Open in interactive viewer → charts, metric popovers & call review

Somnigroup's Best Quarter Ever Came With a Haircut — and a Promise It Already Broke

Record sales, record EBITDA, record EPS — and a guidance cut built on an ERP problem management told investors three months ago was 'maybe another one quarter.'
SGI · Earnings Call · 2026-11-09

On the numbers, Somnigroup International just had the quarter of its life. Net sales of $2.4 billion, adjusted EBITDA of $448 million, adjusted EPS of $1.04, up 10% year over year. CFO Bhaskar Rao did not hedge the framing: “This was the best performing quarter in the company's history.” — Bhaskar Rao, Executive Vice President and Chief Financial Officer · 2026-11-09 The stock, though, has been voting the other way — down 15.3% over the last 90 days and 32.4% below its February peak of $97.99, per the company tape. The last filed 10-Q still showed only $1.8 billion of quarterly revenue, down 3% year over year, with gross margin at 44.8% and a $4.2 billion net debt position — a snapshot that predates the record print and the Leggett & Platt consolidation, and one that makes the gap between operating reality and the equity story unusually wide. That is the setup: a business running better than its own filed financials, in a stock that has already decided it doesn't believe the next two years.

A record quarter with a footnote

The footnote is the guide. Full-year adjusted EPS was trimmed to $2.85–$2.95. Bobby Griffin of Raymond James put the damage plainly: “I think there was about a $0.20 reduction at the midpoint for the guide.” — Robert Griffin, Analyst at Raymond James · 2026-11-09 Bhaskar's reconciliation was refreshingly unvarnished — roughly $25 million of EBITDA lost in the back half to the U.K. operation, about $15 million of transportation inflation across all business units, and then the residual: “the industry just a bit softer than what we had expected.” — Bhaskar Rao, Executive Vice President and Chief Financial Officer · 2026-11-09

Valuation has already absorbed a lot of this. At the last filed quarter SGI traded at 17.3x operating income, down 44% year over year from a 30.9x peak in the third quarter of 2025 — the de-rating happened before this miss, which is what makes the report interesting rather than merely disappointing. The market didn't need a bad quarter to get defensive; it got there on the industry.

The ERP excuse that expired

Here is the tension that matters. On the August call, Scott Thompson was asked directly about the Dreams ERP implementation and whether the complexity of the growing empire was becoming a problem. His answer was dismissive in the way only a confident operator can be:

At most, that's maybe another one quarter. That's not a big issue going forward from my perspective. I think it was just a little choppy.

Scott Thompson, Chairman, President, and CEO · 2026-08-06

Bhaskar added that “the system is functioning. We're just working through it.” — Bhaskar Rao, Executive Vice President and Chief Financial Officer · 2026-08-06 Three months later, Dreams posted a 35% decline in reported sales, international net sales fell 12% on both a reported and constant-currency basis, and the international operating margin collapsed 870 basis points to 9.4%. Scott's own accounting, in the prepared remarks, was blunter than any analyst could have been: “Effective July 1, Dreams completed the rollout of a new ERP platform... Transitioning from a legacy infrastructure to a modern ERP system was necessary for long-term growth and enhanced security, but the near-term execution fell short of our plan.” — Scott Thompson, Chairman, President and CEO · 2026-11-09 The company pegs the cost at roughly $50 million of sales and $25 million of adjusted EBITDA, with recovery expected by early Q1 2027 — a timeline that will be read with appropriate skepticism given the last one. The keyword list tells the story of how fast a dismissed issue becomes a top-30 theme: ERP issue appears in this quarter's extraction after Dreams itself was a high-momentum name three months ago.

And here is the detail that deserves a raised eyebrow. The quarter also booked roughly $20 million of tariff refunds, arriving early relative to plan, which reduced cost of goods sold. Scott noted these effectively offset the Dreams shortfall, making the two one-time items enterprise-neutral. A genuine operating failure was, arithmetically, laundered through an unrelated policy windfall. Investors should treat the clean headline EPS with the appropriate discount.

Bridge to $3.40 — and the levers left

The most consequential sentence of the call was not about Dreams; it was about the shape of 2027. Scott laid out an unusually explicit bridge: $2.90 midpoint, plus $0.25 of Leggett wraparound accretion, plus $0.25 of scheduled synergies. “And so call it a run rate exit that we're targeting for '27 of $3.40,” — Scott Thompson, Chairman, President and CEO · 2026-11-09 he said, before pre-empting the obvious arithmetic — $3.40 to the $5.15 aspirational 2028 target implies a 23% two-year compound. He conceded the admission that matters: “clearly, we're having to use different levers in the business to hit those numbers.” — Scott Thompson, Chairman, President and CEO · 2026-11-09 One of those levers is price. With commodities unrecovered, Scott flagged that “I would anticipate that we would have a price increase sometime in the first half of next year.” — Scott Thompson, Chairman, President and CEO · 2026-11-09 That is a company leaning on its own pricing power to defend a target built on an industry recovery that has now been deferred for at least four consecutive quarters of calls.

Leggett, Stearns, and the waves underneath

What is genuinely new is the shape of the enterprise. Leggett & Platt closed August 26 and is now a fourth reporting segment; the company restructured into direct bedding, wholesale bedding and furnishings, and specialized products; the associate count in the closing remarks jumped from the familiar 20,000 to over 36,000, across 160 manufacturing facilities in 35 countries. The strategic logic is real — vertical integration of innersprings and foam, now branded on the outside of the mattress, and used to underwrite Mattress Firm's new component-qualification standards.

The product cycle is doing the near-term lifting. The new Stearns & Foster lineup is generating roughly 6% incremental slots taken from competitors, not traded between Sealy and Stearns — the single most-quoted keyword of the quarter for this company. A heritage collection at $7,499–$10,999 per queen arrives in late 2027, explicitly requested by third-party retailers. And Mattress Firm, which has drifted from over 3,400 doors to 2,150, is going back on offense: “the opportunity is more like 2,500,” — Scott Thompson, Chairman, President and CEO · 2026-11-09 with roughly 18-month enterprise paybacks. That plan was telegraphed in August — “I expect that we'll be back in store growth mode here pretty soon” — Scott Thompson, Chairman, President, and CEO · 2026-08-06 — so it is a follow-through, not a surprise, and it now has a number attached to store count.

The macro waves are not proprietary. The transportation cost squeeze is a global, cross-sector theme this quarter — diesel and fuel costs show up in the market's keyword set and across airlines, truckers and distributors. The tariff refund is even more clearly a market-wide cycle with dozens of companies harvesting IEEPA refunds; SGI is a small participant, not a pioneer. And the demand question that analyst Peter Keith raised — agentic search reshaping how consumers discover mattresses — got an answer that was candid about the lag: “we've noticed it, noticed it for about 1.5 years or 2 and are working on it.” — Scott Thompson, Chairman, President and CEO · 2026-11-09 Scott's bet is that scale eventually wins the AI funnel, just as it eventually won the bed-in-a-box era. That is a plausible read, but it is also the third consecutive call in which the answer has been 'working on it.'

Somnigroup is using the worst bedding market in North American history to build a vertically integrated global platform — that much is credible, and record EBITDA in a shrinking industry supports it. The open question is whether $5.15 in 2028 survives contact with a consumer whose entry-level tier has simply stopped shopping, a U.K. integration that overran its own one-quarter promise, and a management team now explicitly admitting it needs new levers. Record quarters and broken timelines can coexist; the tape says the market is pricing the second one.