else, what to expect. Even hedged. Even in a customer quote.
- A percentage attached to quality. A number like that reads as a product claim, however you phrase it.
- Compound names in headlines and in URLs. The URL matters even when the visible page is clean. This one catches almost everyone.
- The standard legal disclaimers this category uses. Genuinely awkward, because you may need them on your store for legal and processor reasons. They still get read.
- Visible age-check wording. Not the gate itself — the announcement of it.
- Promises about money or certainty. A separate family of trouble that catches people who thought they were safe.
What to say instead
The angles that clear review are the verifiable ones, and they convert at least as well: third-party tested, batch traceable, ships from the US, shipping speed, packaging, price, support. Lead with the company. Sell the brand, not the product.
Check yourself rather than guess. Our scanner reads any page the way a reviewer would and shows you every flagged line with the reason. Free, no email. Scan a page
Chapter eight
The page behind the ad
You need one page that is written to clear review and does nothing else. Not your homepage. Not a product page. A separate landing page whose only job is to receive ad traffic, say enough to be trusted, and hand the visitor on to your store.
Two reasons. First, your store has to carry things a landing page doesn't — legal text, full catalogue, disclaimers — and every one of those is something a reviewer can object to. Second, a page with one message and one action converts better than a homepage with a menu.
What belongs on it
- The brand name and what you sell, in plain terms
- Trust: testing, shipping, where you ship from, how to reach you
- One clear action
- Nothing else. Every extra sentence is a surface.
The mistake that quietly wastes months
If the page your ad points at sits on a different domain from the one your tracking is installed on, your sales don't get counted. Not fewer sales — none. Orders come in, the platform sees nothing, and it concludes your ads don't work. So it stops showing them to the people who were buying.
We've seen accounts run for months in that state while the owner slowly decided advertising didn't work for them. It's silent, it looks exactly like failure, and it's a setup problem you can fix in an afternoon.
Chapter nine
What actually gets reviewed
Two things get submitted with every ad that most people never look at.
Fields you didn't write
Ad platforms offer to fill in your product details for you. What that means in practice is that they visit your page, take whatever text they find, and put it into fields that are submitted for review under your name. Entity names, phone numbers, age-gate wording, purity claims — things you never chose, in fields you never opened, being reviewed as though you wrote them.
Turn off automatic generation. Turn off automatic enhancements. Leave product detail fields empty on purpose. Then save, reopen, and check they're still empty, because they refill.
Creative you didn't approve
The same systems will generate variants of your ads and run them. In one account we inherited, roughly half the spend went to assets that were auto-created and later deleted — so there was no way to even see what had been running. If you don't control what runs, you can't diagnose what failed.
The standing rule: nothing runs that you didn't write. It costs you a few minutes per campaign and it removes an entire category of failure.
Chapter ten
How the machine learns
Advertising here is not you choosing an audience. It's you teaching a system what your buyer looks like, and the system doing the finding.
Tracking code on your site reports what each visitor did — landed, browsed, added to cart, bought. When someone buys, the platform looks at that account: what they watch, what they follow, their age, where they are, when they're online. Now it knows what one of your buyers looks like. Then it goes and finds more accounts that look like that one.
Why it compounds
One buyer is a coincidence. Ten and it starts seeing what they have in common. A hundred and it knows your customer by type, and has an enormous pool to search for more of them.
That's why the first two weeks are expensive and erratic. It's guessing. Those first sales cost more than they should, and that's not the campaign failing — that's the campaign paying tuition. Most accounts that "didn't work" were switched off during this phase.
Every sale you get today makes tomorrow's cheaper. That's the whole reason to protect the account rather than churn through new ones.
Two things that undo it
- Sales that never get counted. Covered in chapter four. If it can't see the purchase, it learns nothing from it.
- Constant editing. Every significant change sends a campaign back to the start of learning. Some accounts have never once been left alone long enough to finish.
Chapter eleven
When you get rejected
You will. The difference between an inconvenience and a lost account is entirely in what you do next.
The sequence
- Read what it says. The notice names a policy. It's vague, but it tells you which family of problem you're in.
- Look at the page, not the ad. Nine times out of ten it's the destination.
- Fix the specific thing and resubmit once. One change, one resubmission.
- If it rejects again, stop. This is the part people get wrong.
Why stopping matters
Submitting variation after variation of something that keeps getting rejected is how accounts move from "an ad was rejected" to "this account is a problem." The system is counting. Two rejections is a bad day. Ten is a pattern, and patterns get accounts closed.
If you're stuck after one honest attempt, change the approach rather than the wording. And treat an appeal carefully — it puts a real reviewer on your whole account, which is good if you're clearly in the right and bad if you're hoping nobody looks too closely.
Keep a log. Date, what was rejected, the policy named, what you changed, what happened. Six weeks later, that log is the most valuable document you own.
Chapter twelve
Scaling without breaking it
Scaling means spending more per day. The problem is that spending more means reaching further from your best buyers, and if the system only half understands who those are, results decay as you go out. That's the ceiling most brands hit.
It isn't a bigger budget that scales you. It's a system that knows what it's looking for well enough to stay accurate further out.
The boring rules that work
- Raise budget in steps, not jumps. Large increases reset learning and you start again.
- Change one thing at a time. Otherwise you can't attribute the change to anything.
- Judge nothing before the end of the first week. Earlier than that you're reading noise.
- Watch how often the same people see you. Rising frequency with rising costs means you've run out of audience, and more budget makes it worse.
- Cost per purchase is the only number that decides anything. Everything else is diagnostic.
Chapter thirteen
What to do this week
Reading a guide changes nothing. Here's the order that gets you somewhere, whether you do it yourself or not.
- Scan the page your ads point at. Ten seconds, free, and it will find something. Fix what it flags before you spend another dollar.
- Turn off automatic generation everywhere. Then reopen the settings and check it stayed off.
- Confirm your sales are actually being counted. Compare orders in your store against what the ad platform reports for the same period. If those numbers aren't close, nothing else matters until they are.
- Strip the destination page. Every sentence that isn't trust, price, shipping or testing is a liability.
- Then leave it alone for a week. Genuinely the hardest one.
The one thing we didn't teach you
There's a step above all of this: stopping review systems from reading your product pages at all. If they can't read the page, there's nothing on it to object to — and that single change does more than every rule in chapter three combined. It's built at the server, before the page exists, which is why it isn't something you can add to a page yourself.
That's the part we do for people. Everything else in this guide, you can do without us, and you should.
Next step
Start with the scan.
It takes ten seconds, needs no email, and tells you specifically what a reviewer would flag on your page. Most people find something they had no idea was there.
Nothing in this guide is held back except one thing.
We told you what it is in chapter nine. If you want that built, along with the store and the campaigns around it, that's the work.