What I’m trying to understand about mattresses and furniture
I keep coming back to this question. A lot of the bull cases in the home category go something like: housing turnover is weak, consumers are stretched, the cycle bottoms, rates come down, existing home sales recover, people move, and mattresses and furniture snap back to a healthy baseline.
Some normalization is going to happen. That part is roughly right. The harder question is the magnitude and where the new baseline actually lands.
The setup: existing home sales are at a 30-year low. Movers spend roughly five times more on furnishings in their first year than non-moving owners do (NAHB). Total mattress unit purchases per household ran 0.20 to 0.26 from 2015 through 2021. They sit at 0.18 in 2024, slightly above GFC-trough levels (~0.15-0.16) on a per-household basis. The gap to a normal baseline is real, whatever “normal” turns out to be.
One more bullish version of this math, from a slide produced by a former board member/current investor and revealed in the FTC matter, makes the pent-up demand framing more concrete. They peg post-GFC under-trend spend at 25.6 million units, delayed household formation at 9.3 million, and the 2016 to 2022 above-trend bump at 26.6 million. Net remaining pent-up demand: 8.3 million units.
That’s a real thesis. The honest version of the bear case isn’t “demand is dead.” It’s: maybe the recovery is anemic. Maybe cycle relief is real but more mix-driven than volume-driven. Maybe replacement cycles have lengthened. Maybe affordability stays bad enough that turnover doesn’t fully unfreeze even if rates ease.
Looking at the bull and bear cases side by side, there are a few core facts they agree on:
Industry units are roughly 30 percent below the 2021 peak and 20 percent below 2019 levels.
Existing home sales are at a 30-year low.
Promotional intensity (0% APR, 60 to 72 month financing) is critical for driving demand.
COVID pull-forward is long gone.
Independent retailer attrition has been brutal (United, Mitchell Gold, BB&B, Conn’s, Big Lots, LL Flooring, Mattress Land, AFM, CVB, Metro Mattress).
Both sides also basically agree some recovery is coming. The fight is over magnitude:
Replacement cycle. Bulls assume roughly stable 8 to 10 years, so the under-shipped reservoir mathematically clears. Bears argue cycles have lengthened (Mordor pegs ownership stretching from 7 years toward 10; the Mattress Recycling Council’s average age of discarded mattresses is 11.2 years), and the reservoir does not unwind without an income or wealth catalyst.
Housing turnover. Bulls see it as cyclically frozen and waiting for rate cuts. Bears think it has structurally reset lower due to the difficulties (mainly affordability) of home ownership. RH’s framing is jarring here: in 1978 the U.S. sold 4.09 million existing homes with a population of 223 million. In 2024 we sold 4.06 million with a population of 341 million. Same units, 53 percent more people.
Trough framing. Management is converging on “we are at the bottom” (TPX NA same-store +5%, Mattress Firm +5%, Wayfair new orders +7%, Haverty’s three positive comp quarters). The same management is also guiding 2026 to flat-to-modestly-up off self-help, not market.
Channel measurement. Some operators (Bedgear, South Bay) argue ISPA materially understates real demand because of DTC, putting “true” total market closer to 48 to 52M units vs. ISPA’s ~36M (!!!). ISPA participants counter the numbers are directionally right; Purple’s Q4 2025 e-commerce was down 15 percent.
The biggest disagreement, the one most of the others depend on: is the cumulative under-shipment of mattresses deferrable or destroyed? Bulls treat it as a reservoir. Bears treat it as wash, because consumers replace later from a longer-life base (or perhaps already did with mattresses not caught in ISPA data).
I don’t have a confident view on the magnitude.
What would push me one way or the other: sustained sub-6% mortgage rates with a durable rebound in existing home sales, ISPA mattress unit growth that is unit-led not ASP-led, and management commentary shifting from value/promotional/lower-traffic to sustained units with reduced promotion. If we get any of those without the others, I want to know which one is doing the work.
Appendix: how I tried to answer this
I ran the same deep research prompt through both ChatGPT and Claude. ChatGPT came back with a clearer point of view (partial cyclical recovery, not a simple snapback). Claude did not really take a stance, but its decomposition of where bulls and bears actually disagree was useful and I leaned on it above.
Including the prompt below in case anyone wants to run a version of this themselves.
The prompt
I am researching the mattress and furniture industry, with a focus on whether the category is likely due for a meaningful cyclical snapback or whether forward growth is more likely to be structurally muted/anemic.
The core question: are mattresses and furniture genuinely set up for a major recovery as housing turnover normalizes, or has the category’s normalized growth rate/volume baseline changed in a way that makes the “snapback” thesis too simplistic?
Please research this deeply and produce an investment-oriented memo.
Companies to keep in mind: Somnigroup / Tempur Sealy / Mattress Firm, Leggett & Platt, Havertys, Arhaus, RH, Wayfair, La-Z-Boy, Ethan Allen, Williams-Sonoma / Pottery Barn / West Elm, Ashley Furniture (private), Serta Simmons / Beautyrest, Purple, DreamCloud, Nectar, Resident, Saatva, Casper.
Key questions:
Historical demand relationship to existing home sales, new home sales, housing starts, household formation, mortgage rates, affordability, consumer confidence, disposable income, credit availability, remodeling, and moving activity. What share of demand is move-driven vs. replacement? How long are real replacement cycles and are they lengthening? Did COVID pull-forward suppress replacement for years?
Is there a real deferred-demand pool, or has demand been lost? How does big-ticket home compare to autos as a deferral category?
What is “normal” volume now? Use ISPA, Census, retail sales, trade association data. Estimate whether normalized real volumes are below, equal to, or above 2019.
Structural changes: DTC, online price transparency, value brand share, financing availability, household formation, smaller homes, mattress-in-a-box, Mattress Firm post-Somni dynamics, private-label mix.
Housing turnover sensitivity. How much upside if existing home sales normalize? How much is already priced into bull cases? Can replacement recover independently?
Retailer and manufacturer commentary on traffic vs. ticket, units vs. ASP, premium vs. value, financing, promotions, inventory, replacement, housing, move-related demand, trade-down, online vs. store, recovery timing.
Industry-level data sources (ISPA, Census, BEA, FRED, NAR, JCHS, NAHB, transaction data summaries).
The strongest version of the bull case and the 3-5 data points that would confirm it.
The strongest version of the bear case and the 3-5 data points that would confirm it.
Investment implications by company, without giving investment advice.
Output: executive summary, bull case, bear case, evidence table (point/source/supports/why/reliability), company exposure table, key charts to read, follow-up questions, source list. Be skeptical. Treat “housing depressed = snapback later” as the hypothesis to test, not the assumed answer.
ChatGPT executive summary (verbatim)
The weight of the evidence supports a partial cyclical recovery, not a simple snapback. Housing turnover is unquestionably suppressing demand: existing-home sales remain near multi-decade lows, mortgage rates are still above 6%, and movers spend sharply more on furnishings than non-movers. But the category is not just waiting for housing to normalize. Public data and management commentary point to a market that has also changed structurally: COVID pulled forward demand, mobility has been falling for years, replacement cycles appear to be lengthening, online and value channels have increased price transparency, and several operators are winning through share gain, assortment, and channel mix even while the category itself remains soft. My base case is that mattresses recover earlier and cleaner than furniture, but both categories likely normalize to a slower, more mix-driven, more promotional baseline than the bullish “snapback” narrative usually implies.
Confidence: moderate. ISPA’s full data sit behind member-only reports, and there is no clean public source that splits demand into mover/new-household versus replacement.
What would change the view: existing home sales rising materially for several quarters, ISPA mattress units turning clearly positive on a real-unit basis, and company commentary shifting from “choppy / lower traffic / value-oriented” to sustained unit growth with lower promotions and better premium mix.