What the global bed in a box brands learned the hard way
18 March 2026|Written by: Dream Chronicler
In February 2020, Casper Sleep listed on the New York Stock Exchange at a valuation of around US$575 million. That was already half what private investors had once been willing to pay for the company. Less than two years later, in November 2021, it was taken private at US$286 million, half the IPO valuation again. By the time the deal closed in early 2022, Casper had managed, in seven years, to lose roughly three quarters of the value that the bed in a box thesis had originally assigned to it.
Casper was not alone. Eve Sleep floated on London’s AIM in 2017 at a £140 million valuation. By October 2022 it was in administration. Bensons for Beds, a UK retailer with around 166 physical stores, paid £600,000 for the brand, the website, and the intellectual property. Made.com floated on the London Stock Exchange in June 2021 at a £775 million valuation. By November 2022 it too was in administration. Next, the multinational clothing and homewares retailer with both physical stores and online presence, paid £3.4 million for the assets. Noa Home, a Canadian bed in a box brand active in Australia from 2016, withdrew from the Australian market entirely in 2024, while continuing to operate in Canada, the United States, Singapore, and the United Kingdom.
Last month we wrote on this site that the bedding industry has spent the last five years watching other categories get pulled online by brands who built direct channels at the expense of their retail networks, and that a lot of that went badly. This piece is the detailed evidence behind that observation. The pattern is sharper than the headline numbers suggest. And the lesson the survivors are now teaching is sharper still.
What the bed in a box thesis was supposed to do
Around 2014 to 2018, a wave of new bed brands shared roughly the same plan. Cut out the wholesaler. Cut out the retailer. Compress a queen mattress into a box. Ship it from a low cost manufacturing base. Sell it on the internet to a customer who would never have to walk into a store. The savings on the channel, the argument went, would more than fund the customer acquisition cost of building an online brand from zero.
For a few years, that argument looked like it was working. Casper raised over US$300 million in private capital before its IPO. Eve Sleep was the most visible new brand on the London Underground for two summers running. Made.com had Brent Hoberman, one of the founders of lastminute.com, on its founding team. The UK and US tech press wrote about all three as the future of furniture and bedding retail.
Then the unit economics caught up.
What actually happened
Online customer acquisition costs, which were low when these brands started, rose sharply from 2018 to 2022 as more entrants competed for the same searches. Freight costs rose through the COVID supply chain disruption and stayed elevated. Customers who had been willing to buy a $1,500 mattress sight unseen in 2016 became more cautious as the category matured.
The arithmetic on these brands started looking different. A retail consultancy’s analysis of Made.com’s final accounts found that on an average order value of £246, the gross profit per order was £97, the fulfilment cost per order was £45, and after marketing costs there was £14 of contribution left. In the first half of 2022 that £14 disappeared entirely as freight and energy costs spiked. Eve Sleep’s chief executive described the same period as an “economic tsunami”. Casper, which never reached profitability in any quarter of its public listing, simply ran out of patient capital.
Some of this is a story about a particular set of macro shocks. More of it is a story about a structural mismatch between what the original D2C thesis assumed and what the mattress market turned out to want. The thesis assumed that price was the binding constraint on bedroom purchases, and that a price advantage delivered through channel disintermediation would compound into market share. What actually happened was that price was one consideration among several, that the customer’s confidence in the product mattered as much as the price, and that confidence was much harder to build through online channels alone than the founders had projected.
There is a second pattern visible in the failures, and it is worth naming directly because it is the pattern we have been talking about most with our retail partners over the last few months. In the years where these brands were chasing growth, several of them started competing on price against each other. Casper cut prices to defend share against new entrants. Eve Sleep ran heavier and heavier discount cycles. Made.com expanded its range and discounted aggressively to keep volume up. The pattern is familiar and the name for it is also familiar. It is a race to the bottom. And the businesses that ran that race ran out of margin before they ran out of competitors.
What the survivors did differently
Not every brand from the original wave failed. The ones that survived did something specific, and the same thing keeps showing up in different markets.
Koala, the Australian D2C brand that started in 2015 with the same general plan as Casper, paused a planned A$100 million IPO in 2025 and instead opened its first permanent physical showroom in Sydney’s Moore Park Supa Centre in late 2025. The showroom was designed by IF Architecture, the firm behind Attica restaurant and the Jardan furniture showrooms. It is not a half measure or a small temporary site. It is a serious investment in the kind of physical presence that the original Koala plan said was unnecessary.
Ecosa, an Australian brand that had been online only since 2014, opened three physical stores in Australia in 2025: Geelong, Perth’s Midland Mega Centre, and Nunawading in metro Melbourne. The Nunawading store occupies a 2,000 square metre space and is described by its designers as a flagship. Ecosa’s CEO summarised the strategic shift in a recent interview. “We’re probably seen today as an online mattress brand, not a sleep destination.”
The Koala and Ecosa pivots are not panicked retreats. Both brands continue to be predominantly online businesses. Both are still profitable enough to fund expansion. What they have done is recognised that for a category like beds, where the customer is being asked to spend $1,500 to $5,000 on a product they will use every night for a decade, the option to lie down on the product before they buy it is a structural part of the offer rather than an optional extra. The brands that built only the online half of the customer journey ran out of growth before they ran out of money. The brands that built both halves, even if one of those halves was added late in their development, kept going.
This is the pattern we wrote about back in November when we said that the answer the market has been working toward over the last five years is not “online wins”, it is closer to “the businesses with both legs win”. Koala’s decision to invest in a permanent showroom in 2025 is the most public confirmation of that pattern. The internet starts the conversation. The showroom finishes it.
The lesson is sharper than “online failed”
The lesson that bed in a box brands have taught the global market is not that online is the wrong channel. It is not that physical retail is automatically the right channel either. It is that channel structure matters less than channel alignment. A brand that only sells online can succeed, but only if the customer it is built for is genuinely served well without ever touching the product. A retailer that only sells in a showroom can succeed, but only if the customer it is built for is genuinely served well without ever doing research on a phone. And almost no brand in this category, at least in the last decade, has been able to grow at scale without doing both.
To put it more directly: online is not the enemy. Lack of alignment is. The brands that failed did so not because they sold online but because the model they operated was not aligned with the customer they were trying to serve, the supply chain they were running on, or the cost base they were carrying. The brands that survived are the ones that adjusted the model when the data showed the alignment was off.
Smiths City and Kitchen Things, which we wrote about a fortnight ago, failed for a structurally similar reason from the opposite direction. Their cost base was sized for a different volume of customer traffic than what was walking through the door. Casper, Eve, Made.com, and Noa Home failed because their channel structure was misaligned with the customer they ended up serving. The diagnostic is the same in both cases. The match between the model and the customer is the determining variable. Channel is downstream of that.
What this means for the question we have been asking
For most of this News & Updates thread we have been circling around a question without quite asking it directly. The question is whether the model the New Zealand bedding industry has used for the last twenty years is the right model for the next ten. The Smiths City and Kitchen Things stories say that elements of the traditional model are under more pressure than the headlines have acknowledged. The Casper, Eve, Made.com, and Noa Home stories say that the easy alternative, going pure online and skipping the physical channel, has not worked at scale either. The Koala and Ecosa stories say that the most successful adapters are the ones who treat the question as one of alignment rather than channel.
That is the question we have been asking. We will say more about how we think about it ourselves in the coming weeks.
More from us in early April. We will look at what March’s data is telling us about the recovery, and at what it means for the way we have been thinking about pricing through this year.
Sources cited: CNBC, CNN Business, Fast Company, and Sourcing Journal coverage of Casper Sleep IPO February 2020 and Durational Capital acquisition November 2021; Retail Gazette, Big Furniture Group, and Shares Magazine coverage of Eve Sleep AIM listing 2017 and administration October 2022; New Statesman, TechCrunch, Retail Bulletin, and Wikipedia coverage of Made.com IPO June 2021 and administration November 2022; Noa Home Australia closure announcement 2024; The Design Files, Sandbox Group case study, and Inside Retail Australia coverage of Koala Moore Park Supa Centre showroom opening late 2025 and Ecosa physical store rollout 2025; Inside Retail Asia coverage of Koala A$100 million IPO pause June 2025.
