by Trivas.aiEvery beauty founder eventually types some version of "what is ROAS for beauty brand benchmark" into Google at 11pm, usually right after a campaign underperforms. Fair enough. ROAS is return on ad spend: total revenue divided by ad spend, expressed as a ratio. A 4x ROAS means every $1 spent on ads brought back $4 in revenue. Simple math, complicated context.
ROAS itself isn't complicated. Revenue over spend. What gets messy is figuring out whether your number is actually good, and beauty brands ask this question more than almost any other vertical.
Part of it is structural. Beauty catalogs run deep, often 30-100+ SKUs across skincare, color cosmetics, and haircare, each with its own margin and price point. Promos run constantly (GWPs, bundles, holiday sets), which shifts revenue week to week. And beauty leans harder on Meta and TikTok than most categories because the product is visual: swatches, before/afters, texture shots. All of that makes a single flat ROAS benchmark almost useless without context.
There's also a data problem underneath the question. "Good ROAS" changes depending on whether you're looking at platform-reported ROAS (what Meta's dashboard tells you), blended ROAS (total revenue across the business divided by total spend), or MER (marketing efficiency ratio, a close cousin of blended ROAS). This article is specifically about blended benchmarks, because that's the number that actually reflects what's happening in your bank account.
Here's the range you'll see cited most often across the DTC beauty world:
2x-3x
3x-5x
Beauty brands often report higher ROAS than apparel or electronics sellers, and there's a real reason for that beyond luck. Skincare and cosmetics frequently carry 60-75% gross margins, versus much thinner margins on apparel or hardware. Higher margin means a brand can hit profitability at a lower ROAS than a category selling on razor-thin unit economics.
AOV throws a wrench in all of this, though. A $28 AOV lip gloss brand needs a noticeably higher ROAS to hit the same profit dollar as an $85 AOV skincare bundle seller, because fixed costs like shipping eat a bigger percentage of a small order. So if you're comparing your ROAS to an "industry average" without checking whether that average brand's AOV looks anything like yours, you're comparing apples to eyeshadow.
Channel matters as much as category. Here's how it typically breaks down for beauty brands running blended reporting across Meta, TikTok, Google, and Amazon.
Meta/Instagram
TikTok
Google Shopping/Search
Amazon Ads
Two brands, same revenue, wildly different "healthy" ROAS. Here's why.
Contribution margin is the real driver. A brand running 70% gross margin can turn a profit at 2x ROAS. A brand at 40% margin needs 4x or higher just to break even on the same spend. Neither one is doing anything wrong. They just have different cost structures.
Subscription and replenishment models, common across skincare, change the math further. A brand can accept a mediocre first-purchase ROAS because the real return shows up over months two, three, and four as the subscription recovers acquisition cost. Judging that brand's paid performance off day-one ROAS alone tells you almost nothing.
Influencer and affiliate-heavy brands run into a different distortion. Last-click ROAS on paid often looks low for these brands, not because paid is underperforming, but because a TikTok creator video did the actual persuading three days earlier and paid social just picked up the last click. The credit gets assigned to the wrong channel.
And promo periods lie. Launch week and holiday gifting sets inflate ROAS temporarily in ways that don't reflect what a normal Tuesday in March looks like. If you're benchmarking off your best week, you're benchmarking off a number you can't sustain.
Most bad ROAS benchmarks come from bad math, not bad marketing. A few of the usual culprits:
Platform-reported ROAS
Blended ROAS done wrong
Ignoring discounts, refunds, and shipping
Competitor comparisons with no context
Benchmarks are a starting point, not a destination. To actually know where you stand:
Pull spend and revenue from Meta, TikTok, Google, and Shopify into one blended view rather than trusting whatever number shows up in each platform's ad manager. Every platform is graded on its own homework here, and every platform gives itself a good grade.
Segment ROAS by new versus returning customer. This is the split that tells you whether paid is generating incremental demand or just picking off people who were going to buy anyway. A brand with strong blended ROAS but almost no new-customer contribution has a growth problem hiding behind a good-looking number.
Run your own numbers through the ROAS calculator to see where your blended figure lands against the ranges covered above. Takes a couple minutes and beats eyeballing it.
Then set a target based on your actual contribution margin, not an industry average pulled from a brand with a totally different cost structure. If you're running 45% margin, a "good ROAS" pulled from a 70%-margin skincare brand's press release is going to steer you wrong.
Most of the miscalculation problem above comes down to one thing: stitching together numbers from four different platforms by hand, in a spreadsheet, on a deadline. Trivas pulls Meta, TikTok, Google, and Shopify data into one Redshift-backed dashboard and calculates blended ROAS automatically, so you're not reconciling four exports every Monday morning.
The AI Wingman layer sits on top of that and flags it when ROAS on a specific channel or campaign drifts outside your own historical benchmark, which matters more than any industry-wide number. Your brand's normal is the only benchmark that actually predicts your next quarter.
If you're a founder or CEO trying to get a straight answer on marketing efficiency without living in spreadsheets, founders and CEOs is worth a look, and the BI reporting product page covers the dashboard this benchmark data feeds into.
If this kind of breakdown is useful, it's worth subscribing to keep the benchmark data coming as channels and algorithms keep shifting under everyone's feet.
explore more insights
3 min read
3 min read
3 min read
.d53b12e5.png&w=3840&q=75 "Trivas.ai")