Sears Roebuck: The Amazon of 1900 That Took 130 Years to Die

In 1893, a young Minnesota railroad agent named Richard Sears partnered with a watchmaker named Alvah Roebuck to expand a small mail-order pocket-watch business he had been running on the side. Within five years, the catalog they produced together had grown to over five hundred pages and offered, by some counts, more than ten thousand different products — clothing, tools, furniture, sewing machines, bicycles, kitchen ranges, eyeglasses, prefabricated houses, and a substantial portion of every other consumer good then manufactured in the United States. The catalog was distributed to roughly twenty million American households at its peak. For the rural majority of Americans, who lived too far from any town to shop at a department store, the Sears catalog was the entire consumer economy.

By the time the company filed for Chapter 11 bankruptcy in October 2018, Sears had been the largest retailer in the world (briefly, in the 1970s), the largest seller of insurance in America (through its Allstate subsidiary), the largest seller of major appliances in America (through its Kenmore brand), and the operator of one of the largest real-estate portfolios held by any American non-financial company. Each of these positions was sold off, divested, or wound down across the four decades preceding the bankruptcy.

The story of how the original Amazon — a company with a half-century head start on every category Amazon would eventually dominate — failed to become Amazon, is the most expensive case study in retail history. It is worth understanding because the structural conditions that produced the failure are not unique to retail.

What Sears actually was, in 1900

The Sears Roebuck catalog of 1900 was, in modern terms, an e-commerce platform with no e. The customer received a thousand-page catalog (heavier than most modern phone books), browsed it at home, filled out an order form, sent the form along with payment to Sears headquarters in Chicago, and received the goods by mail. The fulfilment infrastructure that made this possible — Sears's enormous Chicago warehouse, the standardised shipping operations, the relationships with railroads and parcel carriers — was, for its time, a logistical achievement comparable to anything Amazon has built in the last twenty years.

The customer base was specifically rural. American cities in 1900 had department stores. American farmers and small-town residents did not. For the roughly seventy percent of Americans who lived outside meaningful urban shopping districts, the Sears catalog was the only practical access to most consumer goods. The catalog's pricing was, in many categories, twenty to forty percent below what local general stores could offer, because Sears was buying directly from manufacturers in volume and shipping nationally. The local general stores hated the catalog. They organised boycotts. They pressured manufacturers not to sell to Sears. None of it worked, because the customers had figured out that the catalog was cheaper, larger, and more convenient than anything else available to them.

By 1908, Sears was selling, through the catalog, complete prefabricated houses — flat-packed, shipped by rail, with detailed instructions for assembly. Approximately seventy thousand of these houses were sold over the next thirty years, of which perhaps thirty thousand are still standing in American suburbs today. The houses were, in a literal sense, mail-order. The catalog had become the platform through which Americans bought essentially anything that could be packed and shipped.

The transition that worked, in 1925

The structural change that should have killed Sears the first time happened in the early 1920s. Cars and improving roads were dramatically expanding the geographic range of urban department stores. The rural-only catalog model was being eroded by customers who could now drive into town to shop. Sears's catalog revenues, which had been growing for decades, began to plateau and then decline.

The company's response, under a new CEO named Robert Wood (who joined from Montgomery Ward in 1924), was to open physical stores. The first Sears retail store opened in 1925 in Chicago. The chain expanded rapidly through the late 1920s and 1930s, focused on suburban and small-city markets where Sears could undercut the existing department stores using the same supply-chain advantages that had powered the catalog. By 1940, Sears had opened approximately six hundred stores. The catalog continued to operate alongside them. The combined model was, throughout the 1940s and 1950s, the most successful retail format in American history.

This is the part of the Sears story that is most often forgotten. The company did successfully transition once, from catalog to physical retail, when the underlying technology change required it. It did so because the leadership in 1925 was willing to invest in a new business model that would, in the long run, cannibalise the existing one. The transition took two decades and substantial capital but it was managed.

The transition that did not work

The technology change that finally broke Sears, in the early 1990s, was the inverse of the one it had successfully managed seventy years earlier. Walmart had, by 1990, built a national chain of large-format discount stores in small cities and rural areas — exactly the markets Sears had built itself on. Walmart's supply chain, more efficient than Sears's, allowed it to undercut Sears's prices in the same way Sears had once undercut general stores. By 1993, Walmart's revenues exceeded Sears's for the first time. By 1997, Walmart was permanently larger.

The internet was a smaller cloud on the horizon at the same moment. Amazon was founded in 1994, selling books out of a Seattle garage. Sears was, in 1994, the second-largest retailer in America, with a hundred-year history of mail-order distribution, an existing relationship with tens of millions of American households, brand recognition that Amazon would not match for another decade, and a category catalog (the Wish Book and the main catalog) that was, in concept, identical to what Amazon was about to become — but rendered on paper instead of on a screen.

The catalog itself, by 1993, was unprofitable. Sears's response was to discontinue it. The general catalog, which had been continuously published since 1888, was killed in January 1993. The Wish Book continued for a few more years before also being discontinued. The decision was, on the catalog's own profit-and-loss statement, defensible. The catalog was losing money in a particular year. Cutting it improved current profitability.

What the decision sacrificed, which the income statement could not show, was the entire infrastructure and customer relationship that would have positioned Sears to compete in e-commerce. Sears had, in 1993, the customer database, the fulfillment network, the supplier relationships, the credit infrastructure, and the brand to be Amazon. It chose, in the same year Amazon was founded, to dismantle the channel that connected all of those assets to its rural and small-town customer base.

The forty-year unwinding

Through the late 1990s and 2000s, Sears made a series of sequential decisions that — each individually defensible against the company's quarterly earnings — collectively accelerated the company's irrelevance.

The Allstate insurance business was spun off in 1995. The financial services arm (Dean Witter, Discover Card) had been spun off in 1993. The Sears Tower in Chicago was sold in 1994. The credit-card portfolio was sold to Citigroup in 2003. The Kenmore appliance brand, which had been an exclusive Sears product line for nearly a century, was eventually licensed for sale through other retailers (including Amazon, in 2017) — a decision that turned what had been a competitive moat into a commodity good available everywhere.

In 2005, Sears merged with K-Mart, in a transaction structured by hedge fund manager Edward Lampert as primarily a real-estate play. The combined company controlled enormous quantities of valuable retail real estate. Lampert, who became CEO, focused increasingly on extracting value from the real estate (selling off store locations, restructuring leases, spinning off real-estate-investment-trust subsidiaries) rather than on operating the retail business. The retail business itself was systematically underinvested in. Stores became visibly run-down. Inventory became inconsistent. Employee headcount was cut repeatedly. The company became, by the mid-2010s, a real-estate vehicle wearing the costume of a retailer.

The 2018 bankruptcy filing was, by that point, an administrative formality. The retail operation had been dying for two decades. The real-estate value had been extracted. There was very little left to bankrupt.

Why the case is harder than it looks

Sears is sometimes presented as a story about complacency — a once-great company that simply failed to keep up with Walmart and Amazon. This framing is wrong in an instructive way. Sears was not complacent. The company actively, repeatedly, made decisions to optimise its current financial performance. Each decision, at the moment it was made, was defensible against the metrics the company was being judged on. The cumulative effect of these decisions, however, was to systematically dismantle the assets that would have made the company competitive in the future.

The catalog was unprofitable in any given year — but it was the channel that connected Sears's customers to its supply chain. Killing it removed a money-losing line item but eliminated the architecture for e-commerce.

The credit-card portfolio was a non-core financial asset — but it was the customer-payment infrastructure that would have made Sears a natural fintech platform. Selling it produced a one-time gain but eliminated a recurring relationship.

The Kenmore brand was being underutilised — but its exclusive distribution through Sears was the loyalty asset that brought customers back into stores. Licensing it to competitors freed up revenue but eliminated the differentiator.

Each individual divestiture made sense on its own. The pattern made no sense at all. But pattern recognition requires a leadership horizon longer than the quarterly cycle that incentivised each individual decision. Sears, after the late 1980s, did not have that horizon.

The lesson that recurs

The deepest lesson Sears teaches operators is that the difference between a temporary downturn and a permanent decline is almost always invisible at the moment the relevant decisions are being made. A profitable company can dismantle the assets that produce its future profits without ever having a single quarter of catastrophic loss. Each quarter looks fine. The trajectory only becomes obvious in retrospect.

Sears in 1993 had, by any reasonable accounting, the assets to become Amazon. The leadership in 1993 chose to optimise the wrong things, in ways that looked sensible at the time, against metrics that did not measure what mattered. Twenty-five years later, the company was bankrupt. There was no single decision that caused this. There were a hundred small ones, each of which traded a long-term competitive position for a short-term financial benefit.

The pattern is recognisable in many other industries. The newspapers that killed their classified-ads business when it was still profitable, only to find that the classifieds had been the customer-acquisition channel for the entire publication. The traditional broadcasters who licensed their content libraries to streaming services for short-term cash, only to discover that the libraries were what would have made them streaming competitors. The retailers who outsourced their warehouses for cost savings, only to find that the warehouses were the asset that would have let them compete with Amazon's logistics.

In each case, the decision was financially rational on a one-year horizon. In each case, the decision was financially fatal on a ten-year horizon. The Sears case is the cleanest example because the company had the longest possible head start and made the longest possible series of these decisions. It is also a useful reminder that the original Amazon — the company that pioneered the catalog-to-doorstep model that Amazon has spent thirty years rebuilding — was not killed by Amazon. It was killed by its own failure to recognise what it was, at every decision point that mattered.