The brutal physics of furniture DTC.

No consumer category punishes bad operators like furniture, and no category has buried more DTC darlings. Bulky freight, double-digit damage and return rates, a purchase cycle measured in years, and a housing market frozen at a thirty-year low — furniture is where the brutal physics of physical goods meet the worst demand backdrop in a generation, now with a 25% tariff stapled on top. And yet some brands are thriving in exactly this environment. The split between who is compounding and who is collapsing is not about taste; it is about model. This is a deep dive into what actually works in interior and furniture e-commerce in 2026, who lived, who died, and why.

The macro: a frozen market with a tariff on top

Furniture demand is downstream of moving, and almost nobody moved. Existing-home sales hit roughly 4.06M in 2025 — a thirty-year low — with mortgage rates parked near 6.5% locking owners into their current homes and out of the new-sofa decision. As WWD put it, “as long as housing turnover is at near record-low levels, there's just less of a market for furniture.” The casualties were real: Conn's closed ~550 stores in 2024, American Freight closed all 328. Then came the tariff. About 61% of US furniture is imported — ~86% of wood furniture — mostly from Vietnam (~35%) and China (~24%), and in October 2025 the administration layered Section 232 tariffs of 25% on upholstered wood furniture and on kitchen cabinets and vanities, with planned escalations to 50%/30% pushed to at least 2027. RH alone cited a ~$45M revenue hit from tariff-driven backorders. For a category already squeezed on demand, the cost side got squeezed too.

The public scoreboard

The listed players tell you the model matters more than the category. The standouts are operating discipline and channel choice, not pure DTC.

Wayfair is the redemption arc — revenue +7.4%, its best Q1 adjusted-EBITDA margin in five years, active customers growing again after three rounds of layoffs taught it discipline. Williams-Sonoma is the quiet machine, with West Elm comping +8.5% and 16%+ operating margins. RH is something else entirely: Gary Friedman has turned it into a gallery-and-hospitality real-estate business — multi-level galleries in Milan and Paris, restaurants, guesthouses, even a luxury yacht charter — trading furniture retail for an ecosystem. And the mattress adjacency is the cautionary corner: Purple issued a going-concern warning, Sleep Number's sales fell ~19%, and even the Tempur Sealy / Mattress Firm combination (now Somnigroup) posted a revenue miss despite +12% headline growth.

The DTC-native cull

The pure-play DTC furniture cohort got sorted hard, and the sorting was financial. The deaths and distress — Interior Define ran its sales “off a cliff” in 2022 and was bought for its IP by Havenly, leaving ~$96M of liabilities behind; Modsy, the virtual-design service, shut down after raising $72.7M when a last-minute acquisition collapsed; Industry West is in distress with a default judgment, pivoting to trade-only. The acquisitions — Burrow (nine-figure revenue) and Floyd both got rolled up. And the survivors share a profile: Article has been profitable since 2015 and runs ~$500M; Castlery grew 63% to ~$180M and posts a profit; Maiden Home is made-to-order with triple-digit online growth. Note what every survivor is now doing: opening physical stores. Article, Castlery, Maiden Home, Lulu and Georgia — all reversing the pure-DTC thesis and adding showrooms, because furniture is a category people want to sit on before they spend $2,000.

Furniture didn't kill DTC brands because the brands were bad. It killed the ones whose unit economics couldn't survive bulky freight, a damaged-on-arrival rate, and a customer who buys a sofa once a decade.

Why the unit economics are so brutal

The structural problem is worth naming precisely because it explains the body count. Returns and reverse logistics are murder: US retail returns hit $890B in 2024, online return rates run toward ~19%, and home-goods carries the highest handling fees of any vertical — processing a furniture return can cost $55–90+ per item, multiples of apparel. Damage and white-glove delivery add $100–300 per item. And the lifetime-value math is unforgiving: a sofa lasts 7–15 years, so for most furniture brands “LTV often equals a single purchase” — against a DTC customer-acquisition cost that roughly doubled after Apple's 2021 privacy changes broke cheap Facebook targeting. A 3:1 LTV-to-CAC ratio, the baseline of a healthy DTC business, is nearly impossible when the customer never comes back. That is why the survivors are the made-to-order, low-inventory, profitable-or-bootstrapped operators — the model removes the working-capital and markdown risk that the category's physics impose.

What's actually working — including the AI layer

Four things separate the winners. Made-to-order / zero-inventory (Maiden Home, Inside Weather, Sabai) lifts conversion and AOV while slashing returns and working capital — though it's fragile to supply shocks, which is what actually killed Interior Define. Trade and designer programs are the margin engine: Arhaus reports designer-assisted orders carry ~4x higher AOV with ~40% becoming repeat, chasing a $27B US design-services market; RH runs a big chunk of revenue through its trade program. Financing and AR move the needle on big-ticket carts — Affirm reports up to 85% higher AOV, and Houzz's CEO said 3D “view in my room” users are 11x more likely to buy. And then there's the disruption layer: a wave of AI room-design tools (Spacely, Collov's $10M Series A, REimagine Home) and retailer-shipped visualization (Wayfair's Decorify and Muse, IKEA Kreativ, Lowe's Style Studio). The most interesting move is Havenly's — it trained its own free AI design tool on ~2.4M proprietary renderings and uses it as a funnel, routing complex jobs to human designers (and free users to paid tiers). The AI-in-interior-design market is projected to grow from ~$1.4B in 2025 to ~$4.6B by 2030. Design services, long a money-loser as a standalone (RIP Modsy), finally work — as a customer-acquisition layer on top of product sales.

Sources: WWD: housing & interiors; White House: Section 232 furniture tariffs; Wayfair Q1 2026; Retail Dive: Havenly/Burrow; Business of Home: Interior Define; TechCrunch: Modsy; NRF: returns; Business of Home: Havenly AI.