In August, e.l.f. Beauty raised prices by $1 across every SKU in its lineup, a roughly 15% increase, citing tariffs that had reached as high as 170% and averaged about 55% over the fiscal year. It was a defensible, even conservative response to a real cost problem. Unit sales dropped anyway, enough that CEO Tarang Amin acknowledged on the company’s own earnings call that the business had “recently seen a more pronounced decline in units.”
What makes this worth studying isn’t that a price increase caused a sales drop, that’s not surprising. It’s the size and speed of the reversal on one product, and what that gap between “the increase was justified” and “the increase still cost us customers” actually means for how a business should price under real cost pressure.
The Tariff Math Was Real
e.l.f. wasn’t padding margin. The company paid roughly $58.5 million in tariffs before the US Supreme Court moved to strike down the relevant tariffs in February 2026, at rates that reportedly reached 170% on some goods and averaged around 55% across the fiscal year. A $1 increase on a company known for sub-$10 price points was, by the numbers, a proportionate response to real, documented cost pressure. Nobody could reasonably call it price gouging.
That’s exactly what makes the unit decline instructive. The price increase was rational at the balance-sheet level and still cost the company real volume, because the customer evaluating a $1 jump on a $7 product isn’t running the same tariff math the CFO is. They’re just noticing the price moved.
What The Reversal Actually Showed
e.l.f. didn’t roll back pricing across the board. It cut the price of one specific product, Halo Glow Skin Tint, from $18 to $14, a $4 reduction. Sales on that product rose 40% almost immediately afterward. That’s not a modest recovery, it’s a number large enough to suggest the original increase had been actively suppressing demand on that item specifically, not just slowing its growth.
The company later received roughly $50 million in tariff refunds after the Supreme Court ruling, and net income surged nearly 100% in the following quarter. Some of that recovery is refund timing, not pricing strategy. But the targeted cut, not a blanket rollback, applied specifically where the data showed the steepest drop, is the part worth copying regardless of how the tariff situation resolved.
What This Actually Means If You’re Not e.l.f.
Most SMB leaders will never negotiate tariff exposure at this scale. But the core decision, how to price under real cost pressure without guessing, applies at any size. Chipotle and Wingstop faced the same inflation and made opposite pricing bets, worth reading side by side with this one: see Chipotle’s Margins Fell to 23.4%. Wingstop’s Hit 25%.
A justified price increase can still be the wrong one. “Our costs went up” is a true statement and a business reason. It is not the same statement as “customers will accept this increase.” Track both separately, because only one of them shows up in your unit sales the following month.
Cut where the data is worst, not everywhere evenly. e.l.f. didn’t reverse the whole price list. It found the one product where the decline was steepest and fixed that specifically. A blanket rollback wastes margin on products that weren’t actually losing customers.
Watch the reaction speed, not just the direction. A 40% lift within roughly a quarter of the cut is fast enough to be a real signal, not noise. If a pricing change takes months to show any measurable reaction either way, you likely don’t have enough signal yet to know if it worked.
This is an Architect-stage decision underneath a Finance-beat story. A Pioneer sets prices by instinct because there’s no data yet to argue with. The Architect stage is where pricing has to become a system, a real feedback loop between cost pressure, unit economics, and actual customer response, checked often enough that a bad call gets caught in one earnings cycle instead of absorbed as a permanent new normal.
So here’s the real question: is there a price on your own list right now that you raised for a good, defensible reason, and have you actually checked whether the unit sales agree with your reasoning, or just assumed they would?