Launching a New Beauty Brand in 2026

Launching a beauty brand has never been cheaper or faster than it is right now. Minimums that used to sit at thousands of units now start in the low hundreds. Development cycles that took a quarter now take weeks. The cost of putting a product in front of a stranger has fallen dramatically. None of which has made it easier — because the hard part was never the product.

This is our 2026 launch guide: the eight steps that take a new beauty brand from idea to shipping, in the order we’d actually do them. One theme runs through all of it: spend the minimum required to become sellable and protected, and save everything else for the only thing that proves you have a business — strangers who buy. Our earlier 2023 guide still stands as a general primer — this edition covers what has changed since: new compliance obligations, collapsed minimums, and a different map of where customers are actually found.

---

In this guide

• Identify Your Niche and Go To Market Fast
• Lock Down the Name Before You Print Anything
• Create a Visual Identity You Can Actually Reproduce
• Find a Great Product — And Size the First Batch to Sell Out
• Secure Your Trademark, Earlier Than Feels Necessary
• Design Labels That Are Legal Before They’re Beautiful
• Create a Digital Presence That Can Take Money
• Build Your Initial Marketing Strategy (This One’s Yours)
• The bottom line

Identify Your Niche and Go To Market Fast

Conviction is what got you this far. It isn’t what gets you paid.

At the start, the market has no opinion about how good your idea is. It responds to execution, and it accepts exactly one form of evidence: strangers handing over money. Not a mood board. Not a business plan. Not a friend telling you the texture is incredible. Until people with no relationship to you have paid, you don’t have a brand — you have a hypothesis with packaging.

We’ve watched this pattern hold across more than a thousand brands. The founders who make it through the first stretch ship before they feel ready and let the market teach them. The ones who stall are always convinced they need a little more time — to perfect the formula, to raise a round, to get the box exactly right. They mistake motion for progress, and the money runs out while the brand is still in rehearsal.

Nobody Is Watching Yet — And That’s the Good News

The hardest thing for a new founder to accept is that nobody knows who you are, and nobody is paying attention.

Your first few thousand orders will not define your legacy. One bad review will not follow you around. An imperfect launch will not close doors permanently, because the doors aren’t being watched. What those early orders will do is hand you the only thing that matters right now: evidence about what people will actually pay for.

That should feel liberating rather than deflating. If nobody’s watching, you are free to launch before you’re ready, free to be wrong in public, and free to treat the first year as practice rather than a permanent record. Chasing perfection at this stage is polishing a boat that never leaves the dock — it looks excellent, and it goes nowhere.

What Actually Counts as Proof

Set the bar somewhere specific, and set it in strangers.

A useful threshold is a thousand people you don’t know choosing to buy. The number matters less than the definition: no friends, no family, no colleagues doing you a kindness. Those orders are the most misleading data in early CPG — they arrive out of loyalty and they tell you nothing about demand. They will also make an unviable product look viable for exactly as long as it takes to run out of relatives.

What you’re really looking for underneath the number is repeatability: the same message, through the same channel, with the same offer, producing sales more than once. That’s the difference between a lucky week and a business.

Why Speed Is Worth More Than It Used To Be

The case for moving quickly has always been about learning. What’s new is how much cheaper moving has become.

Minimum order quantities that once sat in the thousands of units now start in the low hundreds. Development cycles that took a quarter now take weeks. A launch that needed tens of thousands of dollars a few years ago can start in the low thousands. Every one of those shifts lowers the cost of being wrong — and the cost of being wrong is the real currency of an early-stage brand.

The corollary is less comfortable. Waiting has become relatively more expensive. When launching cost a fortune, deliberation was rational. Now that a real, sellable product is within reach in weeks, months spent deliberating are months spent buying nothing.

Start With Who It’s For

None of this works without a decision about who you’re for — and that decision belongs before the product, not after it.

An ideal customer profile isn’t a demographic sketch or a slide with a stock photo named “Ashley, 32.” It’s a working answer to four questions:

• The job: what are they actually trying to get done?
• The pain: what’s frustrating or failing about how they do it today?
• The desire: what outcome would make them feel the money was well spent?
• The context: where are they, what do they already use, and what do they believe?

Get those right and most downstream decisions stop being arguments. Which format to launch, what to say on the label, which channel to test, what price is defensible, which feedback to act on and which to ignore — all of it resolves faster when you know precisely who you’re serving. Founders who skip this step don’t avoid the work. They just do it later, more expensively, in public.

Focus is also what makes a small brand competitive at all. You will not out-spend an incumbent. You can absolutely out-specify one, by being unmistakably right for a narrow group of people who currently feel underserved.

One Caution

Narrow is a starting position, not a life sentence.

The point of picking a niche isn’t to stay small forever — it’s to concentrate limited resources until you have proof, then expand from a position of evidence rather than optimism. And don’t mistake the exercise for certainty: your first customer profile is a hypothesis too. The people who actually show up will edit it, and you should let them.

Define who it’s for. Ship the smallest real version. Let strangers tell you whether you were right.

Lock Down the Name Before You Print Anything

Your domain isn’t a branding exercise. It’s a cheap insurance policy, and the window to buy it is now.

Here’s the failure we watch most often: a founder falls in love with a name, spends months on identity, orders packaging — then discovers the name is taken, trademarked, or already attached to something adjacent. Rebranding after your boxes are printed is one of the most expensive unforced errors in this business.

So do the boring version, in one afternoon, in this order:

1. Check the domain.
2. Check the handles on every platform you might ever use.
3. Run a preliminary trademark search (step 5 — do it now, not later).
4. Then buy all of it.

The old rule was that a name had to be under ten characters and .com or nothing. That has softened. A .com still buys credibility and is worth paying for when you can get it, but plenty of brands now run happily on a .co or a brandable variant and no customer blinks. What still matters hasn’t changed: it should be easy to say out loud, easy to spell after hearing it once, and free of accidental second meanings. Say it into a voice memo. If you have to spell it twice, keep looking.

One rule that hasn’t moved: buy the name before you announce it, not after.

Create a Visual Identity You Can Actually Reproduce

Two things have moved fast here, and they pull in opposite directions.

The floor moved. AI tools will now generate usable logo concepts, palettes and type pairings in minutes, for close to nothing. Posting a brief to a freelance marketplace and waiting a week is no longer the fastest or cheapest way to reach a starting point.

The ceiling didn’t move at all. AI is very good at producing variations of what already exists — which makes it very good at producing a logo that looks like a logo, and much worse at producing one that looks like you. It also doesn’t reliably solve the unglamorous constraints that decide whether an identity survives contact with reality: legibility at thumbnail size, contrast that holds on a phone in daylight, a mark that still reads when it’s printed on a 30ml bottle.

So the practical path for most founders launching now is a hybrid — use AI to explore fast and land on a direction, then pay a human to sharpen it and make it production-ready. You spend a fraction of what a full identity project used to cost, and you skip the part where you pay someone to guess at a direction you hadn’t defined yet.

Keep the scope honest. At this stage you do not need a brand book. You need a logo that works small, two or three colours, one or two typefaces, and a written note on how they’re used — enough that you, a label printer and a freelance editor all produce the same-looking thing.

Consistency is what reads as “real brand” to a stranger. Sophistication is optional. Consistency isn’t.

Find a Great Product — And Size the First Batch to Sell Out

Every founder arrives with a formula in their head. It’s usually vivid, usually specific, and usually the most expensive thing they could possibly build first.

The question that matters at this stage isn’t “what do I want to make?” It’s “what is the smallest real product that lets me start collecting evidence?” Those are different questions, and they have very different answers.

How much to order

One rule should govern your first production run:

Not the amount you can afford. Not the minimum the factory will accept. Not the volume that makes the per-unit price look good on a spreadsheet. The amount you can realistically sell, inside a defined window.

Two reasons that’s the right variable. First, everything beyond that line is a guess about demand you haven’t measured yet — and unsold units are an expensive way to store a guess. Second, inventory in beauty is cash with an expiry date: formulas carry a shelf life and a period-after-opening, so a three-year supply is rarely a three-year asset.

For a brand with no sales history, honest confidence usually points to a small first batch. That’s why “start small” is sound advice for most first-time founders — but it’s a consequence of the rule, not the rule itself. A brand with proven reorder velocity should order more, and under-ordering carries its own cost: stocking out of a winner wastes demand you already paid to create.

The honest test is a sentence, not a spreadsheet: who is going to buy these, through what channel, and how fast? If you can’t answer that with something better than optimism, the batch is too big.

The two doors

Private label means starting from a formula that already exists and already works, and making it yours. You’re not inventing chemistry — you’re choosing a proven base, dressing it in your brand, and getting to market. Industry-standard private label minimums typically run 1,000 to 5,000 units per SKU. Ours is 140. That matters because it lets a first batch match an honest sell-through estimate instead of a factory’s minimum. Expect roughly $2,500 to $10,000 to cover samples, inventory and logistics, and 4 to 6 weeks of lead time.

Custom formulation means building the formula itself — your actives, your texture, your claims, your intellectual property. It’s genuinely better, and it’s genuinely a commitment: development starts at $9,800, with a 500-unit minimum once the formula is finished, and a realistic all-in range of $20,000 to $30,000. Lead time runs 6 to 12 weeks.

Read those two paragraphs again and notice what they are not: a quality ladder. Private label isn’t the cheap seats and custom formulation isn’t the VIP room. They’re a sequencing decision.

The expensive mistake

The most common way we watch idea-stage founders burn their runway is going straight to custom formulation, because it feels like the real brand move.

It’s the most expensive possible way to buy an untested opinion.

Custom formulation is an excellent answer to a question you haven’t yet earned the right to ask: which of my products deserves to be permanent? You cannot answer that before you have customers. Nobody can — not us, not a consultant, not a focus group.

And the budget math punishes it. First-time founders routinely put 80–90% of their capital into formulation, production and packaging, then discover there’s nothing left to go find a customer with. A workable ceiling is 40–50% of your launch budget on product and manufacturing. The rest is runway — and runway is what buys evidence.

If you spend 90% of your money on inventory, you haven’t launched a brand. You’ve bought a garage full of very nice boxes.

How to actually decide

That’s the ladder, and it’s the one almost every brand we’ve watched succeed actually climbed. Not because custom formulation is a prize at the finish line, but because by the time they got there, they knew exactly what to spend it on.

More on choosing what to keep and what to cut: Focus to Scale →

Secure Your Trademark, Earlier Than Feels Necessary

This is the step founders skip because it feels premature, and it’s the one that most often costs real money when skipped.

Start with a preliminary search before you commit to anything printed. You’re looking for identical and confusingly similar marks in your category — somebody selling industrial adhesive under your name may be irrelevant; somebody selling serum under it is not. Searching is free to start and can save you an entire rebrand.

When you file, you’re choosing between two things. A standard character mark protects the name itself in any font or styling — usually the broader and more useful protection for a young brand. A stylized design mark protects one specific visual treatment of it. If you can only do one right now, most founders are better served protecting the name.

Three practical notes:

• Registration takes months, not weeks. Filing early is the entire point — and you can sell while it’s pending.
• Expect scam mail. Once your filing is public, official-looking invoices arrive from private companies that are not the trademark office. Pay nothing you didn’t initiate.
• This is the one step where a professional is usually worth it. An attorney or a reputable filing service costs a fraction of a rebrand.

None of this is legal advice, and classes and filing strategy get genuinely complicated. But the sequencing rule is simple, and it’s the part founders get wrong: search before you design, file before you print.

Design Labels That Are Legal Before They’re Beautiful

This is the part of launching a beauty brand that has changed most in the last two years — and it’s the one where working from outdated advice can actually hurt you.

For most of the last century, US cosmetics ran under a regulatory regime that had barely changed since 1938. That era is over. The Modernization of Cosmetics Regulation Act (MoCRA) is now in force, and it applies to indie brands, not just conglomerates.

What that means in practice:

• Someone is the “responsible person” for each product — the entity named on the label. That’s usually you, not your manufacturer.
• Facilities manufacturing cosmetics for US distribution must be registered with the FDA, and registration is renewed on a recurring cycle.
• Each marketed product must be listed with the FDA, including its ingredients, and kept current.
• You need a way to receive adverse event reports, and records supporting your safety substantiation.

There is a small-business exemption — broadly, for companies under $1 million in average annual US cosmetic sales over the previous three years — but read it carefully before relying on it. It is not partial. Certain product types are carved out entirely, including products used near the eye, products intended to be injected or used internally, and products designed to alter appearance for more than 24 hours. If a single SKU falls into a carve-out, the exemption stops covering your business.

The label itself still needs the fundamentals: product identity, net quantity, ingredients in descending order of predominance, directions, any required warnings, and the responsible person’s contact information. Batch coding isn’t decoration — it’s how you run a recall without destroying the company.

Here’s the good news for a new brand: this is the part you can genuinely lean on a manufacturer for. Working with a registered, compliant US facility means the manufacturing-side obligations are already handled and you inherit a compliance posture instead of building one. What stays yours is the responsible-person role and the accuracy of what your label claims.

We wrote the long version of this: MoCRA Compliance Handbook → Read it before your first production run, not after.

Guidance, not legal advice — requirements change and your product mix matters.

Create a Digital Presence That Can Take Money

The bar here is lower than founders think, and different than it used to be.

You need a place a stranger can buy from without emailing you. That’s the requirement. A clean, fast, mobile-first store on any mainstream platform clears it. What you don’t need is a custom build, a blog, an origin-story page, or six months of design.

What’s changed is where discovery happens. The old assumption was simple: build the store, drive traffic to it, convert. Today a large share of beauty discovery — and increasingly the transaction itself — happens inside social platforms like TikTok Shop, where creators do the selling and the buying decision and the purchase occur in the same place.

That creates a split worth understanding early, because it shapes how you set things up:

• Marketplaces and in-app shops are excellent at putting your product in front of people who’ve never heard of you, and they generally keep the customer relationship. You get the sale; you often don’t get the email.
• Your own store converts a cold audience worse, but the customer is yours — the email, the purchase history, the ability to sell them a second time without paying for the introduction again.

Neither is a strategy on its own. The founders who handle this well treat discovery and ownership as two different jobs, and work quietly to convert the first into the second: an insert in the box, a reason to register, an offer that only exists on their own site.

Two practical things to do now:

1. Claim your handles everywhere, including platforms you don’t plan to use. It costs nothing today and becomes unbuyable later.
2. Set up the boring plumbing before launch, not after — analytics, email capture, and some way to see which orders came from where. You can’t learn from traffic you can’t attribute.

Your site doesn’t need to be impressive. It needs to load fast, work on a phone, say what the product does in one sentence, and take money.

Build Your Initial Marketing Strategy (This One’s Yours)

Here’s where we stop handing you answers.

We can formulate your product, keep it compliant, and put it in packaging that survives a shipping container. We cannot make strangers want it. That isn’t a disclaimer — it’s the actual division of labour, and the founders who understand it early are the ones who make it.

When we look at a new brand, one of the things we’re quietly assessing is whether there’s a marketing-capable operator behind it. Not a marketing budget. An operator. Because the product is the part we solve; demand is the part you own.

So what follows isn’t a plan. It’s a map of where the ground currently sits, plus a few principles that outlast whichever channel happens to be winning this quarter.

The only number that counts

Strangers who buy.

Not followers. Not impressions. Not friends-and-family orders — those are the most misleading data in early CPG, because the people who love you will buy once and tell you nothing about demand. The bar is people with no relationship to you deciding to spend money.

Where the economics sit right now — examples, not instructions

The cost of reaching those strangers has moved, and mostly in your favour.

TikTok Shop has become one of the largest beauty channels in the US — creator-led, with discovery and checkout happening in the same place. Conversion there tends to run well above traditional social feeds, and brands running creator-affiliate programs commonly report acquisition costs materially lower than paid social for comparable products. A frequent launch shape now is seeding a wide pool of small creators rather than front-loading ad spend.

Take all of that as evidence of where the economics currently sit — not as your strategy. We’re a manufacturer. We see a lot of brands and we’ll tell you honestly what we’re seeing work. But any manufacturer who hands you a channel plan is guessing about your business, and channel advice ages faster than anything else in this article.

The principles that don’t rotate

1. Pick the channel you can personally operate. A strategy that depends on a skill you don’t have and can’t yet hire is somebody else’s business model. If you’re compelling on camera, that’s an asset — use it. If you’re better at email, community or in-person retail, those are assets too.
2. Buy evidence before you buy scale. Spend small across two or three approaches, kill what doesn’t move, concentrate on what does. Cheap losses are tuition; expensive losses are endings.
3. Own your customer data. Marketplaces convert well and keep the relationship. Your own store converts worse and keeps the customer. You want both — be clear which job each is doing.
4. Study before you spend. You can see what competitors are running and how long they’ve run it, and ads that stay live are usually working. (How to read competitor ads →)

What “it worked” actually looks like

You’ve found one combination of message, channel and offer that reliably turns strangers into buyers at a price you can sustain.

That’s it. That’s the entire goal of this stage.

It doesn’t need to be scalable yet. It doesn’t need to be elegant, and it doesn’t need to be what you’ll still be running in five years. It needs to be real and repeatable. Once it is, you have the one thing almost no idea-stage brand has: a machine — however small — that turns money into customers.

Everything that comes next gets dramatically easier once that machine exists.

The bottom line

If you take one thing from this, take the sequencing. Almost every step above is designed to be done cheaply and quickly, so that you arrive at the only part that actually decides your outcome — putting the product in front of strangers — with money still in the bank and time still on the clock.

The founders who struggle aren’t the ones who picked the wrong typeface. They’re the ones who spent nine months and 90% of their capital getting to a launch, and then had nothing left to find out whether anyone wanted it.

So: define who it’s for. Lock the name. Build an identity you can reproduce. Order a first batch you’re confident you can sell through. Protect it. Make it legal. Give it somewhere to be bought. Then go find your strangers.

We can carry most of that with you — the formula, the compliance, the supply chain, the boxes that show up on time. The part we can’t do is the part that makes it a brand: getting people to care. That one’s yours, and it should be.