Let's TalkLet's TalkPaid growth for beauty brands where the creative is the campaign - built around a testing engine that produces enough angles to keep acquisition cost falling instead of drifting up.
Beauty is decided by creative, so we lead with creative. Send us your product and your current best performer, and we send back two concepts, each with its hook and the reasoning behind it. Use them with us or without us - they are yours either way.
Four failures we see repeatedly in this vertical, and what each one actually costs.
Most beauty brands find a creative that works, scale it until frequency kills it, and then discover they have no pipeline of replacements. Acquisition cost climbs for a month while someone scrambles. The brands that scale cleanly are running a production line, not searching for a hit.
First-purchase ROAS looks healthy on a discounted trial size and tells you almost nothing. Beauty lives on the second and third order. If nobody is reporting contribution margin after returns and discounts, the account can look like it is winning while the business gets thinner.
Email should carry a fifth of revenue in this category, and in most brands we audit it carries under eight percent because the flows were set up once and never revisited. That gap is the cheapest revenue in the business, and it is usually sitting untouched.
Before-and-after imagery is restricted on Meta, and skincare copy drifts into health claims faster than anyone expects. An agency that has not worked in the category learns this by rejection, and the brand pays for the lesson in downtime during a launch window.
Business outcomes rather than dashboard metrics. None of these is a guaranteed number - where we do guarantee something, it is written into the offer above.
The real deliverables, not a list written to make a proposal look thicker.
Not the general agency pitch. The reasons that only apply to this kind of business.
A reorder prompt sent on a generic thirty-day cycle arrives too early for some products and too late for others. Timing each prompt to the real usage life of the product - how long a serum or cleanser typically lasts - catches customers as they run out, which is when a reorder feels helpful rather than pushy.
Creator content often becomes a brand's best-performing advertising, and the right to run it as an ad needs agreeing in advance. Negotiating rights after a video has already proved itself is slower and more expensive. We set the terms at the brief, so a winner can be scaled the week it wins.
A brand selling across skin tones and types needs creative that shows them. Customers look for someone who resembles them before they trust that a product will work for them, and creative that only shows one kind of skin quietly excludes much of the market the brand is paying to reach.
We look at every ad you have run, sort them by angle rather than by performance, and find which angles you have never tried. Most brands have tested one message forty ways.
Two concepts first, with hooks and reasoning, so you see the thinking before any money moves. Once we are working together, the first full batch follows - hooks, storyboards and production briefs built from the audit.
A fixed cadence of new concepts into a structure that can read them, with prospecting kept separate from retargeting so the numbers mean something.
Welcome, browse, cart, post-purchase and winback flows, written for the category rather than pulled from a template. This is usually where the first significant revenue lift appears.
Budget follows contribution margin after discounts and returns. That is how acquisition cost falls while spend rises, rather than the reverse.
The parts of this that come from having run it in this industry before rather than from running it well in general.
Beauty buyers hesitate over products they have never tried. Minis, samples and discovery sets give a lower-risk first purchase and a natural path to the full size - and they tend to attract customers who intend to use the product rather than those hunting for a discount.
When a product is also sold on marketplaces or in stores, some of the demand an ad creates converts there instead. Judging paid social only on direct-site sales understates what it does. We look at total sales movement alongside direct attribution, so budget decisions reflect the whole picture.
Beauty buyers increasingly research ingredients before they buy. Clear, accurate ingredient information and plain answers to common questions convert those researchers - provided the claims stay within what can be substantiated, since overstatement is both a platform risk and a trust risk.
Customers think in routines - cleanse, treat, moisturise, protect. Bundles built around a routine feel useful rather than promotional, raise average order value and increase the chance the customer sees results, which is what brings them back.
A skincare brand launched with no paid history and a single product. We built a Meta and TikTok creative testing engine that shipped concepts weekly rather than campaigns monthly, and split the funnel so prospecting was never flattered by retargeting. Acquisition cost fell 63% inside 90 days, email grew to 22% of revenue, and the brand crossed seven figures in nine months at 8.4x blended ROAS.
Read the full case studyThe weekly creative cadence is the whole thing. We stopped panicking every time an ad fatigued because there were always three more in the queue.
First agency to show me margin after returns and discounts. Our old ROAS number had been lying to us for about eight months.
They rewrote two ads before launch because of claim wording. I did not know that was a risk until they explained what happens when the account gets flagged.
Enough that a fatiguing ad is never an emergency, which in practice means a fixed weekly batch of new concepts rather than a monthly campaign. The exact volume depends on spend - a brand at 20,000 dollars a month needs fewer angles in rotation than one at 200,000 - but the cadence is fixed either way. Ad-hoc creative requests are how brands end up with one winner and no bench.
Either. We write the briefs and storyboards in every case, because the angle matters more than the production quality in this category. From there we can source and direct UGC creators, edit footage you already have, or hand the briefs to your in-house team. Founder-led content usually outperforms polished studio work, which surprises people.
Because ROAS on a discounted trial size can look excellent while the business loses money on every order. Beauty economics depend on the second and third purchase, so we report acquisition cost against contribution margin after discounts and returns, and we track repeat rate as a first-class number. ROAS is still in the report - it is just not the number the budget follows.
Yes, and it is usually the fastest revenue available. In most beauty brands we audit, email carries under eight percent of revenue when the category benchmark is around a fifth. The flows exist, they were set up once, and nobody has touched them since. Rebuilding welcome, browse, cart, post-purchase and winback sequences typically moves that within the first quarter.
Restricted on Meta, and the restriction is enforced. We plan creative around demonstration, texture, application and result-in-context rather than side-by-side comparison, which performs comparably once the angles are right. We also run a claims review before launch, because skincare copy drifts into health claims easily and a flagged account during a launch window is expensive.
Yes, and the skincare brand in the case study above started from zero paid history. What matters more than existing revenue is whether the product has been validated by anyone who is not a friend of the founder, and whether there is budget to reach statistical significance on a test. Below roughly 10,000 dollars a month in media, a testing engine cannot gather enough signal to be worth what it costs.
This is one part of a bigger service. Here is the whole of it, and the closest neighbours.
No pitch deck, no discovery call you have to sit through. Tell us the situation and we will tell you whether we can help.